The Behavioral Economics Guide 2021

ISSN 2398-2020

Author information:

Alain Samson (Editor)

John A. List (Introduction)

Siri Chilazi & Iris Bohnet (Guest editorial)

Joshua Henk Balsters (Gorilla Experiment Builder), Florian Bauer (Vocatus), Elizabeth F. Beach
(National Acoustic Laboratories), Paloma Bellatin (The Behavioral Insights Team), Radhika Bhalla
(Quadrangle Consulting Services), Philipe M. Bujold (Center for Behavior & the Environment, Rare),
Monserrat Bustelo (Inter-American Development Bank), Chloe Bustin (The Behavioral Insights Team),
Christopher F. Chabris (Geisinger), Benny Cheung (Dectech), Colm Creaner (GetReskilled), Gerard
Creaner (GetReskilled), Sinead Creaner (GetReskilled), Harry Davies (Frontier Economics), Brent Edwards
(National Acoustic Laboratories), Jo Evershed (Gorilla Experiment Builder), Maria Ferreira-Sequeda
(ING), Prasad G (Quadrangle Consulting Services), Megan Gilliver (National Acoustic Laboratories),
Amir Goren (Geisinger), Nidia Hidalgo (Inter-American Development Bank), Kalina Kasprzyk (Frontier
Economics), Kate Laffan (OECD), Abbie Letherby (Dectech), Daniela Marconi (IAG), Nimit Marwaha
(Quadrangle Consulting Services), Michelle N. Meyer (Geisinger), Roger Miles (UK Finance), Faisal
Naru (OECD), Paula Papp (Frontier Economics), Mitesh Patel (University of Pennsylvania), Alice Pearce
(Dectech), Shruthi Ramakrishnan (Quadrangle Consulting Services), Wieke Scholten (UK Finance), Natan
Sklair (Old Mutual), Nathalie Spencer (IAG), Laura Straeter (ING), Erik Thulin (Center for Behavior & the
Environment, Rare), Manuel Wätjen (Vocatus), Mónica Wills Silva (The Behavioral Insights Team), and
Claire Wilson (GetReskilled)

(Contributing authors)

Andreas Haberl (Editorial assistant)

Laeeq Hussain Arif (Cover design)

Published on www.behavioraleconomics.com by Behavioral Science Solutions Ltd

Copyright © by the authors

All rights reserved

Requests for permission to reproduce materials from this work should be sent to
[email protected] or directly to contributing authors

Suggested citation:
Samson, A. (Ed.)(2021). The Behavioral Economics Guide 2021 (with an Introduction by John List). https://
www.behavioraleconomics.com/be-guide/.

Suggested citation for individual chapters/articles:

[Author(s)] (2021). [Chapter/Article Title]. In A. Samson (Ed.), The Behavioral Economics Guide 2021 (with
an Introduction by John List)(pp. nnn-nnn). https://www.behavioraleconomics.com/be-guide/.
In Association With

Contributions By

Further Support By
Table of Contents

INTRODUCTION �������������������������������������������������������������������������������������������������������������������������������������������������������� V

The Voltage Effect in Behavioral Economics (John List)������������������������������������������������������������������������������������� VI

PART I - EDITORIAL�������������������������������������������������������������������������������������������������������������������������������������������������� 1

Behavioral Economics: Promoting Diversity, Equity, and Inclusion Through Goal-Setting

(Iris Bohnet & Siri Chilazi)������������������������������������������������������������������������������������������������������������������������������������������������� 2

PART II - APPLICATIONS������������������������������������������������������������������������������������������������������������������������������������� 14

Behavioral Insights and Gender-Based Violence Prevention: Evidence From a Facebook Trial
in Honduras                                                                         15
Sticky or Not? How COVID Has Changed Consumer Behaviour in Financial Services, and What
Might Happen Next                                                                   30
Using Behavioural Economics to Drive Better Outcomes in General Insurance��������������������������������������40
Brick by Brick: How Targeted Messaging Can Help Customers With Low Savings Build a Buffer����������49
Increasing Retirement Savings in South Africa ������������������������������������������������������������������������������������������58
Nudging to Improve the Timely Payment of Loans: Results of a Natural Field Experiment�������������������65
Behavioural Economics Is Booming in Banking: Should You Join In?��������������������������������������������������������78
Tired of Behavioral Economics? How to Prevent the Hype Around Behavioral Economics From
Turning Into Disillusionment������������������������������������������������������������������������������������������������������������������������85
Crowdsourcing Nudges: Insights and Experiences From Two Healthcare Systems�������������������������������.94
Applying Behavioural Economics to Improve Hearing Aid Conversations Between First-Time Clients
and Clinicians: The Development and Evaluation of Interventions to Increase Informed Choice ������104
BI Org: Fostering a Behavioural Mindset at the OECD ����������������������������������������������������������������������������115
Complex Behavioral Challenges Require Multi-Faceted Behavioral Solutions: Driving Change in
Sustainable Agriculture ����������������������������������������������������������������������������������������������������������������������������122
Gamification in Behavioral Science: An Engaging Prospect for Online Research ��������������������������������132
No Contract, No Problem: The Value of Flexibility in the Subscription Economy ��������������������������������142
Career Coaching Builds Confidence in Experienced Workers: What Governments Should Be
Thinking About for Reskilling Initiatives Post-COVID – and Why ����������������������������������������������������������153

PART III - RESOURCES ��������������������������������������������������������������������������������������������������������������������������������������� 165

Behavioral Science Concepts ������������������������������������������������������������������������������������������������������������������166

Postgraduate Programs����������������������������������������������������������������������������������������������������������������������������218
Other Resources����������������������������������������������������������������������������������������������������������������������������������������228

APPENDIX����������������������������������������������������������������������������������������������������������������������������������������������������������������� 229

Author Profiles ������������������������������������������������������������������������������������������������������������������������������������������230

Contributing Organizations����������������������������������������������������������������������������������������������������������������������232
Acknowledgements

The editor would like to thank John List for writing the introduction to this edition, as well as Siri
Chilazi and Iris Bohnet for their guest editorial.

We are grateful for this year’s partnership and support from the Think Forward Initiative and the
support received from Behavior & Law, The Behavioral Insights Team, Dectech, Frontier Economics,
Geisinger, GetReskilled, Gorilla Experiment Builder, IAG, ING, National Acoustic Laboratories, OECD,
Old Mutual, Quadrangle Consulting Services, Rare, UK Finance, Vocatus, as well as the Chicago
School of Professional Psychology, London School of Economics and Political Science, University of
Pennsylvania, and University of Warwick.

We would also like to thank Shivaun Anderberg, Andreas Aristidou, Susan Broomhall, Edouard
Camblain, Megan Crawford, Antonio Diaz-Guy, Konstantinos Georgalos, James Healy, Alanna
Holowinsky, Lindsey Horne, Chelsea Hulse, Chaning Jang, Nina Kirsten, Hayley Kornblum, Melina
Moleskis, Hans Müller, Arnaldo Nascimento, Mariam Okhai, Marliane Owen, Esra Oz, Sonal Pandey,
Yasmin Reid-Linfoot, Baiba Renerte, Mariliis Tael-Öeren, Michael Tracey, Victoria Valle, Žiga Vižintin,
Nuala Walsh, Thomas Wetherup, and Anton Wrisberg for their valuable feedback.

Behavioral Economics Guide 2021 IV

INTRODUCTION
 The Voltage Effect in Behavioral Economics
John A. List
University of Chicago, ANU and NBER

All happy families are alike; each unhappy family first Harvard matriculants are coming soon, Mason
is unhappy in its own way. boasted at the Lion’s Club pancake breakfast. After
-Leo Tolstoy, Anna Karenina. one year, the results arrived. Mason and Greta pored
over the costs, benefits, and outcomes, as measured
For decades, a school district in the upper Mid- by standardized cognitive and behavioral tests.
west of the US has been struggling with kindergar- The results: unequivocally mediocre. “The BE
ten readiness. Administrators have tried a long list
1
program does not even pass a benefit cost test,
of solutions with little success, leaving the District much less yield the silver bullet that was promised
Superintendent, Greta, at her wit’s end. A new by the original study results. I guess the science got
member of the school board, and a devoted follower it wrong this time; those BE results didn’t scale,”
of behavioral economics (BE), Mason, recently read Mason concluded.
an article about an early education behavioral inter-
vention with impressive results: peer-reviewed by But did the science get it wrong?
academic experts, the study showed large treatment
effects on several school readiness indicators. At I believe that the science likely had it right but
the end-of-the-year school board meeting, while that the original results were overinterpreted. The
others discussed the district’s woes, Mason brought program that Greta and Mason tried to replicate
up the idea of implementing a similar BE program, a could never carry the water that they had hoped.
potential silver bullet to address the district’s per- Indeed, most of us think that scalable ideas have
vasive issues with kindergarten readiness. The ben- some ‘silver bullet’ feature, i.e., some quality that
efit-cost ratio was astronomical, he assured Greta bestows a ‘can’t miss’ appeal. That kind of thinking
and the rest of school board. Armed with the science is fundamentally wrong. There is no single quality
and associated statistical jargon that few could un- that distinguishes ideas that have the potential to
derstand, the school board chose to trust Mason and succeed at scale with those that do not do so. In this
adopt the BE program to scale in their district. manner, moving from an initial research study to
That fall, the school district began to introduce the one that will have an attractive benefit cost profile
program, rolling it out in an experimental fashion so at scale is much more complex than most imagine.
that officials could credibly isolate its impacts and And, in most cases, scaling produces a voltage
prove its benefits to the community. At every fish drop—the original BE insights lose considerable
fry and rotary club meeting they attended, Greta voltage when scaled. The problem, ex ante, is de-
and Mason mentioned the BE program, being sure termining whether (and why) that voltage drop will
to mention that economic thinkers as far removed occur. When scaling ideas, one can look to Tolstoy
as Adam Smith, Herbert Simon, Gary Becker, Dan- for a bit of wisdom, because in my travels I have
iel Kahneman, and Dick Thaler produced BE ideas. learned that all successfully scaled ideas are alike; all
Just wait until these students apply to college—our unsuccessfully scaled ideas fail in their own way. What
this lesson inherently means is that scaling, in the
end, is a weakest link problem: the endeavor is only
1 This Greta and Mason opening passage leans heavily on as strong as the weakest link in the chain. However,
Al-Ubaydli et al. (2020b). As noted there, except for the via theory and empirical work, various colleagues
names and a few other changes, this is a true story. and I (see, e.g., Al-Ubaydli & List, 2013; Al-Ubaydli

Behavioral Economics Guide 2021 VI

John A. List The Voltage Effect in Behavioral Economics

et al., 2017a,b; 2020a,b; 2021; Supplee et al., 2021) has as its most susceptible link an understanding of
find that there are five specific traits that scalable the science of how to use science for policy purpos-
ideas must possess—what I call the ‘BIG5’. These es.
are the five ‘key signatures’ of ideas that scale. A As mentioned previously, several colleagues and I
deficiency in any one can render an idea unscalable, have put together a series of studies that both the-
even for the most ingenious among us. oretically and empirically explore these questions.
Our research advocates flipping the traditional
How We Got Here model, calling on scholars to place themselves in
the shoes of the people whom they are trying to in-
Before immersing ourselves in the details of the fluence. Our general call is for policy research that
BIG5, it is useful to step back. To start, it is impor- starts by imagining what a successful intervention
tant to note that BE and field experiments have would look like fully implemented in the field, ap-
contributed immensely to the “credibility revolu- plied to the entire subject population, sustained
tion” of the last three decades in the social sciences over a long period of time, and working as expected,
(see Harrison & List, 2004). In this way, field ex- because its mechanisms are understood. To accom-
periments have become a useful tool for providing plish this goal, our original experimental designs
causal estimates that are difficult to obtain using must provide insights along five key dimensions, to
other approaches. Yet, since the early 1990s, field ensure that we are actually scaling ideas and polices
experiments have focused primarily on testing BE that have a chance to make a deep impact, or at least
theories, uncovering BE mechanisms, and estimat- keep the promise of their initial results.
ing program effects. This represented a logical first
step, as experimentalists sought to provide deeper Introducing the BIG5
empirical insights and theoretical tests as part of
the credibility revolution of the 1990s. These needs can be broken down into what I call
Nevertheless, what has been lacking is a scientific the BIG5. First is the inference problem: how much
understanding of how to make optimal use of the evidence should be gathered before scaling? I advo-
scientific insights generated for policy purposes. cate that a post-study probability of at least 0.95 is
I denote this as the ‘scale-up’ problem, which re- achieved before enacting public policies (see Mani-
volves around several important questions, such adis et al., 2014).2 In practice, this amounts to three
as: do the BE insights we find in the petri dish scale or four well-powered independent replications of
to larger markets and settings? When we scale the the original finding. In the case of Greta and Mason,
BE intervention to broader and larger populations, perhaps the original BE results they read about were
should we expect the same level of efficacy that we simply a false positive. A first truth about false pos-
observed in the small-scale setting? If not, then itives is that they can be considered ‘lies’ or ‘false
what are the important threats to scalability? What alarms’. These are cases whereby, due to statistical
can the researcher do from the beginning of their error, there was never any voltage in the first place.
scholarly pursuit to ensure eventual scalability? At the most basic level, a false positive occurs when
Providing answers to such questions is necessary, you interpret some piece of evidence or data as
because understanding when—and how—our BE proof that something is true when in fact it is not
insights scale to the broader population is critical so. For example, when I visited a high-tech plant in
to ensuring a robust relationship between scientific China that produced headsets, if a headset working
research and policymaking. Without such an under- properly was marked as defective due to human
standing, empirical research can quickly be under- error, that was a false positive. Unfortunately, false
mined in the eyes of the policymaker, the broader
public, and the scientific community itself. Indeed, 2 Formally, post-study probability is defined as the share
in modern economies the chain connecting initial of true associations which are declared true divided by the
research discovery to the ultimate policy enacted share of all associations which are declared true.

Behavioral Economics Guide 2021 VII

John A. List The Voltage Effect in Behavioral Economics

positives are ubiquitous across contexts; in a forth- data, thus far, suggest that representativeness of
coming book titled The Voltage Effect (List, 2022), I the situation is much more important than repre-
summarize findings that suggest a wealth of poli- sentativeness of the population when it comes to
cies and ideas that fail to scale are simply the result generalizing or scaling (see, e.g., List, 2007). For
of false positives. instance, when Greta’s school district scaled up the
The second element of the BIG5 is representa- BE program, they did it within their infrastructure,
tiveness of the population. Often, this is the result which might have been entirely different from that
of failing to know your audience—or assuming of the original study, in that certain logistical con-
that the small subset of people for whom the idea straints were present that affected the roll out. If
worked originally are representative of the general the original results are dependent on the specific
population that needs to be served, so that when you context, or they are not done in a policy relevant
expand your idea it falls short for a broader set of environment, we can expect the benefit-cost profile
people. Following the vignette above, in the original to change at scale.
study, the researcher might have gathered a sample For example, consider Head Start home-visit-
of students that was much different than the stu- ing services, an early childhood intervention that
dents Greta and Mason had in their district. Greta’s found significant improvements in multiple child
school district might have had students with much and parent outcomes in the original research study
different characteristics, including observables (Paulsell et. al, 2010). However, variation in the
such as demographics and educational background quality of home visits was found on a larger scale,
that did not match the original study. In addition, with home visits for ‘at risk’ families involving
the original researcher might have reached a pop- more distractions and less time on child-focused
ulation of students that minimized participation activities, thereby diminishing program effective-
costs, or perhaps a population that had charac- ness and increasing attrition (Al-Ubaydli et al.,
teristics that might yield a larger treatment effect 2020a). In this case, the voltage effect likely oc-
(a ‘let’s give the idea its best shot of working’ re- curred because the scaled program did not include
cruiting strategy). A medical example of this type the fundamental core components that made the
of selection effect can be found in meta-studies of initial intervention promising.
recruitment, which confirm that those who stand The implementation literature sometimes calls
to benefit most from a medical treatment are more this ‘context-dependence’. Likewise, in conjunction
likely to participate in trials (see Al-Ubaydli et al., with curriculum specialists, the original researcher
2020a). Such selection effects might yield a good created a curriculum for a pre-kindergarten pro-
journal publication and future grant funding, but it gram, trained the teachers, and provided hands-on
portends a voltage drop at scale as the program is support throughout the process. When the school
rolled out to everyone. district scaled up the program, they might not have
In a nutshell, researcher choice/bias, selection used the exact same curriculum and care as the
bias/sorting of the study’s population into the pro- original implementation, due to local constraints.
gram, non-random attrition, and (dis)economies This is often described as ‘program drift’ in the lit-
of scale in participation costs all affect the repre- erature. This third reason behind voltage effects is
sentativeness of the population studied, which in generally caused by not understanding that the in-
turn might affect the promise of scaling (see Bell & itial success depended on unscalable ingredients—
Stuart, 2016). unique circumstances that cannot be replicated at
Third is the representativeness of the situation. scale.
A subtle fact is that the research and policy com- A fourth key aspect pertains to spillovers (network
munities oftentimes generalize results to both a effects) and general equilibrium effects of scaling.
population of situations and a population of peo- Concerning the midwestern school district, spill-
ple, even though we often only speak of the latter. overs could be negative from the treated group to
This is particularly troubling considering that the the control group. While the intervention improves

Behavioral Economics Guide 2021 VIII

John A. List The Voltage Effect in Behavioral Economics

the school performance of students in a given class, define ‘sites’ as having multi-dimensional char-
the control group may, upon seeing an initial im- acteristics, which our theory guides into popula-
provement in the performance of the treated group, tion and situational categories. It is, thus, critical
feel demoralized, inducing a deterioration in their for researchers to describe comprehensively the
performance, accentuating the measured treatment environment in which the research is carried out,
effect (psychologists denote this effect as “resent- going beyond a cursory description. In this spirit,
ful demoralization”). Of course, the effect could run I advocate that original researchers should stratify
in the opposite direction. (block) on situations when doing experiments, just
Related to spillovers are what economists call gen- like we commonly stratify on individual character-
eral equilibrium effects, a term describing shifts in an istics in modern experimentation (for example, we
overall market or system that likely do not manifest typically are sure to include both women and men
on a small scale. To illustrate this notion, let’s say in treatment and control groups, and we do so by
that I conducted an experiment wherein I randomly stratification; we should do the same for potential

chose 100 college sophomores, forced half of them non-negotiables in our programs, such as the ac-

to change their major to Economics, then exam- tual human’s delivering, correct dosage, program,
delivery, incentives, substitutes, etc.).
ined how much they were earning in their first job
Finally, we consider marginal cost considerations.
compared to the 50 students who did not change
This fifth element of the BIG5 represents the ‘sup-
majors. I would likely find the Economics majors
ply-side economics’ of scaling—does your idea have
doing quite well. Now, instead, let’s say that I had
economies or diseconomies of scale? Greta needed
50% of all college sophomores around the world
high-quality teachers to run the BE program she
change their major to Economics and the other 50%
was attempting to scale. While the original study
constituted the control group. What would happen a
only needed 10 teachers, Greta needed 100 for her
few years later when they all entered the workforce?
school district. There was just one problem: the best
Assuming no sudden spike in employer demand for
teachers are also very expensive to retain and hire in
Economics majors, a large influx of new economists
the first place. In this case, teachers are very
on the market (increased supply) would cause their
difficult to scale while retaining a reasonable
wages to plummet: a huge voltage drop.
budget. As you ‘buy’” more of them at scale, the
Here is the rub: our BE experiments typically give
price invariably goes up, unlike the wholesale
us answers along the lines of a small-scale exper-
price of lettuce going down for Costco when it
iment; they don’t speak to large movements, such
buys thousands of heads for its locations every
as everyone, or even 50% of college sophomores
week. Indeed, the opposite happens: teachers
changing majors. Yet, in a very real sense, this is become more expensive. This is because to
exactly what we want to know before we scale, es- attract more high-quality people into the
pecially in the policy world: what are the total ef- teaching profession, you must raise the teacher
fects of my idea in a world where everyone changes salary in order to compete with employers that
and anything and everything else can change? Ideas might pay them more, such as a Wall Street bank or
do not exist in petri dishes. And an innovation can a Silicon Valley tech company.
have negative consequences that are at odds with its This key element calls on the analyst to not
purpose but only become visible at scale. only measure benefits and how they might
Representativeness of the population and the sit- scale, but also carefully consider the cost side.
uation as potential threats to scalability underline This is typically not discussed in the literature,
how fundamental it is to understand ‘sites’ (i.e., but an idea that has economies of scale is much
the environment where the original research was more likely to scale effectively than one with
implemented) to address the scale-up problem. severe diseconomies of scale. The cost side of the
The literature treats ‘sites’ loosely whereby some equation just cannot be ignored, and benefit-cost
disciplines focus on the population of sites while profiles should be computed not only in the petri
others emphasize the situational characteristics. I dish but also at scale.

Behavioral Economics Guide 2021 IX

John A. List The Voltage Effect in Behavioral Economics

So Where Does This Leave Us? one or several deficiencies associated with the BIG5.
Find ideas that do scale, and they will each be devoid
After you clear these BIG5 hurdles, you will know of the BIG5. My work showcases that moving from
that you have an idea that scales. More generally, evidence-based policy to policy-based evidence
while our running example pertained to a public forces the researcher to backward induct from what
policy, I do not view these insights as limited to a successful idea or policy looks like at scale and test
helping policymakers. By highlighting the key po- those features in the petri dish.
tential economic sources threatening the scalability
of programs and bringing them to the attention of Nearly every problem has been solved by some-
researchers, I hope that those preparing to conduct one, somewhere. The frustration is that we can’t
new studies might consider modifying their own seem to replicate [those solutions] anywhere else.
designs such that their reported treatment effect –President Bill Clinton.
estimates more accurately inform what is likely to
occur, should the program be scaled. In this way, The Author
as mentioned above, the new demand on scholars
is that we backward induct when setting up our John A. List is Kenneth C. Griffin Distinguished
original research plan, to ensure accurate and swift Service Professor in Economics at the University of
transference of programs to scale with minimal un- Chicago. His research focuses on combining field
certainty. experiments with economic theory to deepen our
Yet, after the BIG5 are cleared, we are not done. understanding of the economic science. In the early
When the program is actually scaled, the correct 1990s, List pioneered field experiments as a meth-
empirical approach should be taken to measure odology for testing behavioral theories and learning
efficacy, and continuous measurement should be a about behavioral principles that are shared across
priority. The first best approach to estimating the different domains. He was elected a Member of the
effects of the program at scale is to do a large-scale American Academy of Arts and Sciences in 2011, and
RCT. One can then compare these estimates with the a Fellow of the Econometric Society in 2015. List re-
results from the original studies, to explore effica- ceived the 2010 Kenneth Galbraith Award, the 2008
cy at scale. If this approach is untenable, then it is Arrow Prize for Senior Economists for his research
critical to adopt an empirical approach that allows in behavioral economics in the field, and was the
stakeholders to measure its efficacy without unre- 2012 Yrjo Jahnsson Lecture Prize recipient. He is a
alistic assumptions. While an exhaustive summary current editor of the Journal of Political Economy.
of such approaches is beyond the scope of our work,
I point the interested reader to List (2007), who dis- References
cusses various empirical approaches to policy eval-
uation as an empirical spectrum, which includes Al-Ubaydli, O., & List, J. A. (2013). On the general-
examples of econometric models that make neces- izability of experimental results in economics:
sary assumptions to identify treatment effects from In G. Frechette & A. Schotter (Eds.), Methods of
naturally occurring data. Some of these approaches, modern experimental economics (pp. 420-462).
such as interrupted time series designs or regres- Oxford University Press.
sion discontinuity analysis, can get pretty close to Al-Ubaydli, O., List, J. A., LoRe, D., & Suskind, D.
addressing the internal validity that RCTs solve. (2017a). Scaling for economists: Lessons from
In closing, scaling of ideas is not a silver bullet the non-adherence problem in the medical lit-
problem. This is because all successfully scaled ide- erature. Journal of Economic Perspectives, 31(4),
as are alike; all unsuccessfully scaled ideas fail in their 125-144.
own way. I have documented five key reasons why Al-Ubaydli, O., List, J. A., & Suskind, D. L. (2017b).
most (if not all) policies and ideas fail to scale. Find What can we learn from experiments?
an idea that failed to scale, and it will revolve around Understanding the threats to the scalability of

Behavioral Economics Guide 2021 X

John A. List The Voltage Effect in Behavioral Economics

experimental results. American Economic Review,

107(5), 282-286.
Al-Ubaydli, O, List, J. A., & Suskind, D. (2020a).
2017 Klein lecture: The science of using sci-
ence: Towards an understanding of the threats
to scaled experiments. International Economic
Review, 61(4), 1387-1409.
Al-Ubaydli, O., Lee, M., List, J., Mackevicius, C., &
Suskind, D. (2020b). How can experiments play
a greater role in public policy? Twelve proposals
from an economic model of scaling. Behavioural
Public Policy, 5(1), 1-48.
Al-Ubaydli, O., Lee, M., List, J., & Suskind, D. (2021).
The science of using science: A new framework
for understanding the threats to scaling evi-
dence-based policies. In J. List, D. Suskind, & L.
Supplee (Eds.), The scale-up effect (pp. 104-125).
Routledge.
Bell, S. H., & Stuart, E. A. (2016). On the “where”
of social experiments: The nature and extent of
the generalizability problem. New Directions for
Evaluation, 2016(152), 47-59.
Harrison, G., & List, J. (2004). Field experiments.
Journal of Economic Literature, 42(4), 1009-1055.
List, J. (2007). Field experiments: A bridge between
lab and naturally occurring data. The B.E. Journal
of Economic Analysis & Policy, 5(2), 1-47.
List, J. (2022). The voltage effect: How to make
good ideas great and great ideas scale. Penguin
Randomhouse.
Maniadis, Z., Tufano, F., & List, J. A. (2014). One
swallow doesn’t make a summer: New evidence
on anchoring effects. American Economic Review,
104(1), 277-290.
Paulsell, D., Porter, T., Kirby, G., Boller, K., Martin,
E. S., Burwick, A., Ross, C., & Begnoche, C. (2010).
Supporting quality in home-based child care ini-
tiative: Design and evaluation options. Technical
Report 3887af819cdc4b2e9f0e830c0fd3f97a,
Mathematica Policy Research.
Supplee, L. H., Ammerman, R. T., Duggan, A., List,
J., & Suskind, D. (2021) The role of open science
practices in scaling evidence-based prevention pro-
grams. Working paper.

Behavioral Economics Guide 2021 XI

 I
EDITORIAL
 Behavioral Economics: Promoting Diversity,
Equity, and Inclusion Through Goal-Setting
Siri Chilazi* and Iris Bohnet
Harvard University

The COVID-19 pandemic and the #MeToo and Black Lives Matter movements have exposed deep structural
inequities in the United States and around the world. Achieving diversity, equity, and inclusion (DEI) has
become a core concern for many organizations. This requires behavior change and trade-offs between the
short-term costs and long-term benefits associated with DEI. In this editorial, we discuss organizational
DEI goals as an evidence-based tool that can address both the will (motivation) and the way (cognition,
skills, and tools) of behavior change. Goals generate the will to change, because they involve accountabil-
ity; induce personal pride and public recognition; convey social norms; induce competitiveness; and work
in tandem with other motivational mechanisms, such as financial incentives and feedback. Goals help us
tap into the necessary ways to change by focusing attention; prompting effort; inducing persistence; and
mobilizing specific, relevant strategies to reach the target. Fundamentally, we argue that organizations
should manage DEI the same way they manage all their other business imperatives: through performance
goals with deadlines and rewards, underpinned by monitoring and reporting. Reaching DEI goals requires
no more and no less than the use of the same planning, feedback, and accountability processes that are
employed to achieve goals in other areas.

Introduction: The Will and the Way of Thus, identifying evidence-based, effective ways
Behavior Change for organizations to advance diversity, equity, and
inclusion (DEI) is more critical than ever.
The enormous changes in the world due to the Making DEI a reality in organizations requires
COVID-19 pandemic, apparent inequities within changing the way we behave. By default, humans
countries (exemplified in the U.S. and many other tend to be more comfortable with others who look
countries by the #MeToo and Black Lives Matter like them and prefer people who conform to their
movements), and other structural changes induced stereotypical beliefs. We expect engineers to be men
by technology and globalization are forcing organi- and nurses to be women,1 and if confronted with a
zations to reimagine how work gets done and, more different picture, we tend to respond with confu-
fundamentally, what the workplace of tomorrow sion or hostility. We experience costs from being
should look like. The so-called “Great Reset” has exposed to difference, and such costs often need to
made it painfully clear that organizations have not
kept up with the dramatic changes in how and where
we work and in the demographic composition of the 1 We acknowledge that the concept of gender is complex
labor force. A larger and more diverse workforce has and does not exist on a binary notion, and that biologi-
helped produce growth and innovation, but it can cal sex as assigned at birth and gender or gender identity
only reach its potential if organizations overcome are distinct. Nonetheless, today’s academic research and
the systemic inequities that hold back women, peo- popular literature on goals and targets still generally ex-
ple of color, and other underrepresented groups. ist on the woman-man gender binary, because this is how
most data are collected. As such, this editorial examines
* Corresponding author: [email protected] gender in the binary context.

Behavioral Economics Guide 2021 2

Siri Chilazi & Iris Bohnet BE: Promoting Diversity, Equity, and Inclusion Through Goal-Setting

be counterbalanced by benefits, in order for us to be lazi & Bohnet, 2020). In 2014, top U.S. tech compa-
willing to change. nies led by Apple, Facebook, Google, and Microsoft
What makes this even trickier is that the costs started releasing annual diversity reports, detailing
tend to be experienced immediately, while the ben- their workforce composition—at a high level, and
efits tend to accrue in the future, once the DEI work using idiosyncratic metrics selected by each com-
has been done: diverse teams do not outperform pany—by gender and race. The reports invariably
homogenous ones immediately but have the po- revealed workforces that were overwhelmingly
tential to do so in the long run (Phillips, 2014). We white, Asian, and male-dominated, especially in
frequently face these types of intertemporal choic- technical jobs. And progress has been scant: the
es—like how much to save for retirement or wheth- share of women in tech jobs grew from 15-20% in
er to exercise today—that require us to weigh costs 2014 to 20-23% in 2019 at Apple, Facebook, Google,
and benefits occurring at different times (Ericson & and Microsoft (Harrison, 2019). While the data are
Laibson, 2019). In such cases, humans consistently useful in illuminating the problem and garnering
fall prey to cognitive biases and make suboptimal attention—after all, what does not get measured
choices (Bohnet, 2016). does not count—the numbers alone do not seem
We conceptualize DEI in organizations as one powerful enough to increase the will to change
such intertemporal choice problem. It has short- DEI-relevant behaviors and identify the way by
term costs, including the discomfort resulting which such change could happen.
from difference and a perception that DEI could be Thankfully, the behavioral scientist’s toolbox
a zero-sum game where those currently in power offers us another powerful mechanism to drive DEI
could only lose by being more inclusive. Besides, behavior change: goals and targets.
organizational practices and procedures need to be
changed, which means additional costs today. But Goals and Targets: What Do We Know?
DEI may also have important long-term benefits
for both individuals and organizations, including What Are Goals?
more creative problem-solving on teams, more
objective decision-making, and greater fairness, A goal is at once an outcome that one aims for—
with the latter increasingly demanded by investors, and one that may not “otherwise happen without
regulators, customers, and employees (Page, 2007). some kind of intervention” (Berkman, 2018)—as
The question, then, is: How can organizations well as a standard for determining performance
make smarter trade-offs between the short and or judging satisfaction (Locke & Latham, 2002). A
long term when it comes to DEI? diversity goal, in particular, is an “organizational
We argue that behavior change requires transfor- objective to increase demographic heterogeneity”
mation along two key dimensions that individuals along a particular dimension of diversity such as
and organizations can influence: the will and the gender or race (Heilman & Welle, 2006). Goal-set-
way. The will reflects our desire, or motivation, to ting, then, is the process by which individuals and
act (Wiltermuth & Gino, 2013). The way reflects our organizations determine the outcomes that they
means to execute: the knowledge, resources, and want to achieve, as well as the associated time-
tools we draw on to act given the constraints—sit- frames (Barends et al., 2016). Indeed, research has
uational, cultural, temporal, financial—we face shown goals to be effective in generating positive
(Berkman, 2018). Therefore, to generate lasting outcomes in contexts as varied as medicine, health
behavior change, it is imperative to address the will and fitness, negotiations, job searches, learning,
and the way by tackling both the motivational and and training, with time spans ranging from one
cognitive challenges associated with new behaviors. minute to 25 years, and subjects ranging from indi-
A case in point is the public reporting of diversity viduals to groups and organizational units (Locke &
data, which has not meaningfully moved the needle Latham, 1990).
on gender diversity in the technology sector (Chi- Typically, two key types of goals are distinguished:

Behavioral Economics Guide 2021 3

Siri Chilazi & Iris Bohnet BE: Promoting Diversity, Equity, and Inclusion Through Goal-Setting

behavior (or process) goals, which are defined in ficial behaviors more rewarding (since people are
terms of specific behaviors such as interviewing at inherently motivated to achieve goals), more sali-
least one person of color for every open position, ent and memorable, and easier by enabling people
and outcome goals, which are defined in terms of to process information more appropriately (Duck-
accomplishments such as hiring 50% women for worth & Milkman, 2018).
all entry-level roles (Epton et al., 2017). Whereas Goals help motivate us toward particular behav-
outcome goals motivate people to put their exist- iors by generating the will to do them. Below, we
ing knowledge, resources, and tools to work on a offer five categories of mechanisms for this effect:
task, process goals motivate them to develop their
task-related abilities by acquiring new knowledge, 1. Goals involve accountability: Accountability, i.e.,
resources, and tools (Barends et al., 2016). Both the implicit or explicit expectation that one might
types of goals lead to behavior change, although be required to justify one’s actions to others, in-
correlational data suggest that outcome goals are creases the cost of failure, and thus motivation,
more closely related to outcomes than behaviors, since no one wants to appear foolish or failing in
and vice versa (Harkin et al., 2016). Relatedly, the eyes of others (Lerner & Tetlock, 1999). One
learning goals focus on the skills or competencies form of accountability is the comply-or-explain
to be developed, while performance goals focus on approach, which acts as a ‘soft default’, in that it
the attainment of a specific level of performance introduces a reference point that people generally
(Locke & Latham, 2002). For instance, a DEI-relat- dislike deviating from, even though they retain
ed performance goal would be to reach a particular the freedom to do so. This approach has been used
employee engagement score on an annual survey, in the UK, Australia, and many other countries to
while a learning goal would be to discover a certain drive behavior change in areas including wom-
number of new strategies to help boost employee en’s representation on corporate boards, compa-
engagement scores. nies’ diversity policies, and corporate governance
(Bohnet, 2016).
How and Why Do Goals Work? 2. Goals can induce personal pride and public rec-
ognition: Goals make behavior change attractive
Goals have the potential to be a powerful tool for by introducing an element of personal pride (in-
DEI-related behavior change, because they address trinsic motivation) and public recognition (ex-
both the will (motivation) and the way (cognition, trinsic motivation) as a result of accomplishing
skills, and tools) of behavior change. Moreover, them. This, in turn, increases commitment to the
goals are an intervention at the level of both the goals, which is essential for goal-setting to work:
individual or organization (the decision-maker) if managers doubt their ability to reach assigned
and the context (the environment), and research on goals, or fundamentally do not accept the premise
long-term behavior change suggests that the most of the goals, they are less likely to remain com-
successful approaches deploy both of these strat- mitted to achieving them (Whelan & Wood, 2012).
egies concurrently (Duckworth & Milkman, 2018). 3. Goals may convey social norms: Goals can com-
For the decision-maker, goals serve to amplify the municate social norms and thereby lead to herd-
value of goal-related behaviors, reduce the value of ing, where people and organizations imitate oth-
goal-unrelated behaviors, or do both at the same ers’ behavior, because what they are doing is seen
time (Berkman, 2018). As for the environment, as ‘the (socially acceptable) thing to do’. DEI goals
goals act as a situational nudge by making bene-
2
can thus shift perceptions of desirable behaviors
and outcomes, which is, in itself, a powerful in-
2 A nudge is “any aspect of the choice architecture that
alters people’s behavior in a predictable way without tervention must be easy and cheap to avoid. Nudges are
forbidding any options or significantly changing their not mandates. Putting fruit at eye level counts as a nudge.
economic incentives. To count as a mere nudge, the in- Banning junk food does not” (Thaler & Sunstein, 2008).

Behavioral Economics Guide 2021 4

Siri Chilazi & Iris Bohnet BE: Promoting Diversity, Equity, and Inclusion Through Goal-Setting

fluence on behavior. attach consequences to people’s noncompliance

4. Goals can induce competitiveness: Humans tend (Rogers et al., 2014). As such, goals are inherently
to be competitive and driven to compare them- a commitment device, making people more like-
selves against others whose characteristics or ly to stick with a task—provided that they come
circumstances are relevant to theirs. Such social with some kind of accountability mechanism (we
comparisons can have long-lasting positive ef- discuss this below). Indeed, targeted commitment
fects on behavior, such as in the context of energy to a specific action rather than a more general
consumption, where alerting consumers to the principle—for example, reusing hotel towels spe-
consumption habits of their neighbors has been cifically as opposed to practicing environmentally
shown to induce competitiveness and lead to a friendly behavior generally—was shown to drive
sustained reduction in their own energy usage meaningful behavior change, even though people
(Allcott & Rogers, 2012; Duckworth & Milkman, were less likely to voluntarily take on such tar-
2018). geted commitments compared to general ones
5. Goals can be coupled with other motivational (Baca-Motes et al., 2013).
mechanisms: While goals are inherently motivat- 4. Goals mobilize relevant strategies (resourceful-
ing, their effects can be magnified by deploying ness): Goals spur us to tap into and/or discover
them in tandem with other motivational mech- goal-relevant knowledge and strategies. When
anisms or behavior change techniques. In fact, faced with a goal, we draw on our past experienc-
interventions that combine goal-setting with es, existing knowledge and expertise, or external
plan-making, feedback, progress monitoring, help in order to reach it. As such, goals mobilize
and commitment devices result in greater behav- relevant strategies like plan-making and progress
ior change than goal-setting alone (Baca-Motes monitoring, which in turn help us to achieve them.
et al., 2013; Harkin et al., 2016).
Goals and targets, which are aspirational and vol-
Goals affect performance by helping us summon untary, are distinct from quotas, which are manda-
the necessary ways to achieve them. Below, we offer tory requirements for the specific representation of
four categories of mechanisms for this effect (Locke certain groups in designated positions. Quotas are
& Latham, 2002): typically imposed and enforced by a legislative or
regulatory body that also has the power to deploy
1. Goals focus attention (direction): Goals focus our sanctions in case of failure or non-compliance
cognitive and behavioral attention on key activ- (Whelan & Wood, 2012). For example, political quo-
ities that will enable us to reach them (goal-rel- tas that randomly assigned a third of village chief
evant activities), and, by extension, away from positions to women in India successfully increased
other activities (goal-irrelevant activities). For the share of women in local government from 5%
example, people who were given feedback on in 1993 to 40% in 2005, and served to improve at-
multiple aspects of their driving improved their titudes towards women as legislators, as well as
performance only on those aspects for which they community outcomes for women and girls (Pande &
had previously set goals. Ford, 2011). While this editorial focuses on DEI goals
2. Goals prompt effort (energy): Goals energize us that are set at the organizational level, rather than
to expend (physical and/or cognitive) effort to quotas, research evidence on the impacts of quotas
reach them by serving as a reminder of what we can offer important insights about the effectiveness
want to achieve. In other words, people are in- of goals, which have thus far been less extensively
herently driven to achieve goals, with high goals studied in the DEI realm.
prompting more effort than low ones. In economics, the discussion of quotas has cen-
3. Goals are a commitment device (persistence): tered on the question of whether they help to induct
Commitment devices have two key features: more qualified people—who otherwise would not
people voluntarily choose to use them, and they have applied or been selected—into target posi-

Behavioral Economics Guide 2021 5

Siri Chilazi & Iris Bohnet BE: Promoting Diversity, Equity, and Inclusion Through Goal-Setting

tions, or whether they “lower the bar” and thereby there is false progress (Temkin & Itembu, 2020).
result in inferior performance. Real-world evidence Monitoring both goal-related and (seemingly)
in favor of the former hypothesis comes from Nor- goal-unrelated data closely will allow organizations
way, where a 2003 law mandating a minimum rep- to spot quickly signs of any of these three unintend-
resentation of 40% of both women and men on the ed consequences.
boards of publicly listed limited liability companies While goals are distinct from quotas, research
accomplished its representational goal while also on the effects of the latter raises important con-
narrowing the gender pay gap among board direc- siderations for organizations contemplating DEI
tors; importantly, the women appointed to boards goals, especially because DEI quotas have thus far
after the introduction of the quota tended to be been studied more extensively than DEI goals. The
observably more qualified than their female prede- evidence shows that quotas influence policy out-
cessors (Bertrand et al., 2017). Similarly, laboratory comes and increase women’s leadership across
experiments have shown that in a setting where a variety of contexts, but they can also engender
high-performing women failed to enter competi- backlash, especially in the corporate sphere where
tions, introducing a gender quota increased their companies have been shown to occasionally try to
willingness to compete (Niederle et al., 2013). strategically circumvent the intended impacts of
In sum, goals work to increase task performance quotas (Pande & Ford, 2011). Gender quotas have
and promote behavior change, because they simul- also been shown to subject women to sabotage (by
taneously address the will and the way of behavior both male and especially female peers) as part of
change, and may improve outcomes if they enable peer review processes in competitive environments
qualified people who were previously discriminated (Leibbrandt et al., 2018). Moreover, their impact is
against to enter the playing field and contribute. context-dependent: quotas increase average effort
Furthermore, compared to quotas, goals have the and wages in discriminatory environments where
advantage of being able to be set and tracked at the women are disadvantaged in the selection pro-
organizational level, without externally-imposed cess and yet equally suited to become managers,
consequences for underperformance. but they do the opposite in non-discriminatory
environments where women are not initially dis-
Unintended Consequences advantaged. Approval for gender quotas is simi-
larly context-dependent and—in accordance with
In her BE Guide 2019 editorial, Nina Mažar dis- meritocratic principles—higher in settings where
cussed the importance of unintended consequences women are disadvantaged in career advancement,
generally, and like any organizational policy, goals notwithstanding possible gender differences in
can have unintended consequences. A goal backfires performance (Ip et al., 2020).
when it has the opposite of the intended effect. For If organizational diversity goals—or quotas—are
example, a multinational company’s diversity goal’s perceived to advantage one group at the expense of
singular aim could be to increase the representation another, they can be rejected on grounds of unfair-
of people from outside its home market—say, Ger- ness, especially in (perceived) zero-sum contexts.
many—in leadership positions, but their number Even if DEI efforts are actually designed to level a
decreases instead. If the presence of the goal causes playing field that was previously unfairly disadvan-
German colleagues to resent non-Germans in the taging, for example women, if men perceive wom-
company—an undesirable effect on a variable, i.e., en’s increased representation as a threat to their
cross-cultural relations that the goal did not set out current standing in the workplace, they are unlikely
to influence—that entails negative spillover. Finally, to buy into goals designed to increase diversity, eq-
if the goal results in apparent change without real uity, and inclusion (Bohnet, 2016). Such zero-sum
underlying progress—such as if it leads to more perceptions can be persistent, even in the face of
non-Germans having leadership-level titles with- evidence to the contrary. For instance, a recent
out the accompanying scope of responsibilities— study showed that in settings with diversity goals in

Behavioral Economics Guide 2021 6

Siri Chilazi & Iris Bohnet BE: Promoting Diversity, Equity, and Inclusion Through Goal-Setting

place, high-potential women who had the ability to of the organization’s current DEI state. Just as a
reach the top levels of organizations were perceived doctor first diagnoses the causes of an illness before
as more valuable than high-potential men; howev- prescribing the appropriate medicine, so too should
er, while these women received pay premiums as a organizations first use workforce analytics to un-
result, they were still out-earned by men overall, derstand their DEI pain points before setting goals
due to the female wage penalty. Even explicit diver- to address them. By necessity, DEI goals will be
sity goals did not “afford widespread advantages to highly contextual. For example, at the BBC, where
undeserving women” (Leslie et al., 2017). journalists seek to portray accurately a world that
Finally, organizations also need to look out for is gender-balanced, it was realistic and purposeful
another potential downside of goals, which is tun- to set a goal of having women and men equally rep-
nel vision. In other words, what does not get meas- resented in all aspects of their journalism through
ured, does not count, and vice versa. For instance, 50:50 The Equality Project (Rattan et al., 2019). Sim-
focusing exclusively on numerical diversity goals ilarly, when the UK set out to increase women’s
(i.e., attaining specific levels of representation) representation on FTSE 100 corporate boards in
can lead individuals and organizations to neglect 2011, it set a target of reaching 25% representation
other important considerations, such as the quality in four years, considering that it was starting from
of the employees and the organizational climate. a baseline of 12.5% (“The Davies Review,” n.d.).
Similarly, focusing only on the composition of the When that initial goal was reached ahead of sched-
hiring pool—like the U.S. National Football League ule, the target was revised to 33% women on FTSE
does with its Rooney Rule, which requires teams 350 boards by the end of 2020—another milestone
to interview but not necessarily select racial mi- that was reached ahead of schedule (“Targets &
nority candidates for head coach positions—can in Progress,” n.d.).
fact lead to a reduction in the probability of hiring Achieving DEI goals will require organizations to
diverse candidates if evaluation processes are not tap into well-established behavioral strategies that
adjusted (Fershtman & Pavan, in press). It is also can help enhance both the will and the way to reach
possible that focusing on DEI efforts targeted at one the goals. Accountability, monitoring and transpar-
specific demographic group, such as women, may ency, and reporting are key strategies to increase
lead organizations to focus less on efforts targeted organizational and individual will to achieve DEI
at other groups, such as people of color. goals, whereas choice architecture is an important
The take-home lesson is that the design and strategy to operationalize the ways to change.
implementation of a system of goals and targets Accountability (will). Research shows that tasking
matters greatly for its eventual success. In organ- specific people (such as a Chief D&I Officer or a de-
izations, some representational goals instituted partmental diversity manager) or entities (such as
by leadership—especially when tied to compensa- a diversity taskforce) with accomplishing diversity
tion—may be perceived more like quotas by those goals makes those goals more likely to be achieved;
affected by them, making this lesson particularly on the flipside, research also shows that goals are
relevant (Whelan & Wood, 2012). How goals are set more likely to be abandoned when no one is in
and framed makes a big difference, and we discuss charge of them (Castilla, 2015). Besides, this type of
this further below. organizational accountability for goal attainment
is a powerful tool to mitigate backlash against DEI
Setting and Achieving DEI Goals goals (Dobbin et al., 2015). Cross-functional diver-
sity taskforces and committees may be particularly
In order to promote diversity, equity, and inclu- beneficial in this regard, because they bring togeth-
sion more effectively, organizations should set DEI er people from different parts of the organization to
goals that are specific, measurable, realistic yet pursue a collective goal (Kalev et al., 2006). Another
stretching, and time-bound. The basis for these promising organizational accountability mecha-
goals should be a thorough, data-driven analysis nism is social networks: since humans loathe neg-

Behavioral Economics Guide 2021 7

Siri Chilazi & Iris Bohnet BE: Promoting Diversity, Equity, and Inclusion Through Goal-Setting

ative social consequences (such as disappointing bin & Kalev, 2016). For example, when a company’s
colleagues or letting down a whole department in a biased pay raises were disclosed internally along
company-wide competition), they can be leveraged with performance ratings, the gender and racial
as a source of interpersonal and public accountabil- gaps in raises essentially disappeared (Castilla,
ity (Rogers et al., 2014). In the case of the UK’s board 2015). Similarly, 50:50 The Equality Project shares all
diversification efforts, company chairmen served participating teams’ monthly data on a transparent
as an accountability mechanism for each other, internal dashboard, which allows for comparison
pushing their peers for progress. They were aided and friendly competition between them (Chilazi et
by external accountability from the media and the al., 2020).
UK government, which were actively monitoring Progress against goals should be monitored as
progress (Beshears et al., 2017). frequently as is feasible. People closer to goal-re-
Importantly, true accountability means that there lated action (such as hiring) will often be able to
are consequences for not meeting the specified monitor progress more or less in real time as hires
goals. These consequences could be monetary (such are made, whereas people further removed from
as missing out on a bonus); reputational (such as those actions (such as HR executives) may monitor
having your performance compared against that progress slightly less frequently. At least one person
of your peers on a public dashboard, scorecard, or in the organization should always have a granular,
report); relational (such as being asked to explain real-time view of how the organization is tracking
your poor performance in front of your managers); against its goals (such as holding a running tally of
or punitive (such as being removed from a lead- hires by gender against a goal of having all hires be
ership position). Depending on the accountability 50-50 women and men). That person can then share
structure, these consequences could fall on an indi- monthly, quarterly, semi-annual, or annual pro-
vidual, on a team, or on the whole organization. In gress updates with others. Empirical data suggest
any case, organizations need to ensure that a focus that tech companies that review their workforce
on the accountability for meeting DEI goals does not diversity data at least once per month have a higher
overshadow the focus on improving diversity, equity, representation of women than companies that re-
and inclusion itself (Temkin & Itembu, 2020). view it less frequently, with effects being particu-
Monitoring and transparency (will). Document- larly pronounced on the entry-level representation
ing progress has been shown to motivate people of women (“Top Companies for Women,” 2019).
toward goal attainment. The focus of progress In any case, progress should be monitored at least
monitoring influences what gets done: monitor- every six months, both in aggregate and by relevant
ing specific behaviors has the greatest impact on subdivisions, such as by level or function (“Setting
those behaviors, whereas monitoring outcomes has and Achieving Diversity Targets,” 2017).
a significant impact on those outcomes (Harkin et Reporting (will). Disclosing DEI data, especially
al., 2016; Kruglanski et al., 2011). Public monitor- alongside related DEI goals, is important to facili-
ing, where information on progress is generated tate learning, as evidenced by the successful exam-
either with or in front of others (such as managers ples of both 50:50 The Equality Project and the UK’s
sharing with each other the number of women and board diversification efforts. Goals set publicly were
men they are promoting on their teams, or making shown in a recent meta-analysis to be particularly
that information available on a dashboard), has effective at driving behavior change (Epton et al.,
a greater positive effect on goal attainment than 2017). Simply put, goals need to be visible to be vi-
private progress monitoring (Harkin et al., 2016). able, and a public (i.e., external to the organization)
Research also shows that transparency works to commitment makes it more likely that they will be
reduce bias in organizational processes such as achieved (Galinsky et al., 2015). While we cannot
promotions and compensation decisions, in part make causal inferences, Fortune 100 companies
because of social accountability or our desire to look that report DEI data and set goals have been shown
good and fair-minded in the eyes of others (Dob- to outperform the full population of large compa-

Behavioral Economics Guide 2021 8

Siri Chilazi & Iris Bohnet BE: Promoting Diversity, Equity, and Inclusion Through Goal-Setting

nies significantly in terms of female and ethnic/ search suggests that it is actually external sources
racial minority representation in leadership (Motel, of feedback and assessment, such as supervisor
2016). Public tracking of progress against DEI goals evaluation, that exert a greater effect on behavior
also sends a signal internally and externally that change (Barends et al., 2016).3
the company is serious about them. Organizations • Support mechanisms. Support mechanisms like
should not only report their demographic data and checklists, reminders, and scorecards are ef-
related goals, but also share learnings about which fective in promoting goal achievement for both
strategies have been successful and unsuccessful individuals and teams, because they mitigate the
in promoting DEI, and what their impact has been effects of forgetfulness and cognitive biases while
on culture and performance aside from mere rep- ensuring that decisions are given appropriate at-
resentational numbers (Whelan & Wood, 2012). tention (Beshears & Gino, 2015).
Choice architecture (way). Redesigning deci- • Defaults. Advancement opportunities, such as
sion-making contexts to mitigate the negative im- stretch assignments and promotions, typically
pacts of biases and inadequate motivation can help require employees to raise their hands actively by
facilitate behavior change and goal attainment. At self-nominating (i.e., by opting in). However, ev-
their core, behavioral design tools guide people idence is emerging that such systems, which de-
toward decisions that are better aligned with their fault to no participation on the part of employees,
interests, without taking away freedom of choice contribute to gender gaps observed in leadership
(Thaler & Sunstein, 2008). Organizations may con- positions, and that opt-out systems reduce these
sider: gaps. A straightforward example involves consid-
ering everyone for promotion after a set time in a
• Plan-making. Making a concrete plan about how, particular role. Given that even high-performing
when, and where one will complete a desired women are empirically less likely to opt in and put
action helps mitigate procrastination and for- themselves forward for advancement opportuni-
getfulness (Soll et al., 2016). Specifically, mak- ties, opt-out advancement mechanisms can help
ing so-called ‘if-then’ plans regarding intended enlist eminently qualified women for higher-lev-
behaviors (e.g., “If I am confronted with an all- el positions and close the gender gap in leadership
white slate of candidates, then I will ask recruiters (Erkal et al., 2019). This could entail, for example,
to send me more options before making a hiring automatically considering everyone at a certain
decision”) notably increases follow-through, on level for promotion instead of relying on self- or
both a one-off and a sustained basis. Individuals, manager nominations, or defaulting all leaders
teams, departments, and whole organizations to receive career advancement support such as
should create specific plans detailing how they coaching and leadership training.
will achieve their DEI goals, and what strategies • Active choice. As an alternative to opt-in or opt-
they will deploy to overcome anticipated setbacks out schemes, active choice mechanisms force de-
(Barends et al., 2016). cision-makers to select one option or another, in-
• Feedback. Feedback provides a valuable yardstick stead of accepting a default. The benefit of active
for evaluating whether one’s current efforts lead choice mechanisms is that they induce reflection
to acceptable, goal-aligned performance. It is and more thoughtful decision-making, which can
impossible for people to adjust their effort and lead to less bias in the decision-making process
behavior without knowing how they are track- (Soll et al., 2016). In the context of DEI, an example
ing against the target (Locke & Latham, 2002). is a promotion process that asks every employee
While self-evaluation was initially theorized to
motivate people to perform better, the latest re- and reviewing goals do not actually boost the effects of
goal-setting, although the authors note that this lack of
3
This is not a unanimous finding in the research litera- effect could be due to the small number of studies re-
ture, as one recent meta-analysis suggests that feedback viewed (Epton et al., 2017).

Behavioral Economics Guide 2021 9

Siri Chilazi & Iris Bohnet BE: Promoting Diversity, Equity, and Inclusion Through Goal-Setting

to indicate whether they want to be considered for pending on the DEI dimension (e.g., gender, race,
promotion in that round. or nationality) and the existing context (e.g., very
• Evaluation criteria. Selecting decision criteria in low vs. relatively high current representation of the
advance—whether for hiring, promotion, com- target group); and will goals set for a particular di-
pensation, or other DEI-related decisions—has mension of diversity (e.g., gender) crowd out efforts
been shown in several studies to lead to more in other areas (e.g., racial diversity)?
objective and unbiased decisions (Norton et al.,
2004; Uhlmann & Cohen, 2015). If assessment The Authors
criteria are not specified ahead of time, people
tend to select candidates who are similar to them, Siri Chilazi is a Research Fellow at the Women and
or to the prototype, regardless of qualifications. Public Policy Program at Harvard Kennedy School.
Evaluation criteria should be reviewed regularly Her life’s work is to advance gender equality in the
to ensure that they do not generate disparate out- workplace through research and research trans-
comes (Galinsky et al., 2015). lation. As an academic researcher, Siri specializes
in identifying, testing, and documenting specific
Conclusion interventions that work to close gender gaps and
de-bias structures and processes in organiza-
Making meaningful progress on creating diverse tions. As an advisor and speaker, Siri brings these
and inclusive organizations will require the use of evidence-based insights to practitioners and fre-
the same tools that have already proven success- quently collaborates with organizations, including
ful in changing human behavior in other contexts. Fortune 500 companies, top professional services
Through goals, behavioral economics provides firms, start-ups, and leading academic institu-
organizations with an important evidence-based tions. She has presented at numerous conferences
way to address both the will and the way of behav- around the world, and her work regularly appears in
ior change, thereby allowing them to promote di- leading media outlets, including Harvard Business
versity, equity, and inclusion more effectively. For Review, The New York Times, BBC, Fast Company,
practitioners, this high-level process is as straight- and Forbes. Siri has an MBA from Harvard Business
forward as setting goals, selecting the metrics by School, a Master in Public Policy from Harvard
which to track them, and holding people account- Kennedy School, and a BA in Chemistry and Physics
able for achieving them (Hirsh & Tomaskovic-De- from Harvard College.
vey, 2020). In many organizations, managers are
already assigned some types of business-related Iris Bohnet, the Albert Pratt Professor of Business
performance targets—sales goals, customer ac- and Government, is the Academic Dean of Harvard
quisition targets, deadlines by which to launch new Kennedy School. She is a behavioral economist,
products, budgets to manage—which affect their combining insights from economics and psychol-
rewards and for which they are personally account- ogy to improve decision-making in organizations
able. DEI goals need not be any different. and society, often with a gender or a cross-cultural
Broader implementation of goals as a tool to pro- perspective. Her most recent research examines
mote DEI holds great promise, especially since pro- behavioral design to de-bias how we live, learn,
gress has traditionally been slow in the DEI field and and work. She is author of the award-winning book
goals have the potential to change this issue. None- What Works: Gender Equality by Design and advises
theless, many open questions remain regarding DEI governments and companies on the topic around
goals, and we encourage additional research in this the world. Professor Bohnet is the co-director of the
area. Future work should look into questions such Women and Public Policy Program and the faculty
as what specific types of goals (process vs. outcome; chair of the executive program “Global Leadership
learning vs. performance) are most effective in the and Public Policy for the 21st Century” for the World
DEI context; how may the effects of goals differ de- Economic Forum’s Young Global Leaders. She

Behavioral Economics Guide 2021 10

Siri Chilazi & Iris Bohnet BE: Promoting Diversity, Equity, and Inclusion Through Goal-Setting

serves on the boards, advisory boards, or as a patron encourage wise choices. Harvard Business Review,
of Credit Suisse Group, Applied, Edge, We Shape 93(5), 52-62.
Tech, Women in Banking and Finance, and the UK Bohnet, I. (2016). What works: Gender equality by
Government’s Equalities Office as well as numerous design. The Belknap Press of Harvard University
academic journals. She was named one of the Most Press.
Influential People in Gender Policy by apolitical in Castilla, E. J. (2015). Accounting for the gap: A
2018 and 2019, a Leading Thinker of Victoria, Aus- firm study manipulating organizational ac-
tralia, 2016-2019, and received an honorary degree countability and transparency in pay decisions.
from the University of Lucerne, Switzerland, in Organization Science, 26(2), 311-333.
2016. She is married and the mother of two children. Chilazi, S., & Bohnet, I. (2020, December 3). How to
best use data to meet your DE&I goals. Harvard
References Business Review. https://hbr.org/2020/12/how-
to-best-use-data-to-meet-your-dei-goals.
Allcott, H., & Rogers, T. (2012). The short-run and Chilazi, S., Rattan, A., & Georgeac, O. (2020). Ros
long-run effects of behavioral interventions: Atkins and the 50:50 Project at the BBC (A) (London
Experimental evidence from energy conservation. Business School Case CS-20-010). London
NBER Working Paper 18492. https://papers.ssrn. Business School.
com/sol3/papers.cfm?abstract_id=2167595. Diversity Best Practices. (2017). Setting and achiev-
AnitaB.org. (2019). Top Companies For Women ing diversity targets. https://www.diversitybest-
Technologists—2019 Key Findings and Insights. practices.com/sites/diversitybestpractices.com/
https://anitab.org/research-and-impact/ files/attachments/2017/09/goal_setting_re-
top-companies/2019-results/. port_.pdf.
Baca-Motes, K., Brown, A., Gneezy, A., Keenan, E. Dobbin, F., & Kalev, A. (2016). Why diversity pro-
A., & Nelson, L. D. (2013). Commitment and be- grams fail and what works better. Harvard
havior change: Evidence from the field. Journal of Business Review, 94(7-8), 52-60.
Consumer Research, 39(5), 1070-1084. Dobbin, F., Schrage, D., & Kalev, A. (2015). Rage
Barends, E., Janssen, B., & Velghe, C. (2016). Rapid against the iron cage: The varied effects of
evidence assessment of the research literature on bureaucratic personnel reforms on diversity.
the effect of goal setting on workplace performance. American Sociological Review, 80(5), 1014-1044.
CIPD. https://www.cipd.co.uk/Images/rapid- Duckworth, A. L., & Milkman, K. L. (2018). Changing
evidence-assessment-of-the-research-litera- behavior for good. Unpublished working paper,
ture-on-the-effect-of-goal-setting-on-work- University of Pennsylvania.
place-performance_tcm18-16903.pdf. Epton, T., Currie, S., & Armitage, C. J. (2017). Unique
Berkman, E. (2018). The neuroscience of goals and effects of setting goals on behavior change:
behavior change. Consulting Psychology Journal: Systematic review and meta-analysis. Journal of
Practice and Research, 70(1), 28-44. Consulting and Clinical Psychology, 85(12), 1182-
Bertrand, M., Black, S., Jensen, S., & Lleras-Muney, 1198.
A. (2017). Breaking the glass ceiling? The effect of Ericson, K. M., & Laibson, D. (2019). Intertemporal
board quotas on female labor market outcomes in choice. In B. Bernheim, S. DellaVigna, & D.
Norway. NBER Working Paper 20256. https:// Laibson (Eds.), Handbook of Behavioral Economics
www.nber.org/papers/w20256. - Foundations and Applications (Vol. 2, pp. 1-67).
Beshears, J., Bohnet, I., & Sanford, J. (2017). Elsevier.
Increasing gender diversity in the boardroom: The Erkal, N., Gangadharan, L., & Xiao, E. (2019).
United Kingdom in 2011 (A) (Harvard Business Leadership selection: Can changing the default
School Case 918-006). Harvard Business School. break the glass ceiling? SSRN. https://papers.ssrn.
Beshears, J., & Gino, F. (2015). Leaders as decision com/sol3/papers.cfm?abstract_id=3022386.
architects: Structure your organization’s work to Fershtman, D., & Pavan, A. (in press). “Soft” affirm-

Behavioral Economics Guide 2021 11

Siri Chilazi & Iris Bohnet BE: Promoting Diversity, Equity, and Inclusion Through Goal-Setting

ative action and minority recruitment. American Leibbrandt, A., Wang, L. C., & Foo, C. (2018).
Economic Review. Gender quotas, competitions, and peer review:
FTSE Women Leaders. (n.d.). Targets & progress. Experimental evidence on the backlash against
https://ftsewomenleaders.com/targets-pro- women. Management Science, 64(8), 3501-3516.
gress/. Lerner, J., & Tetlock, P. (1999). Accounting for the
FTSE Women Leaders. (n.d.). The Davies Review effects of accountability. Psychological Bulletin,
2011-2015. https://ftsewomenleaders.com/2011- 125(2), 255-275.
2015-the-davies-review/. Leslie, L., Manchester, C., & Dahm, P. (2017). Why
Galinsky, A. D., Todd, A. R., Homan, A., Phillips, K. and when does the gender gap reverse? Diversity
W., Apfelbaum, E. P., Sasaki, S. J., Richeson, J. A., goals and the pay premium for high potential
Olayon, J. B., & Maddux, W. W. (2015). Maximizing women. Academy of Management Journal, 60(2),
the gains and minimizing the pains of diversity: 402-432.
A policy perspective. Perspectives on Psychological Locke, E. A., & Latham, G. P. (1990). A theory of goal
Science, 10(6), 742-748. setting and task performance. Prentice Hall.
Harkin, B., Webb, T. L., Chang, B. P., Prestwich, A., Locke, E. A., & Latham, G. P. (2002). Building a
Conner, M., Kellar, I., Benn, Y., & Sheeran, P. practically useful theory of goal setting and
(2016). Does monitoring goal progress promote task motivation: A 35-year odyssey. American
goal attainment? A meta-analysis of the exper- Psychologist, 57(9), 705-717.
imental evidence. Psychological Bulletin, 142(2), Motel, L. (2016). Increasing diversity through
198-229. goal-setting in corporate social responsibil-
Harrison, S. (2019, October 1). Five years of tech di- ity reporting. Equality, Diversity and Inclusion,
versity reports—and little progress. Wired. https:// 35(5/6), 328-349.
www.wired.com/story/five-years-tech-diversi- Niederle, M., Segal, C., & Vesterlund, L. (2013).
ty-reports-little-progress/. How costly is diversity? Affirmative action in
Heilman, M., & Welle, B. (2006). Disadvantaged light of gender differences in competitiveness.
by diversity? The effects of diversity goals on Management Science, 59(1), 1-16.
competence perceptions. Journal of Applied Social Norton, M., Vandello, J., & Darley, J. (2004).
Psychology, 36(5), 1291-1319. Casuistry and social category bias. Journal of
Hirsh, E., & Tomaskovic-Devey, D. (2020). Metrics, Personality and Social Psychology, 87(6), 817-831.
accountability and transparency: A simple recipe Page, S. (2007). The difference: How the power of
to increase diversity and reduce bias [Manuscript diversity creates better groups, firms, schools, and
submitted for publication]. societies. Princeton University Press.
Ip, E., Leibbrandt, A., & Vecci, J. (2020). How do Pande, R., & Ford, D. (2011). Gender quotas and female
gender quotas affect workplace relationships? leadership: A review. Background paper for the
Complementary evidence from a represent- World Development Report on Gender. https://
ative survey and labor market experiments. wappp.hks.harvard.edu/files/wappp/files/gen-
Management Science, 66(2), 80-822. der_quotas_-_april_2011_0.pdf.
Kalev, A., Dobbin, F., & Kelly, E. (2006). Best prac- Phillips. K. W. (2014, October 1). How diversity
tices or best guesses? Assessing the efficacy of makes us smarter. Scientific American. https://
corporate affirmative action and diversity pol- www.scientificamerican.com/article/how-di-
icies. American Sociological Review, 71(4), 589- versity-makes-us-smarter/.
617. Rattan, A., Chilazi, S., Georgeac, O., & Bohnet, I.
Kruglanski, A. W., Pierro, A., & Sheveland, A. (2011). (2019, June 6). Tackling the underrepresentation
How many roads lead to Rome? Equifinality of women in media. Harvard Business Review.
set-size and commitment to goals and means. https://hbr.org/2019/06/tackling-the-under-
European Journal of Social Psychology, 41(3), 344- representation-of-women-in-media.
352. Rogers, T., Milkman, K. L., & Volpp, K.G. (2014).

Behavioral Economics Guide 2021 12

Siri Chilazi & Iris Bohnet BE: Promoting Diversity, Equity, and Inclusion Through Goal-Setting

Commitment devices: Using initiatives to change

behavior. JAMA, 311(20), 2065-2066.
Soll, J. B., Milkman, K. L., & Payne, J. W. (2016). A
user’s guide to debiasing. In K. Gideon, & G. Wu
(Eds.), The Wiley Blackwell Handbook of Judgment
and Decision Making (pp. 924-951). John Wiley &
Sons.
Temkin, R., & Itembu, T. (2020, January 29). The
unintended consequences of diversity & inclu-
sion initiatives: Lisa Leslie’s recent work and
ideas for putting her framework to use. WAPPP
Wire. http://wapppwire.blogspot.com/2020/01/
the-unintended-consequences-of_29.html.
Thaler, R., & Sunstein, C. (2008). Nudge: Improving
decisions about health, wealth, and happiness. Yale
University Press.
Uhlmann, E., & Cohen, G. (2005). Constructed cri-
teria: Redefining merit to justify discrimination.
Psychological Science, 16(6), 474-480.
Whelan, J., & Wood, R. (2012). Targets and quotas
for women in leadership: A global review of poli-
cy, practice, and psychological research. Centre
for Ethical Leadership, Melbourne Business
School. https://www.yumpu.com/en/document/
view/50333823/targets-and-quotas-for-wom-
en-in-leadership-melbourne-business-.
Wiltermuth, S., & Gino, F. (2013). “I’ll have one of
each”: How separating rewards into (meaning-
less) categories increases motivation. Journal of
Personality and Social Psychology, 104(1), 1-13.

Behavioral Economics Guide 2021 13

 II
APPLICATIONS
 Behavioral Insights and Gender-Based Violence
Prevention: Evidence From a Facebook Trial
in Honduras

Paloma Bellatin,* Mónica Wills Silva and Chloe Bustin

The Behavioral Insights Team

Monserrat Bustelo and Nidia Hidalgo

Inter-American Development Bank

Globally, one out of three women experiences physical and/or sexual violence at some point in their lives.
Evidence on what works to prevent gender-based violence (GBV) is still scarce, particularly in the field of
behavioral science. This paper discusses the Behavioral Insights Team’s and Inter-American Development
Bank’s work to prevent GBV, and the results of an online randomized controlled trial (N= 829,445) promot-
ing early help-seeking among survivors in Honduras. The results suggest that Facebook ads that address
sunk cost bias and uncertainty aversion can effectively encourage women to seek help online. This trial
illustrates three core aspects of a behavioral approach to developing GBV prevention policy: (1) drawing
on evidence, (2) developing a behavioral map for barriers to key behaviors among specific actors, and (3)
tailoring interventions grounded in context. We hope to apply this approach to interventions with different
actors, in order to contribute to GBV prevention.

Introduction marily provides insights related to response, rather

than prevention (Ellsberg et al., 2015).
Globally, one out of three women experiences Although a range of institutional, cultural and
physical and/or sexual violence at some point in societal factors influence the prevalence of GBV
their lives, the majority at the hands of an intimate in a given context, behavioral science may prove
partner (World Health Organization, 2013). Evidence a particularly suitable method for addressing this
on what works to address gender-based violence challenge, as it relies on a context- and people-led
(GBV) is still scarce, particularly when it comes to
1
approach to devising solutions. As such, the first
the contributions of behavioral science: most of the step towards addressing GBV through behavioral
available evidence has been derived from studies science is to approach it not as an entrenched, in-
carried out in high-income countries, and it pri- tractable problem but as an aggregate of behaviors
by different actors, each occurring under specific
circumstances.
1
Gender-based violence (GBV) is physical, psychological, The COVID-19 pandemic has posed a particular
or sexual violence perpetrated against an individual or a challenge for GBV prevention. During times of hard-
group on the basis of gender or gender norms. ship and crisis, such as the COVID-19 lockdowns
and associated economic pressures, GBV is likely
* Corresponding author: to increase. In Honduras, for example, in April and
[email protected] May 2020, 18,284 women reported being victims of

Behavioral Economics Guide 2021 15

Paloma Bellatin et al. Behavioral Insights and Gender-Based Violence Prevention

GBV according to data from 911 calls, an increase of Behavioral Science and Gender-Based
9% compared to reports from April and May 2019. 2
Violence: Key Actors Within the Social
Researchers have found similar trends after other Ecology
crises, such as the 2008 financial crisis (Schneider
et al., 2016), Hurricane Harvey (Serrata & Alvarado, GBV is not a single, entrenched problem but an
2019), and Hurricane Katrina (Schumacher, 2010). aggregate of behaviors and factors pertaining to
The national lockdown in Honduras during the different actors within specific contexts (Heise,
COVID-19 pandemic meant that many women 2011; Heise, 2015; Heise & Manji, 2015). By apply-
found themselves quarantined with an aggressor, ing a behavioral lens to GBV, we seek to understand
or that already tense relationships became further which factors (personal, cognitive, environmental,
strained as some families struggled to meet basic institutional) influence decision-making, in order
needs or had to rebalance household roles and re- to design evidence-based, targeted interventions
sponsibilities. that can prevent and decrease GBV. Our work to date
In 2019, the Behavioral Insights Team (BIT) and has mostly focused on: 1) encouraging bystanders
the Inter-American Development Bank (IDB) pro- to intervene in situations of GBV and 2) promoting
duced a report, Applying Behavioral Insights To early help-seeking behaviors among survivors of
Intimate Partner Violence, to provide policymakers
3
violence.
in Latin America with recommendations on how
behavioral science could improve existing services Interventions With Bystanders
for survivors. In 2020, we pivoted our program of
work to encourage early help-seeking through on- Strategies for violence prevention involving by-
line platforms in Central America, which survivors standers assume that members of the community
could safely access during lockdown restrictions. We have the power to sanction or legitimize acts of vi-
conducted four randomized controlled trials (RCTs) olence. A bystander can be anyone—a colleague, a
in Honduras and El Salvador, and we are currently neighbor, an acquaintance, or a passerby. Bystand-
collaborating in two more (as of April 2021). ers must go through a complex decision-making
In this paper, we present a summary of some of the process in order to intervene in a case of GBV (La-
work BIT has done over the past few years to apply tané & Darley, 1968). They must first notice that an
behavioral science to address GBV, drawing insights event is occurring, interpret it as an emergency, as-
from targeted interventions while maintaining a sume a level of responsibility to act, choose a form
systematic approach to violence prevention. of assistance, and, finally, take action.
We focus in particular on an RCT that BIT con- Research suggests that working with the commu-
ducted in Honduras in partnership with IDB to nity and bystanders to violence can be a very effec-
encourage survivors to seek help. We also discuss tive channel for reducing GBV (Fenton, 2016). By-
policy implications and a research agenda for future standers can create new norms for intervention, as
work (Rafael Almeida et al., 2016; Alexander-Scott well as foster a sense of community responsibility
et al., 2016; Garnelo et al., 2019). and competence. For example, in Bangladesh, BIT
collaborated with BRAC University to place posters
in buses that addressed barriers to bystander in-
2 Data provided by the 911 in Honduras. In April and May tervention and provided timely nudges to passen-
2019, there were 16,775 domestic violence reports made gers (BIT-BRAC, 2021). These posters effectively
to the 911 hotline in Honduras. improved attitudes towards victims of sexual har-
3
https://publications.iadb.org/publications/english/ assment. Similarly, BIT worked with universities in
document/Applying_Behavioral_Insights_to_Inti- Australia and Peru to promote bystander interven-
mate_Partner_Violence_Improving_Services_for_ tion and prevent sexual harassment among college
Survivors_in_Latin_America_and_the_Caribbean_ students. By using the bystander intervention model
en.pdf to address the behavioral barriers faced by students,

Behavioral Economics Guide 2021 16

Paloma Bellatin et al. Behavioral Insights and Gender-Based Violence Prevention

we achieved successful results (BIT, 2019). courage survivors to seek help early.
In order to understand better the barriers to seek-
Interventions With Survivors of Violence ing help, and to develop interventions in response
to these barriers, we conducted qualitative research
Our work with survivors has focused on recogniz- in El Salvador with survivors, the Women’s Institute
ing warning signs, promoting early help-seeking, ISDEMU,4 and staff from Ciudad Mujer centers—
and encouraging safety planning. Since 2018, we one-stop centers providing government services
have been developing a program of work in Hondu- for women, including health care, child care, and
ras and El Salvador aimed at developing and testing violence responses.5
interventions to address high levels of GBV and en- Based on our qualitative findings, we conceptu-

Figure 1: A model of the help-seeking process.

alized the help-seeking process and its barriers in edging the situation of violence they experience.
three stages (Liang et al., 2005): In our fieldwork, we found that women associate
In the first stage, women identify their situation the word victim and violence with negative traits,
as violence. Through our fieldwork, we identified so they neither want to be associated with
critical barriers that prevent women from identify- these words nor be seen entering the violence
ing their situation as violence, including: attention service module in Ciudad Mujer.

• Stigma of being perceived as a ‘victim of violence’. • Moral licensing is a phenomenon in which people
This stigma prevents many women from acknowl- make allowances for “bad” behaviors because of
other “good behaviors” (Simbrunner & Schlege-
4
Isdemu web page. lmilch, 2017). In the context of GBV, women may
5
Although the fieldwork was conducted in El Salvador, we make allowances for violent behaviors because
validated the relevance of our qualitative findings with their partner is, for example, a good father, friend,
key stakeholders in Honduras (high-level officials and or financial provider, is well-respected in the
field staff from the INAM and Ciudad Mujer). community, or is a loving partner at other times.

Behavioral Economics Guide 2021 17

Paloma Bellatin et al. Behavioral Insights and Gender-Based Violence Prevention

In the second phase, women decide to seek help, Applying Behavioral Insights to Encourage
as well as what kind of help they will seek, and when Help-Seeking Behaviors in Honduras
and how to take the first step. Critical barriers in-
clude: In Honduras, 27.5% of women aged 15-49 have
experienced at least one act of physical violence
• Lack of self-efficacy and disempowerment to since they were 15 years old. Additionally, 22% of
change their current situation. Toxic or vio- women who have ever been in an intimate partner
lent relationships often erode women’s sense of relationship have experienced some form of vi-
self-efficacy (that is, their belief in their ability olence from a partner during the past 12 months.
to complete tasks, achieve goals, and overcome However, only 20% of these women sought support
obstacles (Bandura, 1977)). Even after identifying from any institutional service (e.g., the police, the
their situation as violence, women may feel like courts, the Institute of Women).6
there is nothing they can do to change their sit- Given that seeking help at an early stage is crucial
uation. to breaking the cycle of violence (Garnelo, 2019),
• Sunk costs. Individuals commit the sunk cost fal- and taking into consideration mobility restrictions
lacy when they continue a behavior or endeavor under COVID-19, BIT and IDB developed an experi-
as a result of previously invested resources (time, ment with the aim of encouraging early help-seek-
money, or effort) (Arkes & Blumer, 1985). Evi- ing behavior through online platforms among GBV
dence from our qualitative analysis shows that survivors in Honduras.
some women stay with their aggressors because
of the previous effort they invested in the rela- The Intervention
tionship.
We collaborated with the National Institute of
Finally, women take the first step towards access- Women and the Ciudad Mujer communications
ing the support services available and disclose the team to design four Facebook ads looking to ad-
fact that they have been subject to violence. Critical dress the barriers identified during our qualitative
barriers include: research and encourage help-seeking. We describe
the four ads in detail, below.
• Uncertainty aversion. We tend to prefer known
risks to unknown risks (Ellsberg, 1961). When A. Combat availability bias
making the mental leap between the intention
to seek help and the action of doing so, survivors Identified barrier: Availability bias is the tendency
face a high level of ambiguity. Women may feel to estimate the frequency or magnitude of events or
worried not only about what will happen to them behaviors based on how easily they come to mind
and their families in the long term, but also about (Tversky & Kahneman, 1974). The media most
the details of the help-seeking process. commonly portrays GBV as physical violence (e.g.,
• Lack of safety plans in case of danger. Safety women covered in bruises). For victims of other
planning is used to help advocates and women forms of abuse, this may be alienating and prevent
talk through strategies that would increase safety them from realizing that they are experiencing
in given scenarios. When women make plans in abuse, or even from seeking help.
case of emergencies, they are much more likely Solution design: Image portraying different forms
to follow through and successfully escape risky of violence that are less commonly depicted (psycho-
situations.
5 The figure was 27% according to the National Demo-
graphic and Health Survey 2011-2012. (The most recent
ENDESA was taken in 2019, yet its results are not publicly
available).

Behavioral Economics Guide 2021 18

Paloma Bellatin et al. Behavioral Insights and Gender-Based Violence Prevention

logical, economic, control)(Figure 2). The first step women in abusive relationships: the more effort
in the help-seeking journey is recognizing that and time a woman has invested in achieving a har-
violence is occurring. Therefore, the image that we monious relationship, the less likely she is to leave.
developed to address this barrier portrayed various Solution design: Image of a woman who has been
forms of violence. successful in seeking help sharing this outcome with
others (Figure 3). To overcome this barrier, we
B. Debunk sunk costs showed women what their lives could be like if they
sought help, allowing them to imagine a better fu-
Identified barrier: The sunk cost fallacy can affect ture in which they have the potential to gain, not

Figure 2: Availability bias.

just lose, from ending an abusive relationship. decide to seek help, and this can stop them from
taking action.
C. Reduce uncertainty aversion Solution design: Image portraying the outcome
of help-seeking (a woman on the phone receiving
Identified barrier: Uncertainty aversion (Ellsberg, assistance) (Figure 4). The image aimed to clarify
1961; Berger et al., 2013). Women in situations of vi- and demystify the help-seeking process. It clearly
olence are often uncertain what will happen if they portrayed who was on the other end of the line (not

Behavioral Economics Guide 2021 19

Paloma Bellatin et al. Behavioral Insights and Gender-Based Violence Prevention

Figure 3: Sunk costs and counterfactual thinking.

Figure 4: Uncertainty aversion.

Behavioral Economics Guide 2021 20

Paloma Bellatin et al. Behavioral Insights and Gender-Based Violence Prevention

Figure 5: Action-intention gap and safety planning.

a male police officer but a trained operator) and unhealthy relationship at the last moment) (Goll-
explicitly assured anonymity and confidentiality witzer, 1999).
for the caller. The image aimed to convey that the Solution design: Image of safety plan milestones
caller had agency over what happened after their (Figure 5). Behavioral science literature has shown
help-seeking call. that people are more likely to follow through on
their intentions if they plan their actions in ad-
D. Addressing the action-intention gap portraying vance (Belanger-Gravel et al., 2011). The image
a safety plan that we developed to address the intention-action
gap aimed to help women make these plans by pro-
Identified barrier: The intention-action gap refers viding a checklist of the key decisions (safety plan)
to the idea that people do not always do the things they must make when planning to leave a violent
that they intend to do, because they either fail to relationship.
get started or get derailed along the way. People
sometimes fail to follow through on their inten- The Trial
tions, because they procrastinate (e.g., putting off
calling a help center), they miss the opportunity to We ran a five-arm RCT on Facebook, in which we
act (e.g., failing to call the helpline when they have randomly assigned individual women Facebook us-
a moment alone), or they have second thoughts at ers7 to see one of the four ads presented.
a critical moment (e.g., deciding against leaving an A fifth control group was assigned to see a slightly

Behavioral Economics Guide 2021 21

Paloma Bellatin et al. Behavioral Insights and Gender-Based Violence Prevention

Figure 6: Control image.

modified version of the business-as-usual gov- provide full information to survivors in the control
ernment image, which provided information on group. We automated the randomization process,
support services for survivors. To differentiate the using the A/B testing function in Facebook’s Exper-
control image from the behavioral strategies that iments platform.
we include, we intentionally avoided adding narra- The trial ran from August 22 to September 19,
tive graphics. We included a call to action (call 911) 2020. Our sample consisted of 829,445 women—
and added an explanatory message below, to around ~165,000 in each trial arm.8 We measured

7
Our sample included users who: (1) lived in Honduras, (2) On Facebook, this metric is called ‘Reach’ and is an esti-
8

self-identified as a woman, (3) were 18-65 years old, and mate based on the number of people who were exposed to
(4) spoke Spanish as their primary language. one of our ads at least once.

Behavioral Economics Guide 2021 22

Paloma Bellatin et al. Behavioral Insights and Gender-Based Violence Prevention

Figure 7: Treatment effects on entering the Ciudad Mujer website.

effects on online help-seeking behavior, which we on the Facebook ads that directed them to the Ciu-
defined as a woman interacting with the Ciudad dad Mujer website and loading the page.9
Mujer webpage. Specifically, we recorded (1) if a Women who were exposed to the image address-
woman clicked on the Facebook Ad, taking her to ing uncertainty aversion were 19.4% more likely
the Ciudad Mujer website, and (2) if after navigating to visit the Ciudad Mujer page than women in the
to the Ciudad Mujer webpage, a woman then clicked control group, while those exposed to the image ad-
on any of the contact channels for Ciudad Mujer: dressing sunk costs were 12.9% more likely to visit
WhatsApp, email, Facebook, LinkedIn, or Twitter. the Ciudad Mujer page than others in the control
group (both results statistically significant at the
Results 1% level). Furthermore, the images addressing un-
certainty aversion and sunk costs also had signifi-
We used a logistic regression to test the effects of cantly different effects from the rest of the images,
the different messages. suggesting that these results are not simply effects
from a more attractive design but that these behav-
Primary results: Treatment effect on contacting ioral strategies were indeed more effective than the
Ciudad Mujer
As we tested a large number of hypotheses in our prima-
9

We found positive and significant results from ry analysis, we conducted multiple comparison analysis
two of our trial arms on help-seeking behavior, as (MCA), using the Hochberg Step-Up Procedure on these
measured by the likelihood of users clicking a link results.

Behavioral Economics Guide 2021 23

Paloma Bellatin et al. Behavioral Insights and Gender-Based Violence Prevention

Figure 8: Treatment effects on contacting Ciudad Mujer.

rest. We present these results in Figure 7. all our treatment arms on the likelihood a viewer
The regression results can be found in the Annex. would make contact with Ciudad Mujer.
As a robustness check, we have added the ordinary Women who were exposed to the image address-
least squares (OLS) results. ing sunk costs were 80% more likely to contact Ci-
udad Mujer than women in the control group, while
Secondary results: Treatment effect on contacting women exposed to the images addressing availabil-
Ciudad Mujer ity bias, uncertainty aversion, or the intention-ac-
tion gap were, respectively, 62%, 37%, and 30%
In this section, we describe women’s behavior af- more likely to contact Ciudad Mujer (all statistically
ter they clicked on the Facebook ad and arrived at significant at the 1% level). Taken together with
the Ciudad Mujer website. As we have shown that the results above on the likelihood of clicking, this
the ads had differential effects on the likelihood of suggests that the sunk cost message was overall the
clicking, this sample is therefore no longer random, most effective. Given the smaller sample size in this
and differences reported here cannot be interpreted analysis, we did not find a significant difference be-
as causal. This, however, gives a more direct inter- tween treatment arms. We present these results in
pretation of the effects of the ad on the decision to Figure 8, while logistic and OLS regression results
seek help (women then clicking on any of the follow- can be found in the Annex.
ing channels to contact Ciudad Mujer: WhatsApp,
email, Facebook, LinkedIn, or Twitter)—the second Discussion
step in our model of help-seeking.
We found positive and significant effects from Given that we conducted a quantitative online

Behavioral Economics Guide 2021 24

Paloma Bellatin et al. Behavioral Insights and Gender-Based Violence Prevention

trial with limited covariates, our ability to explore dynamics among specific actors, and (3) tailoring
the mechanisms underlying our results is somewhat interventions grounded in context to address these
constrained. Nonetheless, below we share some hy- barriers.
potheses based on our previous qualitative findings. Drawing on these insights, and motivated by the
First, both images that had a significant effect promise of this approach, we look forward to con-
addressing uncertainty aversion and sunk costs tinuing our work in this space, and we would love to
displayed women in positive scenarios. The un- hear from actors interested in further innovations
certainty aversion image showed women receiving in GBV prevention.
help, and the sunk costs image showed women in a
positive situation after leaving an unhealthy rela- The Authors
tionship. On the other hand, the images addressing
availability bias and the intention-action gap pre- Paloma Bellatin is an advisor at BIT North Amer-
sented negative scenarios, namely, women either ica, where she researches and develops behavioral
in shock looking at different types of violence, or in science-informed policy and program innovations
distress thinking about safety. for government partners. She designs, implements,
Second, uncertainty aversion and sunk costs were and analyzes rigorous evaluations to determine the
two of the strongest barriers to help-seeking that impact of policies, with a regional focus on Latin
we identified in our qualitative research. The posi- America and the Caribbean and a thematic focus on
tive results from our trial suggest that tackling these gender-based violence and social protection policy.
barriers (i.e., reducing uncertainty about what hap- She holds an MPA in Economics and Public Policy
pens after you contact a support service, and help- from the Princeton School of Public and Interna-
ing survivors understand how remaining with an tional Affairs at Princeton University, and a BA in
abusive partner affects their well-being) can be an Political Science from the Pontificia Universidad
effective way of encouraging help-seeking. The im- Catolica del Peru.
age addressing sunk costs, which invites thoughts
of a positive future and the end of a negative sit- Monica Wills is a principal advisor and the co-
uation, was also cumulatively the strongest of all Head of International Development at the Behav-
images, leading to the largest impact on women’s ioral Insights Team. Monica has worked for over 8
likelihood to contact Ciudad Mujer. This suggests years in public policy design and evaluation in low-
that taking the time to understand key behavioral and middle-income countries. Her work has mostly
barriers and develop an evidence-based theory of focused on applying behavioral science to complex
change is critical to addressing GBV effectively. policy challenges including gender-based violence,
anti-corruption, and peacebuilding. Monica holds a
Conclusions Masters of Public Policy from the Hertie School of
Governance in Berlin, and a BSc in Economics and a
In the trial presented in this paper, we mapped BA in Political Science from Los Andes University in
barriers to a desired behavior (online help-seeking) her native Colombia.
for a specific actor (GBV survivors in Honduras), in
order to devise and test targeted interventions to Chloe Bustin is a senior advisor working across
address those barriers. The lessons learned from home affairs and international development. Her
this evaluation have now informed the Honduran work primarily focuses on the application of behav-
Government’s widespread communication cam- ioral science to domestic abuse and gender equality
paigns to increase GBV reporting. in the UK, Latin America and Eastern Europe. She
This trial illustrates three core aspects of a behav- has worked on a range of policy areas in the UK and
ioral approach to developing GBV prevention policy: overseas, including poverty reduction, birth reg-
(1) drawing on evidence, (2) developing a behavio- istration, employment, corruption, and tax. Chloe
ral map that describes barriers to key behaviors and holds an MSc in Language Sciences (with a special-

Behavioral Economics Guide 2021 25

Paloma Bellatin et al. Behavioral Insights and Gender-Based Violence Prevention

ization in neuroscience and communication) from ioral Insights Team for their comments and revi-
UCL, and a BA in Modern Languages (Russian and sions: Ruth Persian, Marta Garnelo, Evie Monning-
Hispanic Studies) from Durham University. ton-Taylor, Laura Litvine, and Bridie Murphy.
All opinions in this paper are those of the authors
Monserrat Bustelo is an economist in the Gender and do not necessarily represent the views of the
and Diversity Division of the Social Sector of the In- Behavioral Insights Team, Ciudad Mujer, the gov-
ter-American Development Bank (IDB). Monserrat ernment of Honduras, or the Interamerican Devel-
has more than 15 years of experience leading the opment Bank, its executive directors, and the gov-
development, implementation, and dissemination ernments they represent.
of cutting-edge research on gender issues related to
the female labor market and violence against wom- References
en, as well as providing technical and analytical
support to operations and managing project teams Alexander-Scott, M., Bell, E., & Holden, J. (2016)
and resources effectively. Monserrat has a Ph.D. in DFID guidance note: Shifting social norms to tackle
Applied Economics from the University of Illinois Violence Against Women and Girls (VAWG). VAWG
at Urbana-Champaign, and a master’s degree in Helpdesk.
Economics from the National University of La Plata, Arkes, H. R., & Blumer, C. (1985), The psychology of
Argentina. sunk costs. Organizational Behavior and Human
Decision Processes, 35, 124-140.
Nidia Hidalgo is a lead specialist for the Gender Bandura, A. (1977). Self-efficacy: Toward a unify-
and Diversity Division at the Interamerican Devel- ing theory of behavioral change. Psychological
opment Bank based in El Salvador. She has a mas- Review, 84, 191-215.
ter’s degree in Rural Development Sciences with a Belanger-Gravel, A., Godin, G., & Amireault, S.
specialization on gender, and a Doctoral Degree in (2011). A meta-analytic review of the effect of
Agro-industrial Economic Problems with a focus implementation intentions on physical activity.
on rural financial markets and gender from the Au- Health Psychology Review, 7(1), 23-54.
tonomous University of Chapingo (Mexico). She has Berger, L., Bleichrodt, H., & Eeckhoudt, L. (2013).
worked as a gender consultant for various interna- Treatment decisions under ambiguity. Journal of
tional organizations, has been a speaker at seminars Health Economics, 32, 559-569.
and conferences, and is the author or co-author of BIT. (2019). Take action: Empowering bystanders to
numerous books and articles. act on sexist and sexually harassing behaviours in
universities. https://apo.org.au/sites/default/
Acknowledgements files/resource-files/2019-09/apo-nid272741.
pdf.
The authors would like to thank the Institu- BIT-BRAC. (2021). Nudging bystanders to com-
to Nacional de la Mujer (INAM) and Ciudad Mujer bat sexual harassment in Bangladesh. https://
teams—Ana Aminta Madrid, Rosa de Lourdes Paz, www.bi.team/publications/nudging-bystand-
Karla Alfaro, Patricia Toledo, Lineth Peralta, and ers-to-combat-sexual-harassment-in-bang-
Wendy Fuentes—for their work and collaboration ladesh/.
in the design of the intervention evaluated in this Ellsberg, D. (1961). Risk, ambiguity, and the savage
trial. axioms. The Quarterly Journal of Economics, 75(4),
Likewise, we want to thank other members of the 643-669.
Inter-American Development Bank team—Paola Ellsberg, M., Arango, D. J., Morton, M., Gennari, F.,
San Martin, Nidia Barahona, and Gladis Morena Kiplesund, S., Contreras, M., & Watts, C. (2015).
Gómez—for their support in the completion of this Prevention of violence against women and
project. girls: what does the evidence say? The Lancet,
Finally, we want to thank members of the Behav- 385(9977), 1555-1566.

Behavioral Economics Guide 2021 26

Paloma Bellatin et al. Behavioral Insights and Gender-Based Violence Prevention

Fenton, R. (2016). A review of evidence for bystander partner violence and Hurricane Katrina:
intervention to prevent sexual and domestic violence Predictors and associated mental health out-
in universities. Working paper. https://www2. comes. Violence and Victims, 25(5), 588-603.
uwe.ac.uk/faculties/BBS/BUS/law/Law%20 Serrata, J. V., & Hurtado Alvarado, M. G. (2019).
docs/dvlitreviewproof0.6.forCLR.pdf. Understanding the impact of Hurricane Harvey on
Garnelo, M., Bustin, C., Duryea, S., Morrison, A. family violence survivors in Texas and those who
(2019). Applying behavioral insights to intimate serve them. Texas Council on Family Violence.
partner violence: Improving services for survivors in https://texashousers.org/wp-content/up-
Latin America and the Caribbean. BIT-IDB. https:// loads/2019/08/Understanding-the-Impact-
www.bi.team/wp-content/uploads/2019/11/ Hurricane-Harvey.pdf.
Applying-Behavioral-Insights-to-Intimate- Schneider, D., Harknett, K., & McLanahan, S. (2016).
Partner-Violence-ENG-3.pdf. Intimate partner violence in the Great Recession.
Gollwitzer, P. M. (1999). Implementation inten- Demography, 53(2), 471-505.
tions: Strong effects of simple plans. American Simbrunner, P., & Schlegelmilch, B. B. (2017). Moral
Psychologist, 54, 493-503 licensing: A culture-moderated meta-analysis.
Heise, L. (2011) What works to prevent partner vio- Management Review Quarterly, 67, 201-225.
lence? An evidence overview. DFID. https://www. Tversky, A., & Kahneman, D. (1974). Judgment
gov.uk/research-for-development-outputs/ under uncertainty: Heuristics and biases. Science,
what-works-to-prevent-partner-violence-an- 185(4157), 1124-1131.
evidence-overview. World Health Organization. (2013). Global and
Heise, L. (2015, May 14) Social norms: Introduction regional estimates of violence against women:
to basic concepts. Webinar, London School of Prevalence and health effects of intimate partner
Hygiene and Tropical Medicine. violence and non-partner sexual violence. World
Heise, L. & Manji, K. (2015) Introduction to social Health Organization. https://apps.who.int/iris/
norms: Briefing note for DFID. DFID. handle/10665/85239.
Latané, B., & Darley, J. M. (1968). Group inhibition of
bystander intervention in emergencies. Journal
of Personality and Social Psychology, 10, 215-221.
Liang, B., Goodman, L., Tummala-Narra, P., &
Weintraub, S. (2005). A theoretical framework
for understanding help/seeking processes
among survivors of intimate partner violence.
American Journal of Community Psychology, 36(1-
2), 71-84.
Parker, E. M., & Gielen, A. C. (2014). Intimate partner
violence and safety strategy use: Frequency of
use and perceived effectiveness. Women’s Health
Issues, 24(6), 584-593.
Rafael Almeida, S., Sousa Lourenço, J., Dessart, F.
J., & Ciriolo, E. (2016). Insights from behavioural
sciences to prevent and combat violence against
women: Literature review. Publications Office of
the European Union. https://publications.jrc.
ec.europa.eu/repository/bitstream/JRC103975/
lbna28235enn.pdf.
Schumacher, J. A., Coffey, S. F., Norris, F. H., Tracy,
M., Clements, K., & Galea, S. (2010). Intimate

Behavioral Economics Guide 2021 27

Paloma Bellatin et al. Behavioral Insights and Gender-Based Violence Prevention

Annex Where:
• Yi: is the primary and secondary outcome –
Analytical strategy: Primary and secondary whether instance i9 of the ads results in the user
outcomes Viewing Content in the Ciudad Mujer website; =1
if instance i of the ads results in the user clicking
Given that our outcomes are binary, we used a one of the Contact options on the Ciudad Mujer
logistic regression to test the effect of our treatment web page;
on the primary outcome measure.

• Ad A is a dummy variable = 1 if individual i was And subsequently converted from log odds to
allocated to ad A; probabilities. The inverse of the logit function is:
• Ad B is a dummy variable = 1 if individual i was
allocated to ad B;
• Ad C is a dummy variable = 1 if individual i was
allocated to ad C;
• Ad D is a dummy variable = 1 if individual i was
allocated to ad D. This tells us the predicted relative probability that
an individual in a given treatment group will be
The coefficients of interest are ß1, which measures Viewing Content, with respect to the control group.
the impact of being allocated to ad A compared to
Control, ß2, which measures the impact of being al-
located to ad B compared to Control, etc. Since these
are logistic measures, the coefficients are log-odds
ratios, which are not straightforward to interpret.

To interpret the effect, we used the log-odds ratio:

9
This will be captured by reach. Reach: “The number of
people who saw your ads at least once. Reach is different
from impressions, which may include multiple views of
your ads by the same people.”

Behavioral Economics Guide 2021 28

Paloma Bellatin et al. Behavioral Insights and Gender-Based Violence Prevention

Primary results: Logistic and linear Secondary results: Logistic and linear
regressions10 regressions11

10
Standard errors are in parentheses. We used the Hoch- 11
Standard errors are in parentheses. We used the Hoch-
berg Step-Up procedure for multiple comparison analysis berg Step-Up procedure for MCA.

Behavioral Economics Guide 2021 29

 Sticky or Not? How COVID Has Changed
Consumer Behaviour in Financial Services, and
What Might Happen Next

Paula Papp,* Kalina Kasprzyk and Harry Davies

Frontier Economics

The COVID-19 pandemic has changed how we do many things, including how we interact with financial
services. In this article, we look at three areas of financial services that have been particularly affected:
household savings, use of cash and online banking adoption. We use data to understand what has changed
and what has likely driven the changes. We then draw on insights from the behavioural economics liter-
ature to draw conclusions about the potential “stickiness” of these changes. We apply concepts such as
habit, social norms, convenience and inertia. The conclusion? It’s a mixed picture. High levels of savings
are unlikely to endure once restrictions are lifted, but the decline in cash usage may well be here to stay. And
while the accelerated shift to digital banking will largely persist, the pandemic has reaffirmed the ongoing
importance of some human interaction in banking services.

Introduction convenience and behavioural inertia.

The conclusion? It’s a mixed picture. High levels
The COVID-19 pandemic has caused great up- of savings are unlikely to endure once restrictions
heaval in the financial services industry. It has are lifted, but the decline in cash usage may well be
accelerated the trend towards cashless payments here to stay. And while the accelerated shift to dig-
and online banking, and seen savings jump to un- ital banking will largely persist, the pandemic has
precedented levels. But are these trends temporary? reaffirmed the ongoing importance of some human
Or are they “sticky” changes, set to persist over interaction in banking services.
the long term? These are questions being asked by Our analysis follows below, beginning with the
businesses and policymakers alike. rise in average household savings.
Behavioural economics can help us provide in-
formed answers. By looking at the data we have, The Great Savings Glut
we can understand how consumer behaviour has
changed, and what has driven these shifts. And we How Has Savings Behaviour Changed?
can use insights from behavioural economics to
look at what might happen next. Prior to the pandemic, household savings rates
We applied this approach to three areas of finan- were stable: between 2018 and 2020, the average
cial services behaviour across Europe – household quarterly rate for EU countries was around 12% of
savings, cash usage and online banking adoption – disposable income, with only minor fluctuations.
considering some of the most powerful behavioural In stark contrast, savings rates skyrocketed in
drivers of economic decision-making, namely hab- 2020, up to around 25%. And while that figure
it, social norms, perceptions, buying impulsiveness, dropped in Q3, as lockdown restrictions were eased,
overall savings balances remained at a high level.
* Corresponding author: Figure 1 shows just how widespread these trends
[email protected] were across Europe.

Behavioral Economics Guide 2021 30

Paula Papp et al. How COVID Has Changed Consumer Behaviour in Financial Services

Aggregate household savings data disguise a more and middle-income earners) and retirees were
nuanced picture across consumer groups. People in most likely to report an increase. On the other hand,
more stable income groups were particularly likely those who were self-employed, unemployed or fur-
to report a rise in their savings. In the UK, as shown loughed were more likely to have seen their savings
in Figure 2, full-time employees (particularly high- fall.

Figure 1: Household saving rates during the pandemic. Notes: Savings rates reported as a percentage of household
disposable income. Source: Eurostat, ECB euro area statistics, Bank of England.

Why Did Households Save More? for a further 1%.2

The overall rise in savings rates can be put down Will Savings Rates Remain High? Applying
to two main factors: Behavioural Economics

• “Involuntary” savings, accumulated “by default” To evaluate the “stickiness” of increased savings
as a result of fewer spending opportunities. rates, two concepts in the literature are particularly
• “Precautionary” savings, actively put aside as in- useful: habit and buying impulsiveness.
surance against greater economic uncertainty. Habit is a well-documented driver of higher sav-
ings balances (Alessie & Teppa, 2009; Loibl et al.,
The European Central Bank (2020) evaluated the 2011). Savings decisions are often thought of as de-
relative importance of these two factors in deter- liberate – choosing to save for a car, for example.
mining the household savings rate during the first But there are numerous smaller-scale decisions,
half of 2020, in the five largest euro area countries.1 made on a daily basis, which are subject to less re-
They found involuntary saving was by far the main flection, like switching off the lights when leaving a
driver: an additional 11% of disposable income was room. Here, habit is more important.
being saved involuntarily in Q2 2020 relative to Q4 Buying impulsiveness refers to a sudden desire
2019, while precautionary saving only accounted
2 The ECB used a fixed effects panel model, using house-
hold expectations about future unemployment to esti-
1 The five largest euro area countries by GDP are Germany, mate the impact of precautionary savings and assigning
France, Italy, Spain and the Netherlands. involuntary savings to the residual.

Behavioral Economics Guide 2021 31

Paula Papp et al. How COVID Has Changed Consumer Behaviour in Financial Services

to spend money, without prior intention to do so higher savings rates are here to stay.
(Beatty & Ferrell, 1998). Greater self-discipline can
help temper it, and consequently increase savings Involuntary saving
(Achtziger et al., 2015).
Looking at both involuntary and precautionary Forming a savings habit – i.e. one that will last
saving during the pandemic through the lens of in the long term – requires repeated cue-based re-
these concepts can help us understand whether sponses (Lally et al., 2009). But increased involun-

Figure 2: Household savings balances in 2020 relative to 2019, by income group. Notes: Totals for employment
groups do not sum to 100%, as the proportion of households with no change in savings balances is not displayed.
Source: Frontier analysis of Bank of England NMG Household Survey data for 2020.

tary saving during the pandemic has been a default es in consumption patterns – either because some
outcome of the restriction of spending opportuni- of the spending categories may never recover to
ties, rather than a more conscious response on the pre-pandemic levels (e.g. commuting) or consum-
part of savers. By the same token, involuntary sav- ers discover some easy “savings shortcuts” (e.g.
ing does not involve any increase in self-discipline adjusting to eating out less often). It remains to be
or decrease in buying impulsiveness. seen whether consumers will seize these opportu-
In other words, the behavioural changes that have nities to save more or shift consumption elsewhere.
led to increased levels of involuntary saving are ex- Second, some consumers might carry on with the
pected to be temporary and unlikely to have become defaults they set up to manage their savings (e.g. a
ingrained, which means savings rates will be more regular standing order or joining a rounding app).
likely to fall away once lockdown restrictions are
relaxed. Evidence of this behaviour has already been Precautionary saving
demonstrated by the fall in savings rates during Q3
of 2020, as restrictions were eased during the sum- The rise in precautionary saving is consistent with
mer months across Europe. traditional economic models, which predict that
But there remains some scope for the forced people save more in times of economic upheaval, in
savings generated during the pandemic to bring order to insure against future income uncertainty
about some enduring increase in the savings rate. (Mody et al., 2012).
First, this can happen through permanent chang- We expect precautionary balances will eventually

Behavioral Economics Guide 2021 32

Paula Papp et al. How COVID Has Changed Consumer Behaviour in Financial Services

fall, led by an increase in the consumption of “big rope. Moreover, it is yet to be seen what customers
ticket” items, postponed during the pandemic. The will choose to do with their accumulated savings
speed of this change will be determined, at least balances over time.
in part, by the strength of the narratives around
economic recovery and social proof (Shiller, 2017). Cash or Card?
Shiller gives the example of the narrative during the
2007 financial crisis which drew parallels with the How Has Cash Usage Changed?
Great Depression of the 1930s. He suggests that the
narrative at that time may have invoked expectation COVID-19 has caused a seismic shift towards buy-
of a severe economic shock, which then translated ing online, and with it a decline in cash spending. It
into a more significant downturn. The big question has also prompted consumers to revisit their pay-
around precautionary balances is how long will the ment preferences for in-person purchases.
feeling of uncertainty last in the case of COVID-19? The ECB’s (2020) survey of the euro area (illus-
And the strength of the narrative around the eco- trated in Figure 3) reveals steep, self-reported de-
nomic rebound will be a key factor in how long that clines in cash usage, with 40% of respondents using
feeling lasts. cash less often.3 Inversely, 40% of respondents said
they were using contactless ‘somewhat more often’
What next? or ‘much more often’. Banks we work with have also
reported a significant increase in the use of contact-
Applying behavioural concepts indicates that less payment, facilitated in part by the expansion of
there has been little change in the underlying con- the contactless limit and mobile wallets.
ditions driving savings behaviour. Instead, the sav- These findings indicate that the pandemic has
ings boom has been largely involuntary. For banks catalysed the pre-existing trends of declining cash
and retailers, therefore, the question is likely to be usage and increasing cashless payment.
when, not if, savings rates drop back to “normal” The disconnect between self-reported behav-
– especially given the low interest rates across Eu- iour and observed behaviour is well documented

Figure 3: Payment methods during the pandemic. Notes: Online survey results are for the euro area (all 19 countries).
Source: ECB IMPACT Survey (2020).

Behavioral Economics Guide 2021 33

Paula Papp et al. How COVID Has Changed Consumer Behaviour in Financial Services

(Bertrand & Mullainathan, 2001; Jahedi & Méndez, consumers towards online shopping. UK data from
2014). But in this case, usage data from the UK on the the Office for National Statistics (2021) reveal that
volume of card transactions and cash withdrawals, online retail sales have reached record proportions;
presented in Figure 4, supports the survey results. in February 2020, just prior to the first lockdown,
Prior to the pandemic, we see evidence of the 20% of all sales were online – a year later, the figure
longer-term trends of declining cash usage and had reached 36%.
increasing card usage. Then, as the UK enters lock- Some individuals have also moved away from cash
down, there is a lockstep fall in card and ATM trans- usage for in-person transactions, likely due, at least
actions, accompanied by a partial “bounce back” as in part, to concerns about virus transmission via the
restrictions ease over the summer. Following that, handling of cash. Figure 5 shows the spike in search
the volume of cash withdrawals remains especially terms related to cash usage and COVID during the
low, plateauing at approximately 35% below previ- first UK lockdown, paired with increased interest in
ous-year volumes. contactless payment methods and a drop in search-
es for ATMs.
Why Has Cash Usage Fallen?

One reason for the drop in cash usage is the reduc-

tion in opportunities to use it, as restrictions pushed

Figure 4: Year-on-year percentage change in card transactions and cash withdrawals in the UK. Notes: Figure
shows percentage changes in monthly card transactions and ATM withdrawals (by value and by volume), relative
to the same month in the previous year. Average transaction amounts are adjusted for monthly CPI. Source:
Frontier analysis of UK finance and LINK ATM network data.

3
At present, limited usage data are available on payment insight whilst acknowledging the potential for self-re-
methods during the pandemic, which is why we use this porting bias.
survey evidence to provide a contemporaneous source of

Behavioral Economics Guide 2021 34

Paula Papp et al. How COVID Has Changed Consumer Behaviour in Financial Services

Figure 5: Popularity of Google search terms on payment methods and COVID-19. Notes: Search popularity indexed
to maximum number of UK weekly searches between January 2019 and March 2021. Source: Frontier analysis of
Google Trends data.

Will Cash Usage Continue to Fall? Applying online shopping is expected to have continued ef-
Behavioural Economics fects beyond the pandemic.
• Repeated online purchases during lockdown may
There are two ways in which the reduction in cash also have built new online shopping habits for
usage might continue beyond the pandemic: high some consumers. Habit has been shown in the
levels of online shopping may hold, and low levels literature, often alongside the “seductiveness” of
of cash usage for in-store payments may persist. the online shopping experience, to have a posi-
Let’s take a look at each mechanism in the context tive effect on repurchase intention (Khalifa & Liu,
of habits, convenience and social norms. 2007).

Online shopping In-store payments

We expect convenience and habits to play a key We also expect new habits to have an effect on
role in moderating the dash back to the shops as in-store payments beyond the pandemic. The
restrictions are lifted: pandemic prompted many people to review their
payment choice as they sought to minimise contact
• Many consumers will appreciate the convenience
of online shopping more than they used to, or they 4
OnePoll (2020) found that 20% of people buying grocer-
will have experienced it for the first time. This
4
ies online in the UK in 2020 had never considered doing so
new knowledge of the benefits and costs of previously.

Behavioral Economics Guide 2021 35

Paula Papp et al. How COVID Has Changed Consumer Behaviour in Financial Services

with others. As a result, some customers may have mally use cash, tipping them towards a change in
formed different responses to the same contextual behaviour (Van der Cruijsen & Knoben, 2018).
cues, generating new habits which are likely to per-
sist as long as the behavioural cues are encountered What next?
(van der Cruijsen & van der Horst, 2019; Lally et al.,
2011). The decline in cash usage is likely to continue be-
The pandemic may also have affected two types of yond the pandemic, but it will not be felt uniformly.
social norms around payment (Cialdini et al., 1990): For many, the forced shift in behaviour will have
been enough to form lasting habits. But there will
• Injunctive norms: people’s perception of the type remain some for whom cash remains the principal,
of behaviour expected from them. Many consum- if not sole, method of payment. Given that these
ers reported opting not to pay in cash during the people often belong to the most financially vulner-
pandemic, because the vendor no longer accepted able groups, the challenge for banks and retailers
it, or expressed a strong preference for card pay- will be to continue developing their digital capabil-
ment, due to COVID transmission risks. This may ities in an inclusive way.
contribute to a “new norm” whereby consumers
perceive that all vendors prefer cashless pay- Branching Out
ment – which is likely, once they have invested in
terminals and discovered the convenience of not How Has Online Banking Uptake Been Affected by
having to process cash. the Pandemic?
• Descriptive norms: the influence of the behaviour
of others. Seeing more and more people paying High street banking was already in decline in Eu-
with card may influence those who would nor- rope before COVID, with steadily increasing branch

Figure 6: Proportion of adults using online banking. Notes: Labels show the percentage point increase on the
previous year’s usage. Source: Frontier analysis of Eurostat data.

Behavioral Economics Guide 2021 36

Paula Papp et al. How COVID Has Changed Consumer Behaviour in Financial Services

closures and dwindling consumer demand. At the behaviour, psychological commitments based on
same time, preference for online banking was on the factoring in sunk costs or a desire to maintain cog-
rise. As a result of the pandemic, these trends have nitive consistency (Polites & Karahanna, 2012).
accelerated across Europe – as Figure 6 shows. 5
Before the pandemic, inertia is likely to have
The number of registrations for digital banking is played a role in the reticence of some consumers to
only one metric for monitoring the importance of adopt digital banking. In particular, many will have
the online channel during the pandemic. Many ex- perceived the time and effort required to set up and
isting digital users have shifted more of their bank- learn to use digital banking as a hurdle they were
ing online; for instance, the banks we work with unwilling to overcome, particularly given the ready
reported a significant increase in the use of mobile availability of branch or phone services.
banking for simpler transactions, such as checking But the temporary (or indefinite) closure of
balances or transferring small amounts. branches during the pandemic has altered this con-
But the picture is nuanced. Many customers began text. For some, the perceived cost of not changing
to use multiple channels for the first time – and in behaviour has for the first time outweighed the cost
some cases, this involved “digital-only” custom- of doing so. Online banking may have now become
ers seeking more in-person contact. Data from one the status quo – and we may not see a rush back to
high street bank in the UK showed that a third of in-person banking even when branches re-open.
digital-only customers adopted a new in-person
channel in 2020. Banks have also seen a spike in the Perceptions
share of customers updating their details so that
they can be contacted. New exposure to online banking may also have
This shows the continued importance of some tested, and changed, many customers’ perceptions
human element in banking – even for tech-sav- of it.
vy customers. Among branch-only users, only 7% The Technology Acceptance Model, initially de-
have switched to digital channels during the pan- veloped by Fred Davis (1989), is commonly em-
demic, indicating that substantial barriers to digital ployed in the literature. It focuses on perceptions of
usage remain for some. usefulness and ease of use as important determi-
nants of the adoption of new technologies.
Will Online Banking Continue to Grow? Applying In the case of online banking during the pandemic,
Behavioural Economics there is evidence that these perceptions have been
altered as a result of forced exposure. Mastercard
When it comes to digital banking penetration, two (2020), for example, found that 58% of over-65s in
behavioural concepts are important: behavioural the UK said it was easier to use their banking app
inertia and the role of perceptions. than they first thought, with 23% stating that their
confidence in using digital banking had increased
Behavioural inertia since the pandemic had begun.

When faced with new technologies, deci- What next?

sion-makers are sometimes known to demonstrate
behavioural inertia, i.e. a preference for the status The pandemic has accelerated the trend for in-
quo, even when better arrangements exist (Samu- creased digitalisation in retail banking, with large
elson & Zeckhauser, 1988). This behavioural trait numbers of customers adopting digital for the first
can operate through several mechanisms – ration- time. This trend is likely to stick in the post-pan-
al decisions based on perceived costs of changing demic world. Nonetheless, one of the key insights
to have emerged for retail banks is just how impor-
5 A notable exception is the Netherlands, where levels of tant some human element is in banking. For some
digital banking were already very high before COVID-19. customers, this might be a “nice to have” on top of

Behavioral Economics Guide 2021 37

Paula Papp et al. How COVID Has Changed Consumer Behaviour in Financial Services

digital interaction and might be particularly rele- Economics. She has worked for more than 15 years
vant for more complicated queries where in-person advising a wide range of clients in the financial
advice is more reassuring. How the human element sector in addressing issues ranging from strategy
is incorporated into the otherwise seamless digital design and regulatory compliance through to op-
interaction for tech-savvy customers might be- erational organisation. In doing so, she integrates
come an important differentiator in the market. For the use of behavioural economics to help financial
a small minority of users, however, offline will re- service companies understand better how their cli-
main the only way they engage with their banking. ents decide.
It will be a challenge for banks to continue servicing
this customer segment while marching on towards Kalina Kasprzyk is a senior consultant at Frontier.
greater digitalisation. She has over 6 years’ experience advising clients in
the financial services and consumer industries, and
Conclusion: Understanding Behaviour Will her work covers strategy in the commercial, regu-
Be Key to Post-Pandemic Success latory and innovation spaces. She uses her expertise
in behavioural economics and data analysis to help
Financial services providers have had to respond clients design commercial propositions and regula-
quickly to a fluid and uncertain environment over tory responses.
the past year. Now, looking ahead to the medium
and long term, behavioural economics demon- Harry Davies is an analyst at Frontier Economics.
strates that they will need to consider varying de- Harry holds degrees in economics from the Uni-
grees of “stickiness” in new consumer behaviour. versity of Bath and Durham University, and he has
Changes to savings behaviour – mainly invol- worked for organisations across the financial ser-
untary – are expected to be the least sticky. More vices sector as part of Frontier’s Strategy practice.
enduring behavioural change is expected in digital
banking usage, albeit with a continued role for some References
human interaction, and in cash usage, where shifts
to online shopping and changed payment prefer- Achtziger, A., Hubert, M., Kenning, P., Raab, G., &
ences will have a lasting impact. Reisch, L. (2015). Debt out of control: The links
What should businesses do as a result? In all cas- between self-control, compulsive buying, and
es, further changes to operating models, financial real debts. Journal of Economic Psychology, 49,
products and customer service design will be nec- 141-149.
essary. Alessie, R., & Teppa, F. (2009). Saving and habit
Ultimately, the providers who adapt most suc- formation: Evidence from Dutch panel data.
cessfully will be those who are able to understand Empirical Economics, 38(2), 385-407.
the behavioural changes that have taken place, and Beatty, S. E., & Elizabeth Ferrell, M. (1998). Impulse
what has driven them. Essential to this is accessing buying: Modeling its precursors. Journal of
the right metrics – those which capture changes in Retailing, 74(2), 169-191.
customer behaviour in real time and draw out the Cialdini, R. B., Reno, R. R., & Kallgren, C. A. (1990).
nuanced picture across customers. As for anticipat- A focus theory of normative conduct: Recycling
ing future trends, there are no simple answers here. the concept of norms to reduce littering in public
Predicting the future is not easy. But grounding our places. Journal of Personality and Social Psychology,
hypotheses in the science of human behaviour will 58(6), 1015-1026.
make financial institutions all the stronger for it. Davis, F. D. (1989). Perceived usefulness, perceived
ease of use, and user acceptance of information
The Authors technology. MIS Quarterly, 13(3), 319-340.
European Central Bank. (2020a, September).
Paula Papp is an associate director at Frontier COVID-19 and the increase in household savings:

Behavioral Economics Guide 2021 38

Paula Papp et al. How COVID Has Changed Consumer Behaviour in Financial Services

precautionary or forced? (Issue 6/2020). https:// tem acceptance. MIS Quarterly, 36(1), 21-42.
www.ecb.europa.eu/pub/economic-bulletin/ Samuelson, W., & Zeckhauser, R. (1988). Status
focus/2020/html/ecb.ebbox202006_05~d- quo bias in decision-making. Journal of Risk and
36f12a192.en.html. Uncertainty, 1(1), 7-59.
European Central Bank. (2020b, December). Study Shiller, R. J. (2017). Narrative Economics. American
on the payment attitudes of consumers in the euro Economic Review, 107(4), 967-1004.
area (SPACE)? https://www.ecb.europa.eu/pub/ van der Cruijsen, C., & van der Horst, F. (2019). Cash
pdf/other/ecb.spacereport202012~bb2038bbb6. or card? Unravelling the role of socio-psycho-
en.pdf. logical factors. De Economist, 167(2), 145-175.
Jahedi, S., & Méndez, F. (2014). On the advantag- van der Cruijsen, C., & Knoben, J. (2018). Ctrl C Ctrl
es and disadvantages of subjective measures. Pay: Do people mirror payment behaviour of
Journal of Economic Behavior & Organization, 98, their peers? SSRN.
97-114.
Khalifa, M., & Liu, V. (2007). Online consumer re-
tention: contingent effects of online shopping
habit and online shopping experience. European
Journal of Information Systems, 16(6), 780-792.
Lally, P., van Jaarsveld, C. H. M., Potts, H. W. W.,
& Wardle, J. (2009). How are habits formed:
Modelling habit formation in the real world.
European Journal of Social Psychology, 40(6),
998-1009.
Lally, P., Wardle, J., & Gardner, B. (2011). Experiences
of habit formation: A qualitative study.
Psychology, Health & Medicine, 16(4), 484-489.
Loibl, C., Kraybill, D. S., & DeMay, S. W. (2011).
Accounting for the role of habit in regular saving.
Journal of Economic Psychology, 32(4), 581-592.
Mastercard. (2020, November 19). Life under the
‘new normal’ accelerates digital banking adop-
tion across Europe. Mastercard.Com. https://
www.mastercard.com/news/europe/en-uk/
newsroom/press-releases/en-gb/2020/no-
vember/life-under-the-new-normal-accel-
erates-digital-banking-adoption-across-eu-
rope/.
Mody, A., Ohnsorge, F., & Sandri, D. (2012).
Precautionary savings in the Great Recession.
IMF Economic Review, 60(1), 114-138.
Office for National Statistics. (2021, March 26).
Retail sales, Great Britain – Office for National
Statistics. https://www.ons.gov.uk/businessin-
dustryandtrade/retailindustry/bulletins/retail-
sales/february2021#online-retail.
Polites, & Karahanna. (2012). Shackled to the status
quo: The inhibiting effects of incumbent system
habit, switching costs, and inertia on new sys-

Behavioral Economics Guide 2021 39

 Using Behavioural Economics to Drive Better
Outcomes in General Insurance

Daniela Marconi* and Nathalie Spencer**

IAG

While applied behavioural economics has made large inroads in many industries and locations, there is rel-
atively little research to date on its application to the general insurance industry in Australia. In this paper,
we explore how applying a behavioural economics perspective to general insurance can generate positive
outcomes for both consumers and insurers. While we explore the specific example of underinsurance in re-
lation to the availability heuristic, defaults, and round numbers, the main purpose of this paper is to provide
direction on the opportunity areas we believe would be fruitful for further research and application.

Familiar Challenges, Fresh Perspective amount of money (their premium) to be protected

from a potentially devastating loss.
General insurance is vital for protecting consum- This may sound simple, so why can insurance
ers, communities and businesses. But insurance is decisions be hard to get right? The UK’s Financial
also complex, and consumers do not always get it Conduct Authority (FCA) describes five features that
right in terms of protecting their assets by selecting can make financial decisions particularly difficult to
optimal coverage, renewing their policy, and taking make well (Dambe et al., 2013). They occur when:
steps to reduce the likelihood or severity of a loss
event. When consumers do not get these decisions • Products are inherently complex
and behaviours right, it is incumbent on insurers to • Products involve trade-offs between the present
address this issue in order to prevent poor consum- and the future
er and commercial outcomes. • Decisions may require assessing risk and uncer-
The insurance challenges we face today are not tainty
new, but a relatively new additional approach to • Decisions can be emotional
addressing them is available. With its rich insight • Products permit little learning from past mis-
into human decision-making, and its approach to takes.
behaviour change, behavioural economics provides
a useful lens (among others) through which insur- Insurance – whether to purchase it, for which
ers can understand consumers, mitigate these chal- risks and at what price, excess and sum insured –
lenges, and create good outcomes for all. matches all five of these features.
Readers will be familiar with the behavioural
A Perfect Storm: Insurance has the economics tenet that contrary to traditional eco-
Hallmarks of Difficult Decisions, and We’re nomic models of human behaviour, people’s deci-
Only Human sion-making and behaviour are influenced by our
cognitive limitations, biases, heuristics, emotions,
At its core, the main idea of insurance is to pool and other ‘supposedly irrelevant factors’, or SIFs
and spread risk. Many consumers each pay a certain (Thaler, 2016). In the words of Richard Thaler and
Cass Sunstein (2009), rather than the rational au-
* [email protected] tomaton ‘Econs’ of traditional economic models,
** [email protected] we are ‘Humans’, i.e. the very efficient and yet

Behavioral Economics Guide 2021 40

Daniela Marconi & Nathalie Spencer Behavioural Economics to Drive Better Outcomes in General Insurance

Features General Insurance Relevance Behavioural Economic Relevance

Typical of
Error-Prone
Decisions
Products are Underwriters use complex modelling to determine We have cognitive limitations. Levels of numeracy,
inherently the ‘technical price’ of a risk, using up to “thou- literacy, attention and risk tolerance can affect com-
complex sands of lines of calculations” (AFCA, 2020). The prehension and preferences (how well we understand
price is determined by multiple inputs (relating to the product and whether we want it) (ABS, 2008;
the world (e.g. climate trends), the asset (e.g. its el- see Dolan et al., 2010, and BIT, 2014, for reviews).
evation above sea level), the consumer themselves A complex product will have high search costs and
(e.g. a new driver), and other factors. The value is be hard to evaluate, and hassle factors may impede
uncertain and conditional on probabilistic events. action (Bettinger et al., 2012; see BIT, 2014 for
There are many different attributes to compare, a review). Those who are experiencing scarcity
and the maths is not simple. may find this even more challenging than others
(Mullainathan & Shafir, 2013; Shah et al., 2012).
Products To buy an insurance policy, a consumer trades off People often exhibit present-bias and are myopic,
involve trade- an immediate certain amount of money (the premi- narrowly focused on the short term (Laibson, 1997).
offs between the um) for a future possible payout (the claim) less a We may procrastinate (Madrian & Shea, 2001; Thaler
present and the future cost (the excess). & Bernartzi, 2004; Thaler & Sunstein, 2009 for a
future review).
Decisions Insurance decisions are fundamentally about Not only is the maths involved difficult, but also
require assess- assessing risk under conditions of uncertainty: how other factors, like optimism (Puri & Robinson,
ing risk and best to mitigate the risk of a large loss resulting 2007; Sharot, 2011; Weinstein, 1980) or how easy it
uncertainty from an uncertain event. is to imagine the loss (Tversky & Kahneman, 1974;
Johnson et al., 1993), can affect how we evaluate
risk and make judgments in uncertain and ambigu-
ous conditions.
Decisions can be Insurance protection is arguably also about peace of Emotions and states such as stress, anxiety, affection,
emotional mind and avoiding regret, fear, or anxiety. Imag- fear of loss and fear of regret may weigh heavier in a
ining a future loss may be emotionally uncomfort- person’s decision-making than a purely financial as-
able. Searching for information about policies or sessment of the costs, benefits, and expected value
risks might be boring or unpleasant. of the policy (Zeelenberg, 1999; Hsee & Kunreuther,
2000; Aseervatham et al., 2015). We sometimes
avoid seeking out information that can be psycho-
logically distressing (Karlsson et al., 2009).

Products permit There is a mismatch in frequency between decisions Frequency and timing of feedback matters. People
little learning (monthly payment of premiums, annual renewal) tend to learn best when the consequence of their
from past mis- and rare events against which the decisions are decision is immediate. Due to the mismatch in fre-
takes meant to protect. The consequence of a ‘bad’ de- quency of decisions and loss events, consumers may
cision in insurance is only seen infrequently (e.g. use overly short time horizons and therefore mis-
after theft, fire, flooding), with a delay, and over a interpret the feedback. The time horizon of a rare
long time period. Low probability high consequence event may lead to the ‘wrong’ lesson learned, e.g.
(LPHC) events are likely to affect only a few people ‘there were no floods this year, so I wasted money
relatively infrequently. and should not renew my policy’.

Table 1: Features typical of error-prone decisions (col 1, adapted from Dambe et al., 2013) are all evident in the
context of insurance (col 2); behavioural economics can help explain why these decisions are difficult (col 3).

Behavioral Economics Guide 2021 41

Daniela Marconi & Nathalie Spencer Behavioural Economics to Drive Better Outcomes in General Insurance

sometimes fallible, real people of behavioural eco- Risk Hotspots – the Potential for Poor
nomic models. Consumer and Business Outcomes
With this in mind, Table 1 presents an overview of
how general insurance maps back to the FCA’s five Three hotspots that we believe are particularly
features typical of error-prone decisions along with important for general insurance are illustrated in
key concepts from behavioural economics. Taken Table 2.2
together, the complexity and inherent uncertainty Potential outcomes, for example underinsurance,
associated with general insurance, plus our human are likely the result of both traditional economic
decision-making and behavioural tendencies, paint factors – including affordability, misaligned incen-
the picture that: tives, or information asymmetries – and behav-
ioural economic factors (as described in Table 1).3
• People may not even understand the value equa- Understanding the decision-making processes of
tion of insurance, or they are unable to calculate it consumers buying insurance, and how these pro-
(due to pricing complexity and lack of transpar- cesses interact with how actuaries, underwriters
ency, as well as cognitive errors and limitations) and insurers develop, price and market insurance
• Even if they did know it and could calculate it, products, is increasingly important. While a behav-
other factors (biases and heuristics, sometimes ioural economics view will not be the silver bullet,
based on SIFs) may come into play to influence without it, an insurer will have an incomplete pic-
people’s evaluations of its components ture of the issue at hand and the approaches that
• The economic equation misses out on unquantifi- may help to solve them.
able components that nevertheless matter greatly Let us explore one of these persistent challenges
(emotional influences and potentially competing in more depth: suboptimal levels of insurance.
goals). Poor decisions about coverage can result in a con-
sumer being over-insured or underinsured (includ-
The point of understanding behavioural econom- ing non-insured).
ics by contrasting it with traditional economics is
not to suggest that we are all hopelessly imperfect,
but rather to highlight that certain aspects of in- 2 The hotspots in this paper are based on the deci-
surance may have been inadvertently designed with sion-making and behaviour of the consumer, aka de-
the traditional models in mind . Where traditional
1
mand-side hotspots. There are also supply-side risk hot-
economic benchmark models and the behaviour- spots identified in the academic literature (such as not
al economic models differ is where we might see offering cover despite it making financial sense to do so;
market inefficiency, or what we call “risk hotspots” not offering cover at prices that accurately reflect risk; re-
– areas with high risk of poor consumer and com- insurance prices decreasing after a period without disas-
mercial outcomes. ters) which are outside the scope of this article. See Kun-
reuther et al. (2013).
3 Price is a traditional economic concept, but a customer’s
perception of affordability will be affected by behavioural
1 For example, disclosure documents aim to improve con- economic concepts, as it involves making trade-offs that
sumer outcomes by providing more information. The tra- will depend on the salience and priority of other finan-
ditional economic model predicts that more information cial commitments, value perception, time bracketing, etc.
will lead to a consumer making a better choice. A behav- Price alone is not enough to solve for suboptimal insur-
ioural economic perspective, on the other hand, would ance; for example, Kunreuther & Pauly 2015 found that
suggest that in their current form – several pages long even when insurance is offered at a better-than-technical
and full of jargon – these disclosure documents are argu- price, some people still fail to buy it. The authors suggest
ably not fit for purpose, as they do not cater to our limited using behavioural economic tools to enhance the messag-
attention span and literacy levels. ing and thus to increase adoption.

Behavioral Economics Guide 2021 42

Daniela Marconi & Nathalie Spencer Behavioural Economics to Drive Better Outcomes in General Insurance

Hotspots Risk of Poor Outcomes for

Consumers Insurers

Sub-optimal decisions about coverage Financial devastation [underinsurance, Missed opportunity for consumer
no insurance] growth, revenue, data, economies of
Paying too much [over-insurance] scale [underinsurance]

Insufficient engagement at renewal Renewed policies may not adequately Regulatory scrutiny
time reflect changes to circumstances Eventual attrition

Failure to practice safer behaviour Unpreparedness and longer recovery Higher cost of claims
despite good intentions Higher premiums

Table 2: Key hotspots in general insurance that have the potential for poor consumer and commercial outcomes.

Underinsurance (including non-insurance) is a a consumer defines this factor, and as such, con-
material problem for Australia; by the end of 2020, sumers and insurers might be optimising (or satis-
it was reported that 37.5% of all households were ficing) in relation to different factors and over dif-
underinsured (DBM Consumer Atlas, n.d.). For a ferent time horizons. As a business, insurers rely on
consumer, the consequences of underinsurance can their understanding of underwriting and insurance
be financially devastating, including loss of shel- risk over time in the absence of a consumer’s per-
ter or mobility. For an insurer, underinsurance is sonal or household financial ecosystem, including
a missed opportunity for revenue. Moreover, more their financial constraints. An insurer will
customers or policies in the market has collateral therefore typically view “optimal” insurance for
benefits such as economies of scale and more data, the consumer as sufficient coverage to restore
which could lead to more accurate risk pricing. them to their pre-event financial state. This
Research to date has begun to understand sub- suggests that insurers take a relatively long-
optimal insurance coverage in the following ways: term perspective when working towards positive
analysing the difference between insured losses and consumer outcomes. If the consumer’s main
the total economic loss value (the “protection gap” objective, on the other hand, is to manage a
(ANZIIF, n.d.), identifying where consumers are short-term household budget, their best outcome
paying far too much or too little compared to the may be different than if they were managing a
expected (financial) value of the policy, thus imply- longer-term or (unknowable) lifetime budget.
ing unrealistic levels of risk tolerance/risk aversion, Future research should explore this topic in more
or comparing the number of assets (e.g. households depth.
or cars) in the market to the number of insurance As mentioned earlier, there may be many differ-
policies sold. ent factors contributing to underinsurance, so in
Nevertheless, it is difficult to pinpoint exactly order to identify whether – or where – behavioural
what objectively constitutes suboptimal coverage, economics matters, insurers can begin with a be-
because the value exchange includes aspects that havioural exploration or audit. This may
are subjective and differ between people, such as include looking for anomalous consumer
risk tolerance and peace of mind. behaviour, evidence of SIFs mattering, or where it
Note that the ways in which the industry deter- appears that a consumer’s decision-making and
mines “optimal” coverage may be different to how behaviour are not in their own best interest.

Behavioral Economics Guide 2021 43

Daniela Marconi & Nathalie Spencer Behavioural Economics to Drive Better Outcomes in General Insurance

Availability Heuristic natural experiment, that when the right to sue for
pain and suffering was included in a default motor
People tend to make judgments about the like- insurance policy, the purchase rate of this right was
lihood of an event based on how easily it comes to higher than when not included as a default.
mind, whether through direct experience or media This finding suggests that defaults matter when
stories (Tversky & Kahneman, 1974). deciding about what types of insurance coverage to
Research from other countries finds evidence to purchase, but it is also likely that defaults affect the
support that the availability heuristic influences amount of coverage purchased. More specifically,
insurance decisions. Analysing flood insurance once a consumer has decided to purchase insurance
in the USA, for instance, researchers found that and is going through the process of selecting their
people tend to buy insurance policies after expe- sum insured and excess amounts, they may stick
riencing or hearing about an event (when the risk with the default amounts presented to them on
is more “available” and front of mind), and they screen. In a sample of over 45,000 customers pur-
cancel insurance policies over time if they have not chasing building and contents insurance online, we
experienced a bad event (Gallagher, 2014; Atreya find that over four in 10 customers stick with the sum
et al., 2015; Michel-Kerjan et al., 2012). Gallagher insured value that an online sum insured calculator
(2014) finds that flood insurance take-up increases (consumer calculation tool) presents, which serves
by 9% following a flood in the same area. Despite as a default (IAG internal data). A similar proportion
their mathematical expertise, insurers themselves adjust their sum insured amount downwards from
have also been shown to exhibit the availability the default, which could potentially be an indication
bias, by limiting the supply of coverage after ter- of affordability issues, optimism, myopia, or other
rorist events or natural disasters (Kunreuther & factors yet unexplored. Only 15% adjust their sum
Pauly, 2015). Relatedly, a more vividly detailed risk insured upwards (IAG internal data).
may be perceived as higher risk than a less vividly Defaults are therefore important for insurers,
described risk (Johnson et al., 1993). Johnson et al.’s because how the manner in which different types
laboratory study found that people were willing to of policies or amounts of coverage are presented
pay more for (hypothetical) insurance when the risk might affect the consumer’s ultimate protection.
was more specific, despite the coverage being more
narrow and therefore objectively worth less. Number Roundness
The importance of the availability heuristic for in-
surers is that consumers may be more likely to buy People may prefer round numbers for their cog-
or renew a policy, or engage in safer behaviour, if nitive ease (Kettle & Haeubl, 2010). Wadhwa and
they have recently experienced a loss event or seen Zhang (2015) studied the roundness of numbers in
it covered in local media. Conversely, consumers product pricing and found that people want to “feel
might exit insurance (cancelling or not renewing right” about their decision-making with respect to
policies) prematurely if they do not experience a purchases. For goods that are utilitarian (practical),
bad event. people were happy to settle on non-rounded prices,
whereas for hedonic (luxury) goods, they preferred
Defaults round numbers for their cognitive ease. When con-
sumers already have a high cognitive load, they
Defaults are a well-known behavioural tool that tend to prefer rounded numbers; when there is low
can influence behaviour (Jachimowizc et al., 2019). cognitive load, they prefer exact numbers. While
In insurance, they have been found to influence the this work focused on pricing and not explicitly
take-up of coverage for different risk types (Johnson about sum insured amounts, it is a useful input to
et al., 1993), especially for consumers who are inex- help us hypothesise and explore what this means in
perienced with insurance decisions (Robinson et al., an insurance context.
2021). For example, Johnson et al. (1993) found, in a Insurance is more utilitarian than hedonic, so

Behavioral Economics Guide 2021 44

Daniela Marconi & Nathalie Spencer Behavioural Economics to Drive Better Outcomes in General Insurance

given Wadhwa & Zhang’s study we might expect thereby casting doubt on a potential rationale that
consumers to prefer exact numbers. However, in- picking this sum insured amount might be an at-
surance is also very complex, as described above, tempt to simplify household budgeting.
so consumers are likely to have high cognitive load Taken together, these concepts help us under-
during the quote process when they are inputting stand that factors such as how (and how much) a
details about their assets and making choices about risk is covered in the media, whether risk types are
coverage, excess, and sum insured. Consequently, included or add-on, and digital sum insured calcu-
on balance, we might expect that high cognitive lator amounts could have a meaningful impact on
load may be driving consumers to prefer round a consumer’s decision, and the consumer might
numbers when selecting their sum insured amount. adjust their sum insured value to satisfy a number
Indeed, in a sample of over 45,000 customers using of factors, including affordability, optimism, and
an online channel to purchase insurance, sum in- potentially the roundness of a number.
sured values cluster at $50k intervals (Figure 1; IAG
internal data). Using a Behavioural Economic Approach to
Note that the roundness of the sum insured num- Improve Consumer Outcomes
ber is especially “irrelevant” (as in a Supposedly
Irrelevant Factor), because a round number in this While behavioural economics might enhance our
regard will not necessarily (nor even likely) result understanding of “why” these risk hotspots are
in a round number monthly or annual premium, present, it is also invaluable in providing levers for

Figure 1: Buildings and HPAC New Business acquired through Web Channel, by Building Sum Insured Value.

change. To date, strategies for designing effective have not been enough to solve these seemingly in-
or impactful change in insurance are still heavily tractable problems. We argue that the addition of
reliant on traditional levers such as: (1) financial in- behavioural economic levers will go a long way to
centives, (2) rules and regulations and (3) the pro- mitigating these challenges and driving better con-
vision of information and education (Soman, 2015). sumer outcomes.
Unfortunately, these traditional levers alone Behavioural economics can be leveraged in differ-

Behavioral Economics Guide 2021 45

Daniela Marconi & Nathalie Spencer Behavioural Economics to Drive Better Outcomes in General Insurance

ent ways to improve consumer outcomes, but argu- also present an opportunity: if insurers are able
ably, the most impactful and sustainable way will be to add behavioural economics to their suite of
one that works towards wholesale, systemic behav- existing capabilities and approaches, there is the
ioural and societal change. Although we may focus opportunity to mitigate these persistent chal-
on the addition of choice architecture as a lever for lenges and provide better outcomes to both
change, amongst traditional levers, to “remedy consumers and their own bottom line. Insurers
each bias”, “optimise touch points” or “improve can gain a richer understanding of their cus-
products, services and the consumer experience” tomers and translate this knowledge into real-
(amongst others), we would be remiss if we did not world impacts that are purpose-driven and
ask ourselves “for what purpose?” along the way. measurable. With Australian financial services
To be clear, using a behavioural economics ap- regulation moving towards a focus on fairness
proach is not about getting everyone to purchase (Robinson, 2020) and the scrutiny of how product
more insurance protection per se; rather, it is about development and distribution practices drive
supporting decision-making and behaviour in a consumer outcomes (ASIC, 2018), there is no
way that contributes positively to the emotional, better time than now to ramp up a behavioural
physical and financial wellbeing of an individual economics approach in general insurance.
and society, should something “bad” happen. In
practice, this means placing a greater emphasis on
The Authors
behavioural research, analysis, and design within a
system, as opposed to focusing solely on discrete, Daniela Marconi is Director of Customer Behav-
independent decisions (Khan & Newman, 2021). iour at Insurance Australia Group (IAG), and her
A recent review of the literature by Pitthan and team is responsible for driving and embedding
DeWitte (2021) suggests that the biases and heu- the application of behavioural sciences across the
ristics that affect insurance decision-making are Group. She has over 13 years of experience in both
mostly driven by lack of knowledge or misuse of fi- clinical and commercial applications of psychology.
nancial concepts and products, linked to low levels Daniela holds a master’s degree in Clinical Psychol-
of financial literacy. As such, they propose financial ogy and an Honour’s Degree in Psychology from the
literacy treatments that serve to improve knowl- University of the Witwatersrand, as well as a bach-
edge and use of finance as a potential systematic elor’s degree in Psychology from the University of
solution. While financial literacy treatments may Cape Town. She ran a successful clinical psychology
be an appropriate intervention in some cases, we private practice before moving into the commercial
argue that for an insurer, a systemic, purpose-led sector, where she has applied behavioural science
approach – using both behavioural economics and in management consulting, policy & regulation and
traditional levers (such as awareness, education, financial services.
and community and stakeholder engagement) – is
important in remedying risk hotspots. Nathalie Spencer is Principal, Customer Be-
What matters most is that this approach trans- haviour at Insurance Australia Group (IAG). She
lates into real-world, measurable impacts. Along- is an experienced applied behavioural scientist in
side asking the question ‘for what purpose?’, it is financial services, working to improve customer
imperative to design consumer behavioural out- outcomes. Nathalie has a Master of International
come measures that ensure we are delivering on Economic Studies in Behavioural Economics from
this purpose every step of the way. Maastricht University and a Bachelor of Commerce
from McGill University. She is author of the book
Conclusion “Good Money: understand your choices, boost your
financial wellbeing” and is an active contributor
The same challenges that insurers face with re- and writer in the field of behavioural science for an
spect to consumer decision-making and behaviour international audience.

Behavioral Economics Guide 2021 46

Daniela Marconi & Nathalie Spencer Behavioural Economics to Drive Better Outcomes in General Insurance

References DBM Consumer Atlas. (n.d.). Atlas - DBM

Consultants. http://www.dbmconsultants.com.
ABS Australian Bureau of Statistics. (2008). au/expertise/atlas/.
Adult literacy and life skills survey, sum- Dolan, P., Hallsworth, M., Halpern, D., King, D., &
mary results, Australia 2006 (Reissue). Vlaev, I. (2010). MINDSPACE: Influencing behav-
https://www.abs.gov.au/AUSSTATS/abs@. iour for public policy. Cabinet Office.
nsf/Previousproducts/4228.0Main%20 Gallagher, J. (2014). Learning about an infrequent
Features22006%20(Reissue). event: Evidence from flood insurance take-up
AFCA. (2020). Determination for case num- in the United States. American Economic Journal:
ber 696449. https://service02.afca.org.au/ Applied Economics, 6(3), 206-233.
CaseFiles/FOSSIC/696449.pdf. More info at Hsee, C. K., & Kunreuther, H. C. (2000). The affec-
https://www.insurancenews.com.au/daily/cus- tion effect in insurance decisions. Journal of Risk
tomer-wins-dispute-over-79-premium-spike . and Uncertainty, 20(2), 141-159.
ANZIIF. (n.d.). Global protection gap. https://anziif. Jachimowicz, J. M., Duncan, S., Weber, E. U., &
com/members-centre/the-journal-articles/ Johnson, E. J. (2019). When and why defaults
volume-41/issue-3/the-global-protection-gap. influence decisions: A meta-analysis of default
Aseervatham, V., J. Jaspersen, & Richter, A. (2015). effects. Behavioural Public Policy, 3(2), 159-186.
The affection effect in an incentive compati- Johnson, E. J., Hershey, J., Meszaros, J., &
ble insurance demand experiment. Economics Kunreuther, H. (1993). Framing, probability dis-
Letters, 131, 34-37. tortions, and insurance decisions. Journal of Risk
ASIC. (2018). Design and distribution obliga- and Uncertainty, 7, 35-51.
tions and product intervention power: Revised Karlsson, N., Loewenstein, G., & Seppi, D. (2009).
exposure draft legislation. Submission by The ostrich effect: Selective attention to infor-
the Australian Securities and Investments mation. Journal of Risk and Uncertainty, 38(2),
Commission. https://download.asic.gov.au/ 95-115.
media/4849144/design-and-distribution-obli- Kettle, K., & Häubl, G. (2010). Numeric fluency and
gations-and-product-intervention-power-re- preference. Advances in Consumer Research, 37,
vised-exposure-draft-legislation-submis- 150-152.
sion-by-asic.pdf. Khan, Z., & Newman, L. (2021). Building behavioural
Atreya, A., Ferreira, S., & Michel-Kerjan, E. (2015). science in an organization. Action Design Press.
What drives households to buy flood insurance? Kunreuther, H. C., & Pauly, M. V. (2015). Behavioral
New evidence from Georgia. Ecological Economics, economics and insurance: Principles and solu-
117, 153-161. tions. In D. Schwarcz & P. Siegelman (Eds.),
Bettinger, E. P., Long, B. T., Oreopoulos, P., & Research handbook on the economics of insurance
Sanbonmatsu, L. (2012). The role of application law (pp. 15-35). Edward Elgar Publishing.
assistance and information in college decisions: Kunreuther, H. C., Pauly, M. V., & McMorrow, S.
Results from the H&R Block FAFSA experiment. (2013). Insurance and behavioral economics:
The Quarterly Journal of Economics, 127(3), 1205- Improving decisions in the most misunderstood in-
1242. dustry. Cambridge University Press.
BIT. (2014). EAST: Four simple ways to apply be- Laibson, D. (1997). Golden eggs and hyperbolic
havioural insights. Behavioural Insight Team, discounting. The Quarterly Journal of Economics,
London. 112(2), 443-478.
Dambe, K., Hunt, S., Iscenko, Z., & Brambley, W. Madrian, B. C., & Shea, D. F. (2001). The power
(2013). Applying behavioural economics at the of suggestion: Inertia in 401(k) participation
Financial Conduct Authority. FCA Occasional and savings behavior. The Quarterly Journal of
Paper No. 1. SSRN. https://ssrn.com/ab- Economics, 116(4), 1149-1187.
stract=2930007. Michel-Kerjan, E., Lemoyne de Forges, S., &

Behavioral Economics Guide 2021 47

Daniela Marconi & Nathalie Spencer Behavioural Economics to Drive Better Outcomes in General Insurance

Kunreuther, H. (2012). Policy tenure under the under uncertainty: Heuristics and biases. Science,
US national flood insurance program (NFIP). 185(4157), 1124-1131.
Risk Analysis: An International Journal, 32(4), Wadhwa, M., & Zhang, K. (2015). This number just
644-658. feels right: The impact of roundedness of price
Mullainathan, S., & Shafir, E. (2013). Scarcity: Why numbers on product evaluations. Journal of
having too little means so much. Macmillan. Consumer Research, 41(5), 1172-1185.
NSW Government. (June, 2020) Emergency Services Weinstein, N. D. (1980). Unrealistic optimism about
Levy Insurance Monitor end of office review. future life events. Journal of Personality and Social
https://www.eslinsurancemonitor.nsw.gov.au. Psychology, 39(5), 806- 820.
Pitthan, F., & De Witte, K. (2021). Puzzles of in- Zeelenberg, M. (1999). Anticipated regret, expect-
surance demand and its biases: A survey on the ed feedback and behavioural decision-making.
role of behavioural biases and financial literacy Journal of Behavioural Decision-Making, 12(2),
on insurance demand. Journal of Behavioural and 93-106.
Experimental Finance. https://doi.org/10.1016/j.
jbef.2021.100471.
Puri, M., & Robinson, D. T. (2007). Optimism and
economic choice. Journal of Financial Economics,
86(1), 71-99.
Robinson, D. (2020) Limelight: Enforcing ‘fairness’
– the future of insurance claims handling in
Australia. GC Legal website. https://www.gclegal.
com.au/limelight-newsletters/enforcing-fair-
ness-the-future-of-insurance-claims-han-
dling-in-australia.
Robinson, P. J., Botzen, W. W., Kunreuther, H., &
Chaudhry, S. J. (2021). Default options and in-
surance demand. Journal of Economic Behavior &
Organization, 183, 39-56.
Shah, A. K., Mullainathan, S., & Shafir, E. (2012).
Some consequences of having too little. Science,
338(6107), 682-685.
Sharot, T. (2011). The optimism bias. Current Biology,
21(23), R941-R945.
Soman, D. (2015). The last mile: Creating social
and economic value from behavioral insights.
University of Toronto Press.
Thaler, R. H. (2016). Behavioural economics: Past,
present, and future. American Economic Review,
106(7), 1577-1600.
Thaler, R. H., & Benartzi, S. (2004). Save more to-
morrow: Using behavioral economics to increase
employee saving. Journal of Political Economy,
112(S1), S164-S187.
Thaler, R. H., & Sunstein, C. R. (2009). Nudge:
Improving decisions about health, wealth, and
happiness. Penguin.
Tversky, A., & Kahneman, D. (1974). Judgment

Behavioral Economics Guide 2021 48

 Brick by Brick: How Targeted Messaging Can
Help Customers With Low Savings Build a Buffer

Laura Straeter* and Maria Ferreira-Sequeda

ING

Having sufficient savings is extremely important for dealing with financial shocks and staying financially
healthy; nevertheless, many Europeans do not have such a savings buffer, and so they could use some help.
Building on behavioural science, we test two interventions to encourage vulnerable ING customers (i.e. with
less than €2,000 savings) in the Netherlands to increase their buffer. Focusing on actual transfers to savings
accounts and the net saving balance, we examine the influence that the fear of missing out and temporal
reframing can have on savings growth. The studies demonstrate that the target group transferred €48 more
to their savings accounts when the fear of missing out was activated, and when saving was presented in
bite-size, recurring steps, they transferred €68 more to their savings accounts, resulting in a €25 increase
in the net balance. These findings show how fruitful behaviourally-driven actionable insights can be in
guiding saving behaviour.

Introduction dicated that they did not have any savings, a number
that has been relatively consistent over the past few
Building a sufficient rainy-day fund has always years. Moreover, nearly six out of ten Europeans
been important. Savings are crucial for dealing were dissatisfied with what they saved each month
with financial shocks such as income reduction or (INTRUM, 2020). People with low savings could
large, unexpected expenses, and staying financial- thus use some help to build a buffer, a role that fits
ly healthy. Nonetheless, many people struggle to financial institutions well. At the Think Forward
accumulate savings and fail to build and maintain Initiative and ING, we have taken this responsibili-
a sufficient buffer, making them vulnerable to run- ty and are applying behavioural science to increase
ning into other financial issues such as problematic savings among ING’s vulnerable customers with
debt. less than €2,000 savings in The Netherlands.1
Although there is no general agreement on the Although most people know that saving is impor-
minimum size of a household emergency fund, the tant, for many it is still hard to put a little money
UK Money Advice Service suggests holding three aside. Behavioural change theories help us un-
times the household’s monthly expenses in an in- derstand why people have difficulty in saving or
stant access savings account. NIBUD, the Dutch changing this habit (Fogg, 2009; Michie et al., 2011).
equivalent institution, suggests people put aside For people to change behaviour and start a healthy
10% of their monthly income. Easy and simple rules saving routine, they need to be triggered by the
of thumb, you could say, but saving is not easy for environment, feel motivated and – importantly –
everyone. Some people don’t feel motivated to save,
whereas others simply do not have the capacity to 1
In a five-month collaborative programme, multiple ac-
do so, because their expenses absorb all of their tionable insights were developed and tested in the ING
income. According to the ING International Survey, banking app, aiming to help people with low or no sav-
conducted in December 2020, 22% of Europeans in- ings increase their savings buffer and build saving habits.
The two studies presented in this research are part of this
* Corresponding author: [email protected]. larger experimental programme.

Behavioral Economics Guide 2021 49

Laura Straeter & Maria Ferreira How Targeted Messaging Can Help Customers With Low Savings Build a Buffer

have the ability to save. When one of these elements (Przybylski et al., 2013; Riordan et al., 2015). It is
is lacking, it is harder to start and/or stick to a sav- often believed to affect individuals negatively, as it
ings plan (Fogg, 2009). For example, some people has the power to increase purchase decisions (Abel
may need to spend every spare penny to pay off a et al., 2016; Hodkinson, 2019). But would it also be
consumer loan, which makes it impossible for them possible to use the fear of missing out to boost sav-
to save. However, as soon as they have paid off their ing instead of spending money?
loan, a saving opportunity arises. Finding those This fear seems to be particularly activated in the
sweet spots, and using behavioural techniques to case of partial belonginess, i.e. relatedness with-
encourage individuals, could set the stage for driv- out interaction (Baumeister & Leary, 1995). Due to
ing change. Covid-19, many Europeans have been socially re-
In this article, we test two interventions based stricted: mostly home-bound and allowed to meet
on influential behavioural techniques to help vul- only very few social contacts. However, with the
nerable people (i.e. ING primary customers with vaccination programmes on the roll, there seems
less than €2,000 savings) increase their buffer. We to be light at the end of the tunnel; consequently,
believe that the value of such techniques lies in ap- people could be more receptive to small reminders
plications offered at the right time and with a direct about missing out on future social activities. Nev-
opportunity to take action. ING offers a range of ertheless, would they consider saving money now,
tools in its digital environment, to help customers which they could spend when restrictions are fur-
build a financially healthy future, and sends ac- ther relaxed, to keep their fear of missing out to a
tionable insights to customers in its banking app. minimum?
This combination of targeted communication (to
trigger and motivate customers to save) and linking Temporal Reframing
the right tools (to improve the ability to save) could
make the difference in helping people low in sav- Saving and spending can feel like a trade-off
ings build a buffer. (Frederick et al., 2002). Saving reduces the amount
There are numerous behavioural techniques that free to spend and can be perceived as a direct loss in
can steer people’s decisions. This research does not spending power (Hershfield et al., 2020). Temporal
aim to test all techniques; instead, it hopes to pro- reframing, which could help soothe the psychologi-
vide insights into how some of these concepts can cal pain of reduced spending power (Goldstein et al.,
be employed to help people most in need. Specifi- 2016; Hershfield et al., 2020), involves presenting
cally, we test the influence that the fear of missing large costs as a series of smaller ongoing expenses
out and temporal reframing can have on the amount (Gourville, 1998). It is often used to sell big-ticket
saved among the financially vulnerable. items such as electronics and cars by breaking down
the cost price into daily, weekly or monthly instal-
Fear of Missing Out ments (Bambauer-Sachse & Grewal, 2011).
Although most research on temporal framing
People have a central need to be around others. has focused on cost perceptions in purchases, we
They want to form and maintain lasting, posi- believe, like others (Hershfield et al., 2020; Col-
tive and significant interpersonal relationships by & Chapman, 2013), that it could be a successful
(Baumeister & Leary, 1995; Maslow, 1943). This instrument for driving savings. Hershfield et al.
need to belong is universal and an important driver (2020), for instance, found that presenting savings
of our emotional responses and behaviour. It even in a more granular format encouraged people, par-
causes people to experience the fear of missing out ticularly those with a low income, to participate in
(FOMO; Xie et al., 2018). a recurring deposit programme. Moreover, when
The fear of missing out is the desire of people to presenting a savings goal in weekly subgoals, people
stay continually connected aligned with the uneasy were more willing to forego a small purchase in or-
feeling that they are missing out on social activities der to reach it (Colby & Chapman, 2013). These find-

Behavioral Economics Guide 2021 50

Laura Straeter & Maria Ferreira How Targeted Messaging Can Help Customers With Low Savings Build a Buffer

ings are promising but still leave the question open The customers could transfer money to their sav-
as to whether temporal reframing also increases ings account(s) directly, but they could also do it at
the actual amount of money saved instead of only a later date manually or not transfer any money at
the intention to do so. We will test whether it has all. Customers in the No insight condition did not
this direct consequence on actual savings among receive an insight on saving. In all three conditions,
our vulnerable target group, varying the levels of participants’ savings were tracked over the two-
temporal granularity (i.e. weekly or monthly steps). week period in terms of the amount transferred to
Temporal reframing of saving could be especial- their savings account(s) and the net savings balance
ly important to people in a stretch. Given they are at the end of this period (i.e. the amount transferred
likely to have a restricted income, saving via small to their savings accounts minus the amount with-
recurring steps could be helpful in building habits drawn). We identified 441 participants (0.4%) as
and constructively growing a buffer. outliers and removed them from the analysis.3

Experiment 1: Fear of Missing Out Results

In Experiment 1, we tested whether ING primary Amount transferred into savings account(s). A
customers with small amounts of savings would be one-way ANOVA on the amount transferred into
receptive to insights reminding them of the fear of customers’ savings accounts showed a significant
missing out on future social activities. We expect- difference across the three conditions, F(2,123944)
ed that in the current context of COVID-19, where = 216.60, p <.001. Simple contrasts (t-tests)
many people have been limited in their interactions showed that customers in the FOMO condition (MFO-
with others, customers would be sensitive to such MO
=248.71, SDFOMO=641.82) transferred most money
an insight motivating them to save extra money to during the two-week experimental period, followed
spend in the future. by customers in the No insight (MNone=200.42, SD-
None
=570.72) and the Neutral (MNeutral=171.79, SDNeu-
Method tral
=563.78) conditions. All contrasts were signifi-
cant (t’s(123944) >6.02, p’s <.001).
Participants. In total, 124,388 ING primary cus-
tomers with less than €2,000 in savings participat- Net savings balance. A one-way ANOVA on the
ed in our field experiment (Mage=43.10, SDage=16.54, customers’ net savings balances showed a sig-
50.9% male). Customers were randomly assigned nificant difference across the three conditions,
to one of the three conditions (FOMO, Neutral and F(2,123944)= 4.57, p =.01. Customers in the FOMO
No insight) in a between-subjects design. Both the condition (MFOMO=48.46, SDFOMO=922.17) had the
Neutral and the No insight conditions served as highest net savings balance at the end of the two-
controls. week experimental period, followed by customers
Procedure. Over a two-week period, participants in the No insight (MNone=39.03, SDNone=776.77) and
in the FOMO and Neutral conditions were shown an the Neutral (MNeutral=26.93, SDNeutral=1446.93) condi-
insight (at least once) after they logged into their tions. Simple contrasts (t-tests) showed that only
banking app. In the FOMO condition, participants the difference between the FOMO condition and the
received a slightly different insight than in the Neutral condition was significant (t(123944) = 2.85,
Neutral condition (see Table 1). After receiving the 2
p =.004), but the other contrasts were not signifi-
insight, customers could click on a button which cant (t’s(123944)< 1.433, p’s >.152).
would direct them to the transfer feature of the app.

Customers receiving the insight could choose to accept,

2 3
In Experiments 1 and 2, outliers were detected based on
snooze or delete it. Those who snoozed the insight would the studentised residuals (SRs) of both dependent varia-
see it again within the two-week period. bles. Cases with |SR| > 3 were marked as outliers.

Behavioral Economics Guide 2021 51

Laura Straeter & Maria Ferreira How Targeted Messaging Can Help Customers With Low Savings Build a Buffer

Table 1: Actionable insights per condition in Experiment 1.

Figure 1: Amount transferred to savings account(s) per condition in Experiment 1. Note: Error bars represent ± 1
standard error.

Discussion neutral insight.

Yet, an insight triggering the fear of missing out
The findings demonstrate that an insight trig- did not increase the net savings balance of cus-
gering the fear of missing out on future spending tomers in our target group after the two-week
positively influenced the amount of money trans- experimental period. Net savings balance has two
ferred into the savings accounts of the customers in components: the amount transferred to and the
our target group. Customers receiving this insight amount withdrawn from the savings account. Next
transferred, on average, €48 more to their savings to driving the amount transferred to the savings ac-
account(s) than those who did not receive any in- count, our insight triggering the fear of missing out
sight, and €77 more than those who received the also unexpectedly led to an increase in the amount

Behavioral Economics Guide 2021 52

Laura Straeter & Maria Ferreira How Targeted Messaging Can Help Customers With Low Savings Build a Buffer

Figure 2: Net savings balance per condition in Experiment 1. Note: Error bars represent ± 1 standard error.

withdrawn from the savings accounts of our target we present saving as a series of granular, bite-size
group. Although the FOMO insight suggested to
4
steps towards a larger goal. This aligns with previ-
people to save, linking the insight to future spend- ous research (Colby & Chapman, 2013; Hershfield
ing could have also activated a spending mindset. et al., 2020), which varied in terms of whether the
Further research is needed to validate this assump- saving goal was represented as a monetary amount
tion. However, to prevent the activation of a spend- or a symbolic object (e.g. an iPad or a TV). Therefore,
ing mindset in the next experiment, we decided to next to testing the role of different levels of tem-
test insights that did not relate to future spending. poral granularity, we also decided to test whether
savings depend on the way the goal is represented.
Experiment 2: Saving a Little Adds up
Method
The second experiment aimed to test a behavioural
technique that shifts people’s perception of saving: Participants. In total, 130,740 ING primary cus-
temporal framing. We believe that people with low tomers with less than €2,000 savings participated
savings are more likely to save when they believe it in our field experiment (Mage=39.81, SDage=15.81,
will not heavily reduce their spending power, and so 48.4% male). Customers were randomly assigned
to one of the five conditions in our 2 (Temporal re-
4
One-way ANOVA on the amount withdrawn: F(2,123944)= framing: €5 weekly vs. €20 monthly) × 2 (Goal rep-
31.584, p <.01. Customers in the FOMO condition (MFO- resentation: Monetary vs. Symbolic) + control (No
MO
=200.25, SDFOMO=916.02) withdrew most money during insight) between-subjects design.
the two-week experimental period, followed by custom- Procedure. The procedure was similar to that of
ers in the No insight (MNone=161.38, SDNone=772.55) and the Experiment 1. Customers in the four treatment con-
Neutral (MNeutral=144.87, SDNeutral=1432.23) conditions. All ditions were shown an insight (at least once) after
contrasts were significant (t’s(123944) > 1.972, p’s <.049). they logged into their banking app over a period of
This means that customers receiving the FOMO insight two weeks (see Table 2). In the No insight condition,
withdrew, on average, €39 more from their savings ac- customers did not receive an insight to save. In all
count(s) than those who did not receive any insight, and five conditions, participants’ savings were tracked
€70 more than those who received the neutral insight. over the two-week period in terms of the amount

Behavioral Economics Guide 2021 53

Laura Straeter & Maria Ferreira How Targeted Messaging Can Help Customers With Low Savings Build a Buffer

transferred to their savings account(s) and the net Results

savings balance at the end of this period (i.e. the
amount transferred to their savings accounts minus Amount transferred into savings account(s). A
the amount withdrawn). We identified 408 partic- two-way ANOVA of temporal reframing and goal
ipants (0.3%) as outliers and removed them from representation on the amount transferred by
the analysis. customers into their savings accounts showed
no significant difference in temporal reframing

Table 2: Actionable insights per condition in Experiment 2.

and goal representation, or an interaction effect ferent goal representations did not have the expect-
(F’s(1,130327)< 1.996, p’s >.158). An independent ed effect and worked equally well.
t-test was run on the amount transferred into sav- Net savings balance. A two-way ANOVA of temporal
ings accounts, comparing the No insight condition framing and goal representation on the customers’
to all four treatment conditions, to test whether net savings balance showed no significant differ-
there was a general effect of temporal reframing. ence of temporal reframing and goal representa-
The results showed a significant effect (t(130327)= tion, or an interaction effect (F’s(1,130327)< 0.483,
12.983, p <.001; MTreatment=277.38, SDTreatment=691.51; p’s >.487). An independent t-test was run on the net
MNone=214.02,SDNone=623.23). This means that all savings balance, comparing the No insight condi-
four insights presenting saving as small recurring tion to all four treatment conditions, and showed
steps motivated customers to transfer more money a significant effect (t(130327)= 3.518, p =.010). This
into their savings accounts. Thus, general refram- result indicates that an insight presenting sav-
ing as a concept worked, but the level of temporal ings as a granular activity (MTreatment=44.11, SDTreat-
granularity (i.e. size of the saving step) and the dif- ment
=834.53) had a significant positive impact on the

Behavioral Economics Guide 2021 54

Laura Straeter & Maria Ferreira How Targeted Messaging Can Help Customers With Low Savings Build a Buffer

Figure 3: Amount transferred to savings account per condition in Experiment 2. Note: Error bars represent ± 1
standard error.

Figure 4: Net savings balance per condition in Experiment 2. Note: Error bars represent ± 1 standard error.

net savings balance of our target customers over saving as bite-sized steps were equally effective in
the two-week period (MNone=19.01, SDNone=1327.21). motivating customers to save more money.
Similar to the results for the amount transferred
to savings accounts, all four insights presenting

Behavioral Economics Guide 2021 55

Laura Straeter & Maria Ferreira How Targeted Messaging Can Help Customers With Low Savings Build a Buffer

Discussion The findings of both experiments are in line with

previous research (e.g. Abel et al., 2016; Hodkinson,
The results of experiment 2 indicate that refram- 2019; Hershfield, 2020; Colby & Clapman, 2013). Our
ing saving as small recurring steps has a positive in- first experiment adds to the literature, as it shows
fluence on both the amount transferred into savings that the fear of missing out can be valuable beyond
accounts and the net savings balances of ING cus- its usual implementation. Although it is often used
tomers with low savings. Compared to receiving no to increase sales, our findings also show that ac-
insight, customers who received one of the insights tivating the fear of missing out could help people
transferred, on average, €68 more to their savings transfer more money to their savings account and
account(s) and bolstered their net savings with €25 build a buffer instead. Nonetheless, FOMO insights
more. Although we aimed to unveil what level of need to be used with caution, as people were also
temporal granularity would lead to most savings, found to withdraw more money from their savings
in terms of both money transferred to savings ac- accounts when exposed to them. More research is
counts and increase in net savings balance, all the needed to understand this outcome better.
insights showed to be equally effective for different The second experiment on temporal reframing is
levels of temporal granularity (i.e. saving weekly or particularly valuable to the existing literature, as it
monthly) and ways to represent the saving goal (i.e. measures real savings and not the intention to save.
€250 or a TV). We found that breaking down savings into bite-size
portions was effective in motivating people to save
General Discussion more, compared to people not receiving the insight.
A neutral insight was not included, which could have
We presented two experiments that aimed to help determined the added value of temporal reframing
people with low savings build up a buffer. Targeting over an insight by itself. In addition, we also need
the right people and exposing them to actionable to acknowledge that we did not find the expected
insights based on behavioural techniques in a nat- differences between the levels of temporal gran-
ural environment, namely their banking app, was ularity. Our target group, namely people with less
effective. Specifically, we found that activating the than €2,000 savings, might be more susceptible to
fear of missing out and presenting saving as small smaller steps (e.g. saving per day). Future research
recurring steps motivated customers to transfer would need to investigate these questions further.
more money to their savings account(s), and that Our research findings can be directly applied by
the latter also led to growth in net savings balance. ING. Not only did we help customers who were part
Whereas it is insightful to know what insights of our experimental selections, but the insights
have a direct effect on behaviour (i.e. do customers can easily be scaled to other customers with little
take action?), the mere value of such insights for savings in the Netherlands – and potentially oth-
customers lies in the change they bring in terms er countries. Cultural differences would need to be
of net savings balance over a longer period of time. taken into account. Our research also has implica-
One insight will not have the power to change sav- tions for other financial institutions, since it sug-
ings behaviour consistently, but a range of targeted gests that actionable insights based on behavioural
insights together will more likely do so. Therefore, concepts in general, and two techniques specifical-
besides the short-term contributions of a single ly, can be fruitful in driving saving behaviour. Con-
insight to the amount transferred and changes in sequently, we believe it is important that financial
the net savings balance measured in our research, institutions join ING in the goal to improve people’s
we believe it is crucial to analyse the increase in net financial health and help those with low savings to
savings balance over a longer period of time (e.g. a build a buffer.
few months or a year) to understand the real value
of actionable insights based on behavioural tech-
niques for customers with low savings.

Behavioral Economics Guide 2021 56

Laura Straeter & Maria Ferreira How Targeted Messaging Can Help Customers With Low Savings Build a Buffer

The Authors goals, and reference points. Judgment and

Decision Making, 8, 16-24.
Laura Straeter is a behavioural scientist in ING’s Fogg, B. (2009). A behavior model for persua-
Global Sustainability department, working with sive design. Proceedings of the 4th International
different divisions across the retail organisation to Conference on Persuasive Technology, 40.
apply practically an understanding of behaviour- Frederick, S., Loewenstein, G., & O’Donoghue, T.
al science to promote financial health. She is also (2002). Time discounting and time preference: A
involved in the Think Forward Initiative, fuelling critical review. Journal of Economic Literature, 40,
innovations and start-ups with consumer research 351-401.
insights. She holds a PhD in Marketing Manage- Goldstein, D. G., Hershfield, H. E., Benartzi, S.
ment from the Rotterdam School of Management, (2016). The illusion of wealth and its reversal.
Erasmus University. Laura received her bachelor’s Journal of Marketing Research, 53, 804-813.
degree in Social Psychology and her research mas- Gourville, J. T. (1998). Pennies-A-Day: The effect of
ter’s degree in Social and Behavioural Sciences from temporal reframing on transaction evaluation.
Tilburg University. Before joining ING in January Journal of Consumer Research, 24, 395-408.
2018, she was a lecturer at Maastricht University Hershfield, H. E., Shu, S., & Benartzi, S. (2020).
and worked for marketing research consultancy Temporal reframing and participation in a sav-
firms. ings program: A field experiment. Marketing
Science, 39, 1033-1201.
Maria Ferreira is a senior economist and research Hodkinson, C. (2019). ‘Fear of Missing Out’ (FOMO)
professional at ING, where she works as part of the marketing appeals: A conceptual model. Journal
global sustainability department’s financial health of Marketing Communications, 25, 65-88.
team. Her role sits across the retail organisation Intrum (2020). European consumer payment report
and the Think Forward Initiative to deliver on the 2020. https://www.intrum.com/media/8047/
ING purpose of empowering people to stay ahead ecpr-2020-special-edition-white-paper.pdf.
in life and in business, and to make better financial Maslow, A. H. (1943). A theory of human motivation.
decisions. Maria holds two master’s degrees in Eco- Psychological Review, 50, 370-396.
nomics and Social Sciences, and a PhD in Economics Michie, S., Van Stralen, M. M., & West, R. (2011). The
from Maastricht University. Before joining ING, she behaviour change wheel: A new method for char-
worked in academia in the fields of Applied Micro- acterising and designing behaviour change in-
economics, Labour Economics and Development terventions. Implementation Science, 6. https://
Economics. doi.org/10.1186/1748-5908-6-42.
Przybylski, A. K., Murayama, K., DeHaan, C. R.,
References Gladwell, V. (2013). Motivational, emotional,
and behavioral correlates of fear of missing out.
Abel, J. P., Buff, C. L., & Burr, S. A. (2016). Social Computers in Human Behaviour, 29, 1841-1848.
media and the fear of missing out: Scale devel- Riordan, B., Flett, J., Hunter, J., Scarf, D., & Conner,
opment and assessment. Journal of Business & T. S. (2015). Fear of Missing Out (FoMO): The
Economics Research, 14, 33-44. relationship between FoMO, alcohol use, and al-
Bambauer-Sachse, S., & Grewal, D. (2011). Temporal cohol-related consequences in college students.
reframing of prices: When is it beneficial? Journal Journal of Psychiatry and Brian Functions, 2. http://
of Retailing, 87, 156-165. dx.doi.org/10.7243/2055-3447-2-9.
Baumeister, R., & Leary, M. (1995). The need to be- Xie, X., Wang, Y., Wang, P., & Zhao F. (2018). Basic
long: Desire for interpersonal attachments as a psychological needs satisfaction and fear of
fundamental human motivation. Psychological missing out: Friend support moderated the me-
Bulletin, 117, 497-529. diating effect of individual relative deprivation.
Colby, H., & Chapman, G. B. (2013). Savings, sub- Psychiatry Research, 268, 223-228.

Behavioral Economics Guide 2021 57

 Increasing Retirement Savings in South Africa
Natan Sklair*
Old Mutual

Upon leaving one’s employer, many South Africans have one-off access to withdraw funds from their re-
tirement pot. Despite being taxed, many use this opportunity to use these funds for immediate consump-
tion, but they fail to plan appropriately and instead withdraw unnecessarily high values from their fund.
Often, the chosen value is the result of naïve diversification. In order to increase the value of funds preserved
for retirement, Old Mutual’s Behavioural Economics Team designed a Randomised Control Trial (n=6207)
to test the effectiveness of a behaviourally-informed contact centre conversation. The conversation used
behavioural principles, such as framing and anchoring, to change members’ decisions upon leaving their
employer, from three options (full withdrawal, partial withdrawal and full preservation) to two options:
full preservation or electing to take out a percentage of the fund, from 0% to 100%, as a withdrawal for an
emergency cash need – while providing anchors of 5% and 10% as examples to the members.. A key result
saw a 33%-point increase in the average proportion of funds preserved.

Introduction Upon investigating trends relating to Old Mutu-

al retirement fund members2, it was observed that
When South Africans leave their employers before many withdraw approximately 50% of their avail-
the age of 55, they have a one-off opportunity to able funds while keeping the rest in a preservation
withdraw a lump-sum from their employer-linked fund. One key hypothesis was that they were sim-
retirement fund . Unfortunately, many fail to plan
1
plifying their decision by employing naïve diversi-
properly, and so they choose to withdraw unneces- fication – a mental heuristic whereby individuals
sarily high amounts. While access to these funds can split their funds equally into the number of options
be useful for meeting individual emergency needs, available to them (in this case, preserve and with-
drawing down too much can negatively affect their draw) (Gigerenzer & Gaissmaier, 2011).
wealth when they actually retire. To reduce the value of funds withdrawn before
Old Mutual provides retirement fund options for retirement, Old Mutual’s Behavioural Economics
employees of South African businesses. These em- Team (BET) designed a randomised control trial
ployees are known as “members” of the fund. There (RCT) (n=6207) to test whether behavioural in-
are many ways in which a fund can be structured; terventions would help members preserve their
however, the important point for this case study is retirement funds for the future. By properly under-
that when a member leaves their employer, there is standing the South African context, the individuals
a value accessible to the member from which they affected, and the structure of the organisation, the
can withdraw for present consumption. team was able to design a contact centre conversa-
tion for retirement fund professionals that would
1 Different funds have different rules, and as of March help Old Mutual members safeguard their retire-
2021, the rules have been amended slightly; however, for ment savings.
the purposes of this reading, there is always a maximum A key strategy in the intervention included chang-
withdrawal value available that can be withdrawn from a
fund. 2 Members are individuals who belong to the retirement
fund. This can include either their employer standalone
* [email protected]. fund or Old Mutual’s umbrella fund.

Behavioral Economics Guide 2021 58

Natan Sklair Increasing Retirement Savings in South Africa

ing the decision frame in a conversation with mem- ered around zero, but it has mostly been negative.
bers, from choosing between three options (full In a country where the majority struggle to build
withdrawal, partial withdrawal or full preservation) wealth, and access to capital is limited, opportu-
to two options: full preservation or electing to take nities to increase savings rates are taken seriously.
out a percentage of the fund, from 0% to 100%, as These low savings rates, if not improved, can nega-
a withdrawal for an emergency cash need – while tively affect economic growth and place a financial
providing anchors of 5% and 10% as examples to burden on the government to provide for the elderly
the members. population (Ting & Kollamparambil, 2015).
The key results from the intervention saw: Importantly, under apartheid, almost 90% of the
population did not have access to resources such as
• A 33%-point increase in the average proportion of jobs, capital and land. Therefore, for many families,
funds preserved for retirement (with 92% of the the current generation is the first to have access to
fund value preserved in the pilot test group) opportunities that can build wealth.
• A 20% increase in the volume of funds preserved More recently, South Africa has seen an increase
• A 10% increase in the average NPS score3 in what has been dubbed the “Sandwich Gener-
• A successful rollout of the intervention across the ation” (Old Mutual, 2019). This term refers to in-
organisation dividuals who support both their younger family
members (children, nieces, nephews, etc.), as well
Old Mutual’s Behavioural Economics Team as older family members (parents, grandparents,
(BET) aunts, uncles). With South Africa being one of the
most consistently unequal countries in the world
The BET at Old Mutual was set up as an internal (Stats SA, 2019), initiatives that aim to improve na-
consultancy to assist the business with key chal- tional savings are particularly important.
lenges and to create opportunities for its custom- Therefore, retirement savings play a key role in
ers, in order to improve their financial wellbeing. the country, as the funds are more difficult to access
The team has one goal, namely to apply behav- before retirement – especially when individuals fail
ioural economics across the organisation. In this to save via other financial vehicles4. However, South
way, it has the freedom to explore the entire busi- Africans are allowed to withdraw from their em-
ness, without being siloed in one area. With access ployer retirement fund if, before the age of 55, they
to 30,000 Old Mutual employees, and over 11 million leave their employer due to resignation, retrench-
customers across 14 countries, the team is well- ment or dismissal5.
placed to design interventions that make a differ- For some individuals, access to these funds offers
ence at scale. welcome relief – especially if they are entering a
period of income uncertainty and require addition-
South Africa’s Savings Problem
4 For the employed population, pension and provident
South Africa has one of the lowest household sav- fund options are commonly provided through their em-
ings rates among developing countries, ranging be- ployer. These funds can be structured in different ways
tween -2.31% and 0.2% over the course of 2010 and depending on the employer, but are most typically a com-
2019 (OECD, 2020). Effectively, the metric has hov- pulsory plan where the employee selects a percentage of
their pre-tax monthly salary toward their retirement sav-
3 NPS stands for “Net Promotor Score.” Customers are ings.
asked, “How likely are you to recommend us to a friend 5 Until 1 March 2021, South Africans could withdraw the
or colleague?” on a zero-to-ten scale. The NPS is the full value of these funds, and an appropriate tax would be
percentage of customers who are promoters (those who applied. However, since 1 March 2021, the amount acces-
scored 9 or 10) minus the percentage who are detractors sible for withdrawal by individuals has become limited
(those who scored 0 to 6). (Old Mutual, 2021).

Behavioral Economics Guide 2021 59

Natan Sklair Increasing Retirement Savings in South Africa

al funds. However, our findings suggest that many tion fund (the member defaults to full preserva-
individuals fail to budget properly for their future tion if no decision is made)
during this process, and they end up withdrawing
unnecessarily high amounts from their retirement Once the team understood the environment, the
fund. qualitative and quantitative data aligned to tell a
clear story. That is, members were taking out mon-
Understanding the Affected Individuals ey because:

Essential to the success of the project was the ap- • They had no other resources to provide them sup-
proach taken by the team, investing time in under- port
standing the affected members, their journeys, and • They were unaware of their alternative options for
consolidating the relevant behavioural economics emergency cash needs
tools that might apply to their decisions. • They were not considering exactly how much they
To understand behavioural barriers, the team needed for immediate consumption, leading them
employed a series of methods, each with their own to withdraw an amount that was larger than nec-
objective: essary

1. Mapping out the full member decision journey, Regardless of personal context, members going
from the time the member joins their employer, through an exit process may be experiencing a great
to the time they leave their employer deal of uncertainty or financial stress despite some
2. Reviewing all existing communications sent out of them moving to new employment. Among the
to Old Mutual fund members three broader reasons for withdrawing from one’s
3. Listening to live and recorded calls retirement fund, a number of behavioural factors
4. Conducting interviews with stakeholders across were also identified as barriers to better savings
the affected areas of Old Mutual decisions.
5. Mapping out the different areas of the business
affected The Timing of the Decision Was Critical
6. Analysis of the data to understand the core prob-
lem and possible target area Members effectively had one major decision
7. Customer feedback point regarding their retirement fund: a phone call
confirming their final decision on their retirement
By creating an in-depth decision map for each plan (point 3 in Figure 1, below). Despite being able
type of member exit (resignation, retrenchment to complete a form prior to a conversation with a
and dismissal), the team was able to map discrete retirement fund professional, many members are
choices along the decision path to making sound only confronted with their need to make a decision
financial decisions regarding savings. Combined on receipt of a call conducted by a retirement fund
with a clear understanding of the decision context, professional.
this allowed the team to expose gaps in the existing
process/approach. Using the 1/N Heuristic to Plan for Retirement
Importantly, members effectively have three op-
tions when it comes to their retirement fund deci- Looking at the data for members who had partially
sion when they leave their employer. They can: preserved their funds, the team noticed that people
were, on average, withdrawing just under 50% of
1. Withdraw the full value of their fund their fund value. This (almost) equal split led to the
2. Withdraw part of their fund, and preserve the rest hypothesis that, upon needing to make a decision,
in a preservation fund many members were subject to naïve diversifica-
3. Preserve the full value of their fund in a preserva- tion, i.e. splitting their withdrawal and preservation

Behavioral Economics Guide 2021 60

Natan Sklair Increasing Retirement Savings in South Africa

in half, because it simplified the decision. This is servation is a classic case of hyperbolic discounting,
consistent with the theory put forward by Giger- whereby individuals confronted with the choice
enzer & Gaissmaier (2011), also known as the “1/N hold a preference for a smaller level of consumption
heuristic,” whereby individuals tend to assign equal in the present period while trading off a higher level
weight to the alternatives available to them. of consumption in the future (Laibson, 1997).
Therefore, it appeared that members were per-
Mental Accounting to Justify Early Withdrawals: A forming some mental accounting by thinking of
Clear Case of Hyperbolic Discounting access to their funds as a pool of income relative
to their present consumption needs –rather than
Despite being subject to strong tax penalties for calculating their needs accurately (Thaler, 1999).
early withdrawals (up to 36% of the fund value), Members were allocating some of their retirement
many members were willing to trade off the fu- funds to the here and now rather than the intended
ture value of their fund for present consumption. future purpose of income at retirement – despite
Our research showed that members often thought having access to the calculations that would allow
of their fund value as the post-tax fund value, i.e. them to view their potential future fund value. This
the amount available to spend, and they struggled is congruent with the behavioural life-cycle hy-
to think about themselves at retirement. This ob- pothesis, whereby wealth is assumed to be divided

Figure 1: Key touchpoints in the member decision journey.

into three mental accounts: current income, current proving their future retirement savings. Having
assets and future income (Shefrin & Thaler, 1988). identified the conversation with the retirement
professional as the primary member decision point,
Lack of Salient Alternatives the team designed a conversation blueprint for
retirement fund professionals. In addition to the
As some members were entering a period of fi- conversation blueprint, the team also changed the
nancial uncertainty, especially those who had been form to reflect the options presented in the behav-
dismissed or retrenched, an urgent need for cash iourally-informed conversation; however, the form
could have been justified. However, upon investi- was excluded from the trial.
gation, many members were unaware of the alter- Figure 2, below, summarises the main behaviour-
native financial solutions available to them when al economics principles used in the conversation
withdrawing from their retirement fund that could blueprint, in order to address the challenges iden-
help them with their immediate financial needs. tified at the outset of the project.
Given that people tend to revert to simple heuris-
The Intervention tics, or mental shortcuts, when making decisions in
uncertain situations (Tversky & Kahneman, 1974),
To improve member decision-making, the team the behavioural principles used in the intervention
designed a behavioural intervention aimed at im- simplified the decision-making process by either

Behavioral Economics Guide 2021 61

Natan Sklair Increasing Retirement Savings in South Africa

Figure 2: Behavioural principle applied to the conversation blueprint.

accounting for or taking advantage of these heuris- fund professionals were able to provide members
tics. Expanding on Figure 2, below is a description with a list of alternative solutions that could help
of how the key behavioural principles were used in them meet their immediate needs.
the intervention.
In addition to the above principles, the team em-
• Framing: In the original conversation/decision, ployed other methods in the conversation, includ-
members had three options (fully preserve, par- ing changing the order in which the options were
tially preserve, full withdrawal). The new, behav- presented to the member (presenting full preserva-
iourally-informed, conversation provided only tion as the first option), simplifying and shortening
two decision options to the member: full preser- the wording in the conversation, having the more
vation or elect to take a percentage of the fund, important conversation upfront and using social
from 0% to 100%, as a withdrawal. By present- norms as a guide.
ing the option slightly differently, members were
encouraged to think about each percentage of the Implementation of the Intervention
fund that they would need to withdraw.
• Availability heuristic: Knowing that individuals There were three key components to the inter-
tend to make judgements based on how easy an vention:
example comes to mind (Botha et al., 2014), it was
clear that members would often struggle to think 1. Scripting a behavioural economics-informed
of an example for actually using the withdrawn conversation embedded into the system
money despite intending to make the withdraw- 2. Designing conversation cue cards, to increase the
al anyway. The updated conversation provided a salience of key conversation points
clear reason for the withdrawal portion, that is, 3. Upskilling the retirement fund professionals
“an emergency cash need.”
• Anchoring: The conversation provided 5% or 10% To ensure sustained development in the con-
as an example of how much to withdraw for emer- tact-centre environment, the team shifted away
gency cash needs. Combined with framing, the from the traditional approach of simply training
hypothesis was that the anchoring effect would agents on the new script. Instead, they focused on
counteract the naïve diversification observed in encouraging the agents to adopt a new, behav-
member behaviour. iourally-informed way of thinking – ensuring the
• Salience: As members may not realise they have agents internalised the behavioural principles.
other options if they urgently need funds, the To this end, the training process, which was con-
script included questions regarding the main fi- ducted using a combination of prompt cards and a
nancial concern. Once established, the retirement behavioural toolbox, involved group sessions, role-

Behavioral Economics Guide 2021 62

Natan Sklair Increasing Retirement Savings in South Africa

play, dry runs with the management team, live and organisation and pursue the ongoing expansion of
recorded call listening (with feedback) and ongoing their role and influence.
one-on-one coaching, which was finally handed
over to the retirement fund professional manage- Reflections on the Intervention
ment team.
Since conducting this study, as of 1st March 2021,
Testing the Intervention & Key Results the South African government has changed the
percentage of funds available for withdrawal upon
To test the effectiveness of the intervention, the leaving one’s employer. While this will help South
team set up a randomised control trial (n=6207). The Africans maintain a higher value of their funds for
RCT consisted of splitting the sample of members income at retirement, they will still be subject to the
into two groups: the first would receive a behav- same challenge, should they leave their employer.
iourally-informed conversation, while the second In other words, despite the percentage of funds
received the existing, business-as-usual process. from which one can withdraw being smaller, people
This was achieved by only assigning one group of will still be subject to the same temptations and bi-
retirement fund professionals the additional train- ases when faced with a similar decision.
ing and material that included the behavioural eco- Therefore, there were two key lessons from the
nomics principles for the new conversation. Both intervention. First, individuals’ environment and
groups of agents had similar levels of experience context at the point at which a decision is made are
and previous training. important contributors to retirement savings out-
The pilot ran for six weeks and captured 6,207 comes. That is, despite financial planning, or a lack
customer interactions. Using a difference-in-dif- thereof, the conversation with members appeared
ference approach, the final results saw: to be highly influential in one’s retirement fund
decision-making. This means that measures ought
• A 33% point increase in the average proportion of to be put in place to ensure that these conversations
funds preserved for retirement (with 92% of the account for the mental biases and heuristics that
fund value preserved in the pilot test group) fund members may be experiencing when the de-
• A 20% increase in the volume of funds preserved cision is made.
• A 10% increase in the average NPS score Firms responsible for conversing with members
with regards to their retirement fund upon leaving
One result that stands out most significantly is their employer are encouraged to employ similar
the change in the average proportion of funds pre- behavioural principles in their conversations or
served. While the team did not test each behavioural relevant decision points.
technique individually, there is a strong argument Second, part of the success of the intervention is
to be made that members were previously using attributed to the training provided to the retirement
the 1/N heuristic to make their partial withdrawal fund professionals – the ones taking the calls with
decisions. Providing members with anchor points, the members. The key lesson here was develop-
and prompting them to think about each percent- ing a training programme that would allow them
age point withdrawal, allowed them to be more to internalise the behavioural principles – rather
conscious about their thinking and make more fu- than simply being exposed to them. By immersing
ture-orientated decisions. the professionals in behavioural economics, and
Following the success of the pilot, the interven- providing them with consistent feedback, they
tion was then rolled out to the rest of the relevant were able to account for any sudden or unexpected
business unit. Importantly, the pilot demonstrated changes to the flow of the conversation with mem-
a convincing success for the business. From a be- bers.
havioural capability-building perspective, this al- Ultimately, the intervention conducted by Old
lowed the team to increase its legitimacy within the Mutual’s Behavioural Economics Team highlights

Behavioral Economics Guide 2021 63

Natan Sklair Increasing Retirement Savings in South Africa

the importance of addressing the target or prob- OMC_Annuitisation_note_Q&A_17_

lem behaviour at the points which matter most to Feb_2021.pdf.
the final decision, together with the people who are Shefrin, H. H., & Thaler, R. H. (1988). The behavioral
most influential in interacting with the affected in- life-cycle hypothesis. Economic Inquiry, 26, 609-
dividuals. 643.
Statistics South Africa – Stats SA. (2019). Inequality
The Author trends in South Africa: A multidimension-
al diagnostic of inequality. http://www.stats-
Natan Sklair leads Old Mutual’s Behavioural Eco- sa.gov.za/publications/Report-03-10-19/
nomics Team. Since 2016, he has been applying Report-03-10-192017.pdf.
behavioural economics across both the private and Thaler, R. H. (1999). Mental accounting matters.
public sectors in Africa and the UK. Since joining Old Journal of Behavioral Decision Making, 12(3), 183-
Mutual in 2017, Natan has helped grow a team that 206.
provides behavioural economics consulting servic- Ting, L., & Kollamparambil, U. (2015). Nature and
es to the entire group. The team delivers projects determinants of household retirement savings
on external, customer, internal and employee out- behaviour in South Africa. Development Southern
comes. Natan has a Masters in Applied Economics Africa, 32(6), 675-696.
from the University of Cape Town. Tversky, A., & Kahneman, D. (1974). Judgment
under uncertainty: Heuristics and biases. Science,
References 185(4157), 1124-1131.

Botha, M., Rossini, L., Geach, W., Goodall, B., Du

Preez, L., & Rabenowitz, P. (2014). The South
African financial planning handbook 2014.
LexisNexis.
Gigerenzer, G., & Gaissmaier, W. (2011). Heuristic
decision making. Annual Review of Psychology,
62(1), 451-482
Laibson, D. (1997). Golden eggs and hyperbolic
discounting. Quarterly Journal of Economics, 112,
443-477.
Old Mutual. (2019). Savings and investment mon-
itor. https://www.oldmutual.co.za/docs/
default-source/personal-solutions/finan-
cial-planning/savings-and-monitor/om-
sim-2019-general-findings-only.pdf.
Old Mutual. (2021). Retirement re-
form. https://www.oldmutual.co.za/
v3/assets/blt0554f48052bb4620/blt-
d2aadec9399e2140/60335362919f-
9b52ae2fe0d6/OMC_Annuitisation_
note_Q&A_17_Feb_2021.pdf.
Organisation of Economic Co-operation and
Development – OECD. (2020). Household
savings. https://www.oldmutual.co.za/v3/
assets/blt0554f48052bb4620/bltd2aade-
c9399e2140/60335362919f9b52ae2fe0d6/

Behavioral Economics Guide 2021 64

 Nudging to Improve the Timely Payment of
Loans: Results of a Natural Field Experiment
Radhika Bhalla,* Shruthi Ramakrishnan, Prasad G and Nimit Marwaha
Quadrangle Consulting Services

At the commencement of this study, ~31.5% of customers of DMI Finance, an Indian B2B non-banking
financial company (NBFC), had failed to repay their loans on time. We used insights from the literature
on biases, heuristics and nudge theory to positively influence customers’ payment behaviour. We created
persona-linked, personalised communication to increase awareness about payment processes, to make
call-to-action links easily accessible and to provide education about the consequences of delays. The inter-
vention led to an increase of 7% - 11% in the timely repayment of loans in the experimental groups. It also
reduced delinquency on average by 2 days a month.

Introduction mencement of this study, despite DMI Finance’s

robust customer communication program, which
India’s per capita income is estimated to increase includes text messages (SMS), e-mails and on-call
to US$ 3,500 by 2024 (a FICCI NBFC Committee Re- assistance via the company’s service centre, ~31.5%
port, 2020). Movement of the population to high- of customers had failed to repay their loans on time,
er-income categories, a rise in the working age thereby endangering their own credit scores and in
population to ~1200 million by 2031, the rampant turn their chances of qualifying for further loans.
spread of mobile usage and technology will drive
both consumption in the economy and associated Our Approach
credit need. While traditional banks certainly con-
sider risk in servicing the ~147 million creditwor- Human behaviour is partially driven by reasoned,
thy customers currently under the age of 24 years, deliberate cognition and partially by prior experi-
DMI Finance has designed financial products for ences, inherent biases and heuristics. This study
this underserved segment of society. With better explores the use of nudges to counter some of these
data utilisation and unique risk underwriting, it has biases and to influence customers’ behaviours, us-
harnessed the opportunity to provide low-value, ing our context-agnostic and experiment-led 4-D
unsecured loans to these customers. methodology.
DMI Finance acquires customers through multiple
digital partner apps. Once the customer is deemed D1 – Diagnose
eligible for a loan, disbursement communication is
led by DMI Finance, as is management of the cus- The diagnosis phase included the collection of
tomer’s journey and the collection of debt. Given information around default patterns through data
that average unsecured loan amounts are as low as reviews and meetings with internal teams, from
INR 5000 (~ US$ 68), the cost involved in servicing customer success, credits, collections and IT, to
customers is kept low by leveraging the digital me- data science. Qualitative insights from interviews
dium to manage customers’ journeys. At the com- were cross-referenced against data taken from DMI
Finance’s CRM platform, customer service centre
* Corresponding author: records and website analytics, in order to review
[email protected]. demographic indicators (education, age, gender,

Behavioral Economics Guide 2021 65

Radhika Bhalla et al. Nudging to Improve the Timely Payment of Loans

Figure 1: 4-D Methodology.

history of borrowing, location, etc.), patterns of It was hypothesised that:

payment delays, the nature and frequency of que-
ries and complaints raised by customers, etc. A cus- 1. Payment delays and defaults can be reduced by
tomer journey map (CJM) was created to highlight creating awareness about the payment process,
critical engagement points between the customer ensuring accessibility to call-to-action payment
and DMI Finance. A combined review of the CJM and links and blending education on the consequences
the data led to the discovery of critical gaps in cus- of missed payment.
tomer communication: 2. Timely payment can be encouraged with more
personalised communication to different cus-
• Missing & Complex Communication: Commu- tomer segments, thus addressing specific barriers
nication was irregular and not present at critical to repayment.
journey points. Messaging via SMS and E-mail 3. Cost to service via the customer service centre
was complex, and it used mature banking termi- can be reduced by refining and promoting simple,
nology that was not congruent with the financial digital self-service channels.
experience of the customer profiles.
• Fragmented & Inconsistent Communication: D2 – Discover Persona
Different digital platforms did not provide con-
sistent information, thus adding to the call vol- Effectively communicating with customers about
ume burden at the customer service centre. repayment, saving and attitude to financial plan-
• Missing Education: Customer education material ning required unearthing the drivers of borrowing
was missing from the journey points. debt, repaying loans and the relationship between
• Misaligned Positioning: Customers were on- lifestyle needs and money. A literature review (Fig-
boarded via a partner app and were unaware of ure 2) revealed the influence of financial literacy
their relationship with DMI Finance. Information (FL) and lifestyle needs (LN) on customer repay-
from both sources was often found to be overlap- ment behaviour.
ping and conflicting. We identified the specific pillars of FL and LN
from the literature to aid the study. Interactions

Behavioral Economics Guide 2021 66

Radhika Bhalla et al. Nudging to Improve the Timely Payment of Loans

Figure 2: Literature reviewed on financial literacy and lifestyle needs.

Figure 3: Sample illustration of one of the five personas.

Behavioral Economics Guide 2021 67

Radhika Bhalla et al. Nudging to Improve the Timely Payment of Loans

between these pillars create five distinct customer Design outcomes

personas – personas A, B, C, D and E. For each per-
sona, possible barriers to repayment were averred. Segmentation of customers, using the profiler form
Level 1 barriers emerged due to a lack of sufficient
and clear information about interest rates, late fees, The segmentation of customers into the five
the consequences of non-payment, etc. These were personas was achieved via a profiler form that was
not distinct for different personas. Level 2 barriers tested on 7,714 customers in two phases. The form
surfaced due to distinct interactions between cus- consists of 11 scorable multiple choice questions
tomers’ FL and LN, creating unique payment and mapped to FL and LN. Based on the scores, a cus-
default patterns that emerged from their unique tomer is classified into high, medium or low seg-
motivations.D3 – Design ments in relation to FL and LN. Together, the cate-
We designed an integrated customer communi- gories culminate in the five personas. If customers
cation programme, using the three critical levers do not answer all or none of the questions, they are
(Figure 4) to nudge customers to pay on time and assigned persona E. Once customers are segmented
leverage self-service digital channels by: into one of the five personas, a tailored SMS series
is triggered.
1. Creating a profiler form for the segmentation of
customers, in order to personalise communica- Nudge-based text messages (SMS)
tion.
2. Deploying nudge-based text messages (SMS) for SMS forms an important tool for customer com-
consequence education, payment reminders, ac- munication, especially in India, where there are
cessible links to self-service digital channels, etc. ~713 million cellphone subscribers. SMS content
3. Refining communication and user design across was repurposed as a nudge and linked to specific
the websites, portal, e-mail and education videos, journey points. It included reminders to address
to promote self-service. availability bias, deadlines to counter informa-
tion avoidance, default payment options to tackle

Figure 4: The three critical levers of design.

Behavioral Economics Guide 2021 68

Radhika Bhalla et al. Nudging to Improve the Timely Payment of Loans

Figure 5: Nudge-based SMS.

choice overload and direct payment links to counter (since nearly 27% of the call centre volume was
procrastination and to reduce switching costs. The generated from assumed delays in system up-
nudges were designed to either reinforce timely dates)
repayment behaviour through the use of positive
strokes or undermine defaulting behaviour through Website and portal. All information pertaining to
the use of social proof and expression of disap- payment and self-service options were standard-
pointment. SMS design was constrained within the ised across the communication channels. Structural
160-character limit.Refining communication and and navigational changes were introduced that in-
user design across digital channels cluded rewording of the FAQ section to reflect cus-
E-mails. Nearly 73% of the customers called the tomers’ vocabulary (limiting friction for customers
service centre more than once to seek resolution when they use a custom search), the clustering of
to a query that was previously resolved via e-mail queries based on their similarity, placing frequently
communication. We rephrased contractual/tech- asked questions at the top of the page and arrang-
nical information used in the e-mails and redevel- ing the navigation buttons to the left, since research
oped the e-mail template to enhance its visual ap- suggests viewers spend 80% of their time viewing
peal and simplicity, to ensure the following critical the left side of the screen (Therese, 2017).
information was addressed by the customers: Education videos. Five videos were created to en-
courage customers to pay on time and adopt the
• Loan ID self-service digital channels. Videos were inserted
• Contact information (sequenced such that the at various points in the customers’ journey, namely
portal was prioritised over the customer service on-boarding, first EMI payment, subsequent pay-
centre information) ment and once when all EMIs are paid.
• Call-to-action link to the FAQs page on the web-
site
• Standard operating time taken to update records

Behavioral Economics Guide 2021 69

Radhika Bhalla et al. Nudging to Improve the Timely Payment of Loans

D4 – Deploy tion of repaying the loan; yet, some do not follow

through on their intentions. This intention-action
Pilot and results gap, or the discrepancy between what people claim
they will do and what they actually do, provides a
It is reasonable to believe that when customers fertile window for behavioural change.
contact lending partners, they do so with the inten- The study tested the efficacy of using various

Figure 6: Sample refined e-mail template.

Figure 7: Customer education videos.

Behavioral Economics Guide 2021 70

Radhika Bhalla et al. Nudging to Improve the Timely Payment of Loans

Table 1: Pilot – customer segments.

nudges as almost cost-free methods (Jensen et al., • C – when customers received messages in the
2018) for improving timely loan repayment and en- presence of a penalty
couraging the use of digital channels for self-ser-
vice, without attaching a substantial financial Segment A
reward or penalty to the process. Pilots were con-
ducted on four segments of customers to test two Overall, 772 (11%) more customers paid on time in
hypotheses, wherein control groups received regu- the experimental group than in the control group.
lar SMSs, and experimental groups received newly Significantly more customers paid on time in the
designed nudge-based Persona E SMSs (since a experimental group as compared to the control
majority of customers in the profiler form testing group in month 1 (31% vs 23%, Ӽ2 [1, N = 4535] =
phase were segmented into Persona E). Hypothesis 38.12, p < .001), month 2 (32% vs 25%, Ӽ2 [1, N =
1– The new SMS series will nudge more customers 4535] = 29.27, p < .001) and month 3 (32% vs 25%,
in the experimental group to pay on or before time, Ӽ2 [1, N = 4535] = 31.89, p < .001). Hence, the results
compared to customers in the control group. support hypothesis 1.
To test hypothesis 1 across segments A, B & C, a
chi-square test of independence was performed in Segment B
both groups to examine the relationship between
making repayments and receiving messages. Overall, 87 (7%) more customers paid on time in
Nudge-based SMS effectiveness was evaluated the experimental group than in the control group.
under three circumstances: This difference, however, was not found to be sig-
nificant between the experimental group and the
• A – when customers had a previous default histo- control group across month 1 (47% vs 41%, Ӽ2 [1, N
ry and had experienced “consequences of default- = 838] = 3.27, p = .07), month 2 (53% vs 48%, Ӽ2 [1,
ing” N = 838] = 1.72, p = .19) and month 3 (58% vs 51%,
• B – when customers were newly on-boarded and Ӽ2 [1, N = 838] = 3.25, p = 0.07). Hence, the results do
thus had no default history not support hypothesis 1.

Behavioral Economics Guide 2021 71

Radhika Bhalla et al. Nudging to Improve the Timely Payment of Loans

Segment C the control group than in the experimental group.

Significantly more customers paid on time in the
Overall, 17 (<1%) more customers paid on time in experimental group as compared to the control

Table 2: Pilot results for Segments A, B & C.

Behavioral Economics Guide 2021 72

Radhika Bhalla et al. Nudging to Improve the Timely Payment of Loans

group in month 1 (29% vs 27%, Ӽ2 [1, N = 9968] = lected for new messages were D-5, D-3, D-1 day(s)
6.78, p = .009). However, no significant differences before the due date (D) and D+4, D+8 days after the
were found between the experimental group and the due date (D) (highlighted by blue circles in Figure
control group across month 2 (36% vs 36%, Ӽ2 [1, N 8). Greater numbers of customers were expected to
= 9968] = 0.04, p = .83) and month 3 (42% vs 41%, pay on the day the messages were sent and on the
Ӽ2 [1, N = 9968] = 0.14, p = 0.69). Hence, the results following day. Such a trend was observed across
do not support hypothesis 1.The journey points se- all three months between D-1 and D in segments A

Figure 8: Segments A, B & C – trends in customer payment behaviour.

Behavioral Economics Guide 2021 73

Radhika Bhalla et al. Nudging to Improve the Timely Payment of Loans

& B. On average, an overall decrease of 2 days was text message reminders and redesigned summons
observed in the delinquency of customers in the forms to reduce the failure to appear in court for
experimental group, compared to the control group low-level offenses by 13% - 21%.
for segment A. The results for segment B customers were direc-
The pilot for segment A customers demonstrates tionally similar to segment A, although differences
that leveraging the levers of design – awareness, between the experimental and control groups were
accessibility and education – in communication marginal and insignificant. It was observed that
strategy can influence customer behaviour. It also customers (segment A) who had previously experi-
reveals that using nudge-based messages signif- enced the consequences of default, such as a decline
icantly increased timely repayment and reduced in their credit score (and thus creditworthiness) or
delinquency on average by 2 days a month. nagging collection calls from lenders, responded
Consequently, according to the findings of this more favourably to nudge-based messages than
study, not all customers default on their payments those (segment B) who had never experienced the
on purpose. A large number of them seem to pay late consequences of default.
for a variety of reasons, including a lack of atten- For segment C, the desired shift in customer be-
tion to payment schedules, a lack of understanding haviour cannot be attributed to the nudge-based
about the implications of late or non-payment and messages. Both the control and the experimental
a lack of clarification about the payment process, groups had nearly equal numbers of customers who
among others. Supposedly irrelevant and subtle paid on time. Customers in segment C, like those in
elements such as the use of simple language, em- segment A, had defaulted in the past and had expe-
phasis on critical information, the elimination of rienced the consequences of doing so. However, the
unnecessary clutter and sharing easy-to-access presence of a systematic penalty (in this case, the
action links can be used to overcome such barri- retailer locking the purchased smartphones if pay-
ers. These simplification strategies make it easier ment was delayed) overshadowed the SMS nudges
to interpret information, by focusing attention on and pushed customers in both groups to pay on
the most critical aspects of a message and thereby time. This had an impact on the results, particularly
reducing the likelihood of misunderstanding and because the messages never mentioned the possi-
procrastination (Service et al., 2014). bility of smartphones being locked.
These trends calibrate with other studies, in With respect to repayment behaviour, it remains
which the net difference created by the use of to be explored to what degree the messages’ effec-
nudge-based messaging was between 6% and 21%. tiveness diminishes in the presence of a stronger
Ximena Cadena and Antoinette Schoar (2011) test- vehicle of influence, such as a penalty or an incen-
ed the effectiveness of incentives versus reminder tive, especially at a time when coercive debt collec-
SMS for loan repayment for a micro lender in Ugan- tion techniques are pushing consumers into debt
da. Their messages improved timely repayment by traps and escalating the need for ethical introspec-
7% - 9% relative to the control group, an effect size tion. Meanwhile, research on pro-environmental
similar to the effect of reducing the cost of the loan behaviour suggests that although penalties can de-
by 25% for customers who repaid in full. They also ter behaviours in some instances, they can lead to
found that the average days of delinquency dropped negative affect and defensive responses if deemed
by 2 days a month. Similarly, Karlan et al. (2016) unreasonable (Bolderdijk et al., 2012). Thus, it is
showed that reminder messages to save increased often desirable to turn instead to positive behaviour
total bank savings and savings goal attainment by change strategies (White et al., 2019).
6%, and Medina (2020) demonstrated how remind-
ers for upcoming credit card payments reduced Hypothesis 2 – The new SMS series will nudge
late-payment fees by 14% for a financial manage- customers in the experimental group to use the self-
ment platform in Brazil. Although in a different service digital channel, compared to customers in the
context, a study by Fishbane et al. (2020) also used control group.

Behavioral Economics Guide 2021 74

Radhika Bhalla et al. Nudging to Improve the Timely Payment of Loans

To test hypothesis 2 in segment D, a chi-square Segment D

test of independence was performed in both the
groups to examine the relationship between digital Overall, 300 (7%) more customers used digi-
adoption and the messages. tal channels in the experimental group than in
It is important to note that the scope of the study the control group. Significantly more customers
was to promote digital channels among customers reached out via the digital channels in the experi-
who would otherwise call the service centre for mental group, compared to the control group, in
queries. The study did not intend to increase en- month 1 (14% vs 6%, Ӽ2 [1, N = 4192] = 76.20, p <
gagement with customers not reaching out to DMI .001), month 2 (11% vs 5%, Ӽ2 [1, N = 4192] = 54.73,
Finance via either channel. p < .001) and month 3 (10% vs 4%, Ӽ2 [1, N = 4192]

Table 3: Pilot results for Segment D.

= 55.72, p < .001). Hence, the results support hy- had limited access to them.
pothesis 2.When information was simple to under- The findings of this study, in general, point to
stand, consistent across media and digital channels the benefits of incorporating behavioural science
were extensively promoted via social media, more insights into customer communication. Such inter-
customers chose to use them for self-service. Also, ventions have a direct impact on both collection and
over a four-month period, a reduction of 8% in service costs. The key is to understand the psycho-
queries, such as how to make a payment, requests logical barriers to targeting behaviour and to use
for NOC and welcome letters, was observed follow- specific attention-grabbing and persuasive strat-
ing the refinement of communication across digital egies to overcome them. Additionally, the study
platforms. Additionally, call volume pertaining to explores the opportunity to expand financial liter-
payment status updates reduced by 5% in the same acy among an under-served population, especially
period. This supports the underlying assumption since message content can be easily ported to other
that customers were not using the digital portals technologies.
because they were unaware of their existence and

Behavioral Economics Guide 2021 75

Radhika Bhalla et al. Nudging to Improve the Timely Payment of Loans

Conclusion possibilities. Once a correlation between perso-

nas and payment behaviour is determined, DMI
As Daniel Kahneman famously said, ‘moving for- Finance will be able to distinguish occasional late
ward in behavioural change should be a mix of apply- payers from deliberate late payers, thereby effec-
ing insights from literature and learning from applica- tively deploying both traditional and non-coercive
tion’. This research raises new concerns that should tools, such as nudges, to drive desirable behaviour.
be investigated further in the future. The use of WhatsApp for business communication
The study was affected by uncontrollable external would expand this research to include over 100,000
factors. The socio-economic fallout of the COV- customers, thereby presenting an opportunity to
ID-19 crisis heightened, thereby affecting custom- positively influence their repayment behaviour and
ers’ ability to repay. As a result, the Reserve Bank of subsequently increase collection rates.
India introduced a moratorium on loans when the
pilots were first launched, resulting in higher cases The Authors
of payment delays and defaults that would not yield
any penalty fees from customers. Radhika Bhalla has over 20 years of experience
The study also experienced technical constraints in behavioural research and management consult-
from the external and internal environment. With ing. As a co-founder of Quadrangle Consulting, she
changes in the Telecom Regularity Authority of helps solve complex talent and customer prob-
India’s regulations, message content had to be fre- lems, using the principles of Psychology. Radhika’s
quently adapted to meet new criteria. Also, remind- strengths lie in qualitative research methods and
er messages could only be sent at predetermined psychometrics.
intervals, because the technical system required
minimum three days to provide a payment status Shruthi Ramakrishnan is a behavioural research-
update. This implied that if the system failed to er at Quadrangle Consulting. She is an incoming
update records on time, many customers who may student at the University of Pennsylvania for the
have paid would still receive messages regarding Masters in Behavioral & Decision Sciences (2021-
non-repayment, thereby pushing them to call the 2022). Using insights from behavioural science, she
service centre. The extent to which these factors designs retention, development and engagement
influenced the efficacy of the messages is yet to be programmes for employees and customers.
determined.
Further, since a nudge in itself creates a subtle Prasad G is a management consultant with deep
change in context that could trigger a small in- interest in the design of customer engagement
cremental change in outcome, the study resorted programmes. Prasad leads on-field studies around
to the use of multiple nudges in a message series customer personas and journey maps. He builds in-
spread across a period of three months. This limited terventions that affect the customer lifecycle, from
the study’s capacity to pinpoint which nudge was pre-sales to re-engagement.
the most effective. Nonetheless, it is important for
future studies to measure isolated effects of single Nimit Marwaha is a customer success profession-
nudges, as small nudges may have large effects. al with over 17 years of experience. At DMI Finance,
Moreover, because the intervention aimed at cre- he leads the design and deployment of special initi-
ating a long-term change in a recurring behaviour, atives targeted at increasing customer engagement.
it is likely that habituation will have an impact on Nimit specialises in customer communication pro-
the effectiveness of the messages. As such, SMS grammes.
content may need to be redesigned after 12 months
or four cycles of use, to continue to produce similar References
results.
The future of this research presents substantial Banks, J., O’Dea, C., & Oldfield, Z. (2010). Cognitive

Behavioral Economics Guide 2021 76

Radhika Bhalla et al. Nudging to Improve the Timely Payment of Loans

function, numeracy and retirement saving tra- et. al. (2014). EAST: Four simple ways to apply
jectories. Economic Journal, 120, F381-F410. behavioural insights. The Behavioural Insights
Bolderdijk, J. W., Lehman, P. K., & Geller, E. S. (2012). Team.
Encouraging pro-environmental behaviour with Sunstein (2016). People prefer System 2 nudg-
rewards and penalties. In L. & J. Steg, J. I. M. de es (kind of). SSRN. http://dx.doi.org/10.2139/
Groot (Eds.), Environmental psychology: An intro- ssrn.2731868.
duction (pp. 273-282). John Wiley & Sons. Therese, F., (2017, October 27). Horizontal atten-
Bridges, S., Disney, R., & Gathergood, J. (2008). tion leans left. Nielsen Norman Group. https://
Drivers of over-indebtedness: Enterprise and regu- www.nngroup.com/articles/horizontal-atten-
latory reform. Department for Business UK. tion-leans-left/.
Cadena, X., & Schoar A. (2011). Remembering to pay? White, K., Habib, R., & Hardisty, D. J., (2019). How to
Reminders vs. financial incentives for loan pay- SHIFT consumer behaviors to be more sustaina-
ments. NBER Working Paper No. 17020. ble: A literature review and guiding framework.
Federation of Indian Chambers of Commerce & Journal of Marketing, 83(3), 22-49.
Industry (FICCI), & Ernst & Young. (2020). Non-
banking finance sector in India. http://ficci.in/sp-
document/23386/EY-FICCI-NBFC-oct.pdf.
Fishbane, A., Ouss, A., & Shah, A. K. (2020).
Behavioral nudges reduce failure to appear for
court. Science, 370(6517), eabb6591.
Herispon, H. (2019). An empirical analysis of house-
hold debt behavior determinants. Economics and
finance in Indonesia 65(2), 132-148.
Jensen, N. H., Reuss, L. F., & Rasmussen, S., (2018).
Increasing public debt collection with nudg-
ing: Results of two natural fields experiments.
Scandinavian Journal of Public Administration
22(4), 45-64.
Karlan, D., Morten, M., & Zinman, J. (2016). A per-
sonal touch in text messaging can improve mi-
croloan repayment. Behavioral Science & Policy,
1(2), 25-31.
Lusardi, A., & Tufano, P. (2009). Debt literacy,
Financial experience and overindebtedness. NBER
Working Paper No. 14808.
Lusardi, A., de Bassa Scheresberg, C., & Avery, M.
(2018). Millennial mobile payment users: A look
into their personal finances and financial behaviors.
GFLEC working paper.
Medina, P. C., (2020). Side effects of nudging:
Evidence from a randomized intervention in
the credit card market. The Review of Financial
Studies. https://doi.org/10.1093/rfs/hhaa108.
Muruganantham, G., & Bhakat, R.S. (2013). A review
of impulse buying behavior. International Journal
of Marketing Studies, 5(3), 149-160.
Service, O., M. Hallsworth, D. Halpern, F. , & Algate

Behavioral Economics Guide 2021 77

 Behavioural Economics Is Booming in Banking:
Should You Join In?

Roger Miles* and Wieke Scholten

Conduct and Culture Academy, UK Finance

Financial regulators around the world have turned to behavioural science to identify, contain and predict
human sources of risk in financial supply chains. The most scientifically literate regulators have introduced
“culture assessment” tools, designing their own behavioural experiments and observational fieldwork and
commissioning independent research. Lingering public suspicion of the financial sector after the 2008 fi-
nancial crash bailouts and continuing events of misconduct in the decade since, besides the social stresses
of the pandemic, have led financial firms to reflect on “socially purposeful culture”. Responding to these
challenges, firms are adopting a “behavioural lens” approach, recruiting in-house specialists in social
psychology, behavioural economics and predictive analytics. As a result, for behavioural science graduates,
there has never been a better time to consider a career in the financial sector.

Introduction see why.

Have you ever stopped to wonder which industry Early Behavioural Investigations by the
is showing the fastest increase in rates of recruit- Regulator
ing behavioural scientists, right now? Though you
might guess (reasonably enough) that it’s user For its first few years of existence (2013-19), the
experience specialists in tech businesses, one of the regulator that protects UK financial customers, the
sharpest rises in new roles for applied behavioural Financial Conduct Authority (FCA), ran a series of
insights is in fact in financial services. A new “be- behavioural experiments in consumer protection. It
havioural track” is opening up in our sector, and investigated first: how are financial providers’ sales
we would like to highlight this opportunity, in case activities exploiting consumers’ biases – present
you’d never considered putting your skills to work bias, overconfidence, reference-dependence and so
in the finance industry. on? Way back in 2013, the FCA published a pathfind-
During the past decade, financial regulators in the er guide (FCA, 2013) identifying a “Top 10” of be-
UK and internationally have bought into behaviour- havioural biases that interfere with good practices
al science big-time, looking to the science to solve in retailing financial products. Such as:
longstanding problems of persistent abuses of cus-
tomers (and others) in this notoriously competitive • present bias induces consumers to overspend on
– and historically troubled – field. The good news their credit cards, as they chase immediate grat-
for behavioural scientists is now that the new so- ification
called “conduct approach” to financial regulation • overconfidence is widely found in consumers’
(conduct = behaviour, right?) is established to a (and indeed professionals’) excessive belief in
point where banks are now themselves hiring be- their own skill at ‘picking a “winning stock”’
havioural specialists to pre-empt the attentions of • halo effect leads many customers to follow any
a conduct regulator. Let’s consider the landscape, to financial advice given, ‘because the adviser is
likeable’.
* Corresponding author: [email protected].

Behavioral Economics Guide 2021 78

Roger Miles & Wieke Scholten Behavioural Economics Is Booming in Banking

Next, the regulator recruited an in-house behav- al., 2020; FCA-Rathi, 2021; Foss, 2020; Nabeel &
ioural research team, who then started to question Miles, 2021). From next year, they are about to start
whether financial firms were (knowingly or other- applying behavioural principles to measure how
wise) exploiting patterns of behavioural weakness financial firms’ staff interact in the workplace and
among customers, such as: how far they are really serving their customers’ best
interests.
• Why do customers run up expensive overdrafts, As one result of all this: if you’re a new behav-
when they could easily avoid this by making a ioural science graduate, it is a great time to pursue a
pre-emptive call to their bank? career in finance as firms scramble to get ready for
• Why do credit card customers, who could afford this new “culture assessment regime.” You might
to do so, not bother to pay off their card balances also be tempted by the notion of applying your
sooner? hard-won B/Sci skills to make the world a better
• Why do so many people save so little for their re- place, by cultivating corporate social consciences in
tirement? a field previously seen as somewhat barren of these
ethics.
The regulator has continued to debrief and publish The new breed of behavioural science-powered
a range of resources urging banks and firms to pur- financial regulator is flourishing – more than 50
sue their own behavioural experiments and to act on such agencies around the world describe themselves
the outcomes from these investigations1, hence the as ‘applying behavioural insights’ to develop mar-
nascent boom in financial firms hiring behavioural ket rules (IOSCO, 2017). They also espouse empirical
skills in-house: partly to ‘keep ahead of the regu- testing of ‘what actually happens’ and are suitably
lator’, as new types of regulatory investigation are sceptical of financial brands’ lofty claims to “pos-
rolled out (Miles, 2021), and partly because they are itive social values”. Several regulators’ inspectors
rediscovering how far applied behavioural insights make direct in-workplace behavioural observations
bring all kinds of benefits to product marketing. (“floor-walk tests”) and sponsor primary research,
looking at the banking system from outside-in to
The New Focus: ‘Culture Assessment’, avoid data bias arising from in-group sources they
Coming to a Bank Near You would have previously relied on (the discredited
‘self-assessments’) (Miles, 2021). Perhaps best of
From 2022, aligning with momentum among all, regulators see the new science as fully aligned
financial regulators internationally (DNB, 2015; with their own mission to support economic in-
MAS, 2020; APRA, 2021), the FCA will be pursuing tegrity and financial stability, whilst “retrieving”
a grander ambition: to observe and assess “human post-pandemic public trust in financial firms
risk” across entire firms. This initiative (FCA, 2020; (FCA-Woolard, 2020).
FCA-Steward, 2021) will see regulators deploying a Though the initial change in thinking was polit-
range of new tools to assess behaviour. These tools ical, it is now all about the science. As a simple ex-
and related indicators will evaluate factors such as ample of regulators’ applied behavioural thinking,
leadership integrity, active open-mindedness, psy- a regulator’s inspector will now challenge financial
chological safety (including ‘speak-up’ and ‘an- firms’ employees to describe how they frame or
ti-bystanding’), cognitive diversity, bias-aware- “mentally account” for what their financial firm
ness and reflexivity (Edmondson, 2019; Ewing et is doing, for the public good – and for how long
they’ve ever paused to think about this point. Back
1 Our good friend and colleague Alexandra Chesterfield, in the day – before the 2008 global financial crash
former lead behavioural researcher at the FCA, gives an and/or before the pandemic, whatever your pre-
excellent summary account of these experiments and ferred set-point – people tended to perceive banks
their findings on UCL’s behavioural insights web strand, as an obscure but somehow necessary presence in
Changing Minds (UCL, 2020) our lives.

Behavioral Economics Guide 2021 79

Roger Miles & Wieke Scholten Behavioural Economics Is Booming in Banking

Since then, the system has faced societal upheavals A second pivot-point for regulators has been rec-
in the form of “stakeholder capitalism” emergent ognising a phenomenon that B/Sci types noted ages
after the 2008 crash, and “social equity” questions ago: the proportionate linkage between social proof
during the pandemic. There have been competitive that ‘legitimizes wrongdoing’, and the severity
jolts, too, with the arrivals of “challenger banks”, of events of misconduct (Vaughan, 1999). People
cryptocurrency, service bots and disintermediated break rules more easily when their mates at work do
(peer-to-peer) customers. so as well; the teams we work in, and feel part of,
All of these factors stack on top of a long-term have a huge impact on whether we tend to do the
rumbling public disillusionment with a financial right thing (Scholten, 2018), hence the impetus for
sector that is widely perceived as a habit-bound, two new workstreams among conduct regulators:
opaque, anti-competitive, introspective boys’ club “culture assessment” and “detecting non-financial
that’s possibly past its use-by date. Hence, people misconduct”.
might welcome a chance to reframe their mental Our industry’s rule-makers – and a few of its
picture of banking, for it to be something, well, business leaders – have also meanwhile seen in the
less like what we think of as “banking” and more advent of Covid-19 not only a public health hazard,
like any other utility – the water supply, say, or the but also a second shot at redemption for the finan-
power company, or the Wi-Fi. Just as you flick a cial sector (FCA-Woolard, 2020). Following the
switch and expect the light to come on, so you now 2008 crash, as the public saw it, banks had simply
expect to switch on the bank (app) and get a line of helped themselves to a big chunk of taxpayers’
credit. money, run away with it and used it to rebuild their
shattered balance sheets. As a result – in the UK
From ‘Dark Side’ to Bright Side at least – they had been seen to cause ten years of
hated austerity, generating a wave of vox populi risk
Initially, conduct regulators used behavioural (Fordham, 2016) that has fuelled mass public pro-
analytics to help chase the bad guys and identify tests and electoral shocks. A pandemic, by contrast,
financial misconduct – essentially, selling people might present an unpopular sector with an oppor-
the wrong stuff, or selling stuff to the wrong people, tunity to shine. In its own way, Covid added mo-
or both. Then, regulators moved the focus beyond mentum to an existing regulatory push for ‘socially
narrow prosecutions for mis-selling, to start look- useful’ banking (Carney, 2014): how might firms
ing harder at “non-financial misconduct”. This use behavioural analytics to show us delivering on a
followed recognition, informed by behavioural sci- “fair social contract”, so making the world a better
ence, that all kinds of misconduct can present a val- place? Which analytics proving what outcomes?
uable cultural “tell” to the regulator’s inspector. As Even before the pandemic struck in spring 2020,
one of our research respondents pithily defined it, several countries’ regulatory agencies had already
in a workplace where “behaving like a jerk to people back-flipped their conduct policy (and rhetoric)
you work with” is normal, that behaviour is itself a from “punishing misconduct” towards “promoting
reliable proxy indicator that the firm tolerates gen- exemplary conduct”. Now that Covid has perversely
erally unhealthy attitudes towards customer care handed us a socially aware reset, we might well ask:
and ethics. Firms’ historic tolerance for cultures have banks stepped up to the opportunity to serve
of abusive behaviour, for as long as badly behaved the public good, and to “highlight the bright side”
individuals were selling lots of product, is after all of their work? How’s that “behaviour assessment”
a staple of popular culture; we know it when we see work going; is it proving you’ve changed for the
it2 . better?
Now, all you B/Sci types will probably be me-
2 One need look no further than Hollywood: Wall Street, The ta-analysing that development: is this a policy shift
Wolf of Wall Street, Rogue Trader, The Boiler Room, Glengarry or simply a reframe, a bit of regulatory theatre ma-
Glen Ross, The Big Short, Margin Call… need we go on? nipulating the labels? Is it a sly attempt at instru-

Behavioral Economics Guide 2021 80

Roger Miles & Wieke Scholten Behavioural Economics Is Booming in Banking

mental conditioning, at building new muscle mem- factors, using B/Sci experimental designs and de-
ories for good behaviour by displacing bad habits in veloping AI-assisted predictive tools.
favour of “pro-social patterns”? Is it a long overdue So, surely, this time really is different? Don’t hold
acceptance of the failings of classical punishment your breath. It is, however, just possible that the
theory, by endorsing the power of intrinsic motiva- rollout of the new tools for culture assessment will
tion and recognising the universal human longing break the established trend, which is why we here
to sustain a positive self-image? Maybe just a clas- call on our BE colleagues to support them. Time,
sic action-bias-driven watchdog intervention? Or then, to look at some components of the new be-
availability bias, using new analytic tools “because havioural lens approach that’s going to apply.
they’re there”? All of the above? Or something else?
If we next set out some highlights of the regula- Putting Your BE Skills to Good Use to Make
tors’ “new behavioural agenda” for financial firms Banks Work Better
for the 2020s, you can make up your own mind how
far these present engaging opportunities for BE- Around the world, from Australia and Singapore
ists. to the UK, the Netherlands and Ireland, and onwards
to the USA and Canada, financial conduct regulators
The Mission: Can We Please not Repeat voice a common message: to prevent new scandals,
Previous Crises? we need to assess firms’ culture (Miles, 2021). Pre-
venting scandals is an important way to build trust
As you will know, much B/Sci research addresses in banking – as a foundation for financial stability
this question, albeit meta-analyses (such as Rein- and the economic integrity of nations. Who can dis-
hart & Rogoff, 2011; FICC, 2018) offer few reassur- agree with that?
ances or practical remedies. It seems that, ironically Regulators host regular conferences on culture
enough, the lesson of history for financial markets and behavioural analytics, gathering together
(and humanity in general) is that we’re really bad bankers and encouraging them to engage. As a few
at internalising any lessons from history. Whilst regulators have been doing (such as the Dutch,
regulatory design has the noble aims of trying to Australian and Canadian regulators), many more
keep markets working smoothly, and to stop firms are now taking the next step of actually going into
ripping off customers, it always faces an epistemic financial service firms to run face-to-face culture
problem, in that it serially addresses the previous assessments. Around the world, banks are respond-
crisis, never the next one. ing by investing in their own behavioural risk and
We may therefore predict with confidence – trag- culture assessment capabilities. Behavioural risk
ically – that between today’s publication of the 2021 teams are popping up, staffed with entrepreneurial
BE Guide and its next edition in 2022, there will be behavioural scientists who apply insights to cor-
at least another three major banking scandals, since porate practice, in order to prevent future issues
that’s the observed rate of event risk for such things (Wood, 2021). As a matter of public record, major
happening, over many years. (Just consider the cur- brands, including HSBC, Standard Chartered, ING,
rent business year 2020-21: Wirecard, Wells Fargo ABN AMRO and NatWest/RBS, have all recruited
[part 2], Greensill Capital; not even thinking about behavioural research leaders from academia and/or
many lesser, near-miss events.) Every regulator in regulatory agencies.
our field wants to prevent the next banking scandal The impulse for this article, in fact, was conver-
– has always wanted to – but we have to expect that sations these authors have had with professional
they will continue to fail in this endeavour. friends who are pioneering the application of B/
Yet we do now know that banking scandals are Sci to financial market practice in those brands and
the outcome of poor cultures and behaviours. The elsewhere. Just as there are various different ways
new regulators call this insight ‘behaviour-at-risk’ for regulators and banks to go about this task, so
and are working on identifying the drivers of these there are different points where a canny BE practi-

Behavioral Economics Guide 2021 81

Roger Miles & Wieke Scholten Behavioural Economics Is Booming in Banking

tioner within a financial firm may choose to plug in havioural data, they will find potentially corrupting
their own personal skillset. team cultures and intervene to fix these before bad
One such point involves applying behavioural behaviour becomes a systemic norm. It is rarely
insights to the way a firm makes new products the case that misconduct (bad behaviour) is down
for customers (just as BE is already doing in many to a single ‘bad apple’ (Scholten & Ellemers, 2016),
other consumer product fields, in fact). Too many since misbehaviour has often grown within a team
financial products in the past – unintentionally or or subculture that encourages – sometimes unin-
not – tempted customers to spend more money tentionally – people to bend the rules. Social proof
than they could afford in the long run. Thanks to (“everybody does it”) is of course a strong fuel for
misbehaviour by some brands, our industry has left misconduct in close-working teams – as are unfair
an infamous trail of exploitation of customer biases treatment by a line manager, skewed incentives
such as short-termism, reference-dependence and (cash bonuses for “hit-and-run” selling) and
availability. As we’ve just seen above, behaviour- “stretch target” pressures to recoup revenue lost
al specialists are now working in many banks to during the pandemic.
redress the balance towards the customer’s good, In fact, we may sort these behavioural drivers into
to help prevent mis-selling and improve product organisational drivers (such as strategy, steering, in-
design by the better framing of purchase decisions. centives, codes), social drivers (moral climates, psy-
Now, if a bank thinks of, say, a new type of mort- chological safety, shared beliefs), individual drivers
gage or savings app, the people involved in creating (cognitive biases, motivation) and contextual drivers
it will need to test that the product is genuinely (such as market conditions, the pandemic, Brexit)
helpful to customers and won’t harm anyone. As (DNB, 2015). Behavioural science, of course, illumi-
with any new product development, behavioural nates how these factors may lead to rule-breaking
science can help uncover pitfalls and unexpected behaviour.
consequences, to reveal weaknesses in product or B/Sci shows us where to look for the “tells” that
process. BE can model and predict how the custom- identify the early stages of misconduct and inter-
er might buy the new product; is it friction-free and vene with targeted actions to deter misbehaviour.
yet also designed to ensure that the customer makes Behavioural or cultural risk assessments often com-
a sound, rationally informed choice? bine quantitative and qualitative research methods.
As you might also know – if you have followed By all means keep on doing surveys, but combine
news events such as the Panama Papers or the Fin- them with confidential informal conversations with
Cen Files leak – criminals try (and sometimes suc- employees, depth interviews and observations of
ceed) to use banks to “launder” their illegal earn- daily work situations. The analysis is strongly based
ings. To prevent this from happening, the knack on the type of scientific research methods (coding,
is to detect a crook in time; to identify patterns of grounded theory, statistical analysis) that make any
behaviour that reveal a criminal who’s disguised as behavioural scientist feel right at home.
a “legitimate” new client, or who may have slipped
into the bank’s existing client system undetected Looking to the Future: Even More
some time ago. Increasingly, behavioural analytics Opportunities for BE-ists
are used to prevent financial crime and detect vul-
nerabilities. As a simple example, even well-intend- We see opportunities for BE practitioners only
ed bank managers are prone to halo bias, believing expanding, as our industry gathers pace with these
that there could not possibly be any crooks amongst reforms. On both sides – regulators and firms –
their clients: ‘I’ve known all of them for years, there there is an exploding demand for designing and
is no way even one of them is dishonest!’. interpreting cultural assessments; we’ve just pub-
Finally regulators, and banks themselves, are lished well-received research on this very point
now using behaviour-based culture assessments (Miles, 2021). Let’s leave you with three example
to forestall future problems. Searching various be- work opportunities to consider:

Behavioral Economics Guide 2021 82

Roger Miles & Wieke Scholten Behavioural Economics Is Booming in Banking

1. Regulators all around the world are joining in the The Authors
hunt for new assessment tools with greater capabil-
ity to detect social drivers of (good and bad) culture Dr Roger Miles is behaviour-at-risk research
and behaviour. If your research work has produced lead, co-founder and head of faculty at UK Finance
any new model insights, such as previously undis- Conduct and Culture Academy and author of Conduct
covered cause-effect linkages, pitch away. Risk Management: A Behavioural Approach (Kogan
2. Any analytic tools that help shorten the dis- Page, 2017) and Culture Audit in Financial Services
tance and/or time lag between an event of misbe- (Kogan Page, 2021).
haviour and the early detection of that event will
be eagerly welcomed. Such tools include improved Dr Wieke Scholten is senior behavioural specialist
observation technologies (yes, we mean surveil- at &samhoud consultancy, a former senior super-
lance – which is, of course, controversial for many) visor of behaviour and culture, Dutch Central Bank,
and pattern-recognition tools (such as AI, massive and author of Banking on Team Ethics (2018) and
parallel processing and social network analytics). co-author of Supervision of Behaviour and Culture
Behavioural science may help to catch misbehav- (DNB, 2015).
iour at an early stage, stopping the slippery slope by
which an initial innocent mistake can morph into References
serial unethical actions. Predictive analytic tools are
increasingly drawing on behavioural experimental Australian Prudential Regulatory Authority (APRA)
design to improve the accuracy of their forecasts of (2021). APRA’s evolving approach to super-
bad behaviour, to prevent “norms of misconduct” vising risk culture. https://www.apra.gov.au/
from gaining traction. apra%E2%80%99s-evolving-approach-to-su-
3. Finally, there is plenty of work to do to pull pervising-risk-culture.
some of BE’s mighty theories into the practical Carney, M. (2014, May 27) Inclusive capitalism –
application space that our financial brands are now creating a sense of the systemic. Mark Carney,
so keen to expand. We know of banks who would be Governor of the Bank of England, speech to BIS
interested in talking to anyone who has managed to Conference on Inclusive Capitalism. https://
develop robustly empirical and universal indicators www.bis.org/review/r140528b.pdf.
for human risk factors such as bystanding, psy- De Nederlandsche Bank (DNB) (2015). Supervision
chological safety, cognitive diversity, situational of behaviour and culture: Foundations, prac-
awareness, vox populi risk, hysteresis and moti- tice & future developments. https://www.dnb.
vated reasoning. Though we have been following nl/media/1gmkp1vk/supervision-of-behav-
the experimental literature on all of these elements iour-and-culture_tcm46-380398-1.pdf.
– thanks Alain – as yet we have seen only a limited Edmondson, A. (2019). The fearless organisation:
number of reliable prototype scorecards. Who’s up Creating psychological safety in the workplace
for this challenge? for learning, innovation and growth. Harvard
As practitioners, we are constantly reminded Business School / John Wiley & Sons.
what a privilege it is to be working in this field at Ewing, P., Bloomfield, C., Fahy, O., Ben Haddel, N.,
such a formative time. The science has shown this & MacDonald, T. (2020, June 10). Conduct, cul-
somewhat late-arriving sector the value of de- ture and Covid-19. FCA Insight. https://www.fca.
ploying behavioural insights to intervene in and org.uk/insight/conduct-culture-and-covid-19.
improve financial firms’ recognition of their wider FICC-Markets Standards Board. (2018). Behavioural
value to society. Further ahead lies a promised land cluster analysis: Misconduct patterns in finan-
of predictive behavioural indicators for all of the cial services, Report. http://www.fmsb.com/
above – and more. But before we dive into that pool wp-content/uploads/2018/07/BCA_v32_1.pdf.
and go all Minority Report on you… wait until next Financial Conduct Authority (FCA) (2013). Occasional
year’s BE Guide. paper 1: Applying behavioural economics at the

Behavioral Economics Guide 2021 83

Roger Miles & Wieke Scholten Behavioural Economics Is Booming in Banking

Financial Conduct Authority. https://www.fca.org. www.ivey.uwo.ca/leadership/for-leaders/lead-

uk/publication/occasional-papers/occasion- ership-blogs/2020/05/leader-character-in-
al-paper-1.pdf. isolation-and-for-mental-health/.
Financial Conduct Authority (FCA) (2020, March Miles, R. (2021). Culture audit in financial services.
5). Discussion paper 20/1: Transforming culture Kogan Page. http://www.koganpage.com/cafs.
in financial services – driving purposeful cultures. Monetary Authority of Singapore (MAS) (2020).
https://www.fca.org.uk/publications/discus- Good culture and conduct: MAS and industry as-
sion-papers/dp20-1-transforming-culture-fi- sociations established steering groups to pro-
nancial-services-driving-purposeful-cultures. mote sound culture and strengthen standards of
Financial Conduct Authority (FCA) (2020, April conduct for banks and insurers. Annual Report
20). Chris Woolard, Interim Chief Executive, in- 2019-2020, Monetary Authority of Singapore.
terviewed in Financial Times: UK regulator says https://www.mas.gov.sg/who-we-are/annu-
coronavirus is first test of post-2008 banking rules. al-reports/annual-report-2019-2020/respon-
https://www.ft.com/content/0ec16280-423f- sible-and-trusted-financial-centre/good-cul-
405a-a081-b51fee20d6c7. ture-and-conduct.
Financial Conduct Authority (FCA) (2021, March 17). Nabeel, H., & Miles, R. (2021). The new mindset
Speech by Nikhil Rathi, CEO: Why diversity and and language of culture: Assessing financial
inclusion are regulatory issues. https://www.fca. and non-financial misconduct. In R. Miles (Ed.)
org.uk/news/speeches/why-diversity-and-in- Culture Audit in Financial Services (pp. 154-171).
clusion-are-regulatory-issues. Kogan Page.
Financial Conduct Authority (FCA) (2021, April 26). Reinhart, C. M., & Rogoff, K. (2011): This time is dif-
Speech by Mark Steward, Executive Director of ferent: Eight centuries of financial folly. Princeton
Enforcement and Market Oversight: Compliance, University Press.
culture and evolving regulatory expectations. Scholten, W. W. (2018). Banking on team ethics: A
https://www.fca.org.uk/news/speeches/compli- team climate perspective on root causes of miscon-
ance-culture-and-evolving-regulatory-expec- duct in financial services. Doctoral dissertation,
tations-mark-steward. Leiden University.
Fordham, T. (2016). Vox populi risk: A future where Scholten, W. W. & Ellemers, N. (2016). Bad ap-
aggregate economic growth no longer guaran- ples or corrupting barrels? Preventing traders’
tees political stability. Citigroup / World Economic misconduct. Journal of Financial Regulation and
Forum Foresight. https://reports.weforum.org/ Compliance, 24(4), 366-382.
global-strategic-foresight/tina-fordham-citi- University College London (2020, July 20). Banking
group-vox-populi-risk/. on behavioural insights: Alexandra Chesterfield.
Foss, B. (2020, August 11). Cognitive diversity Changing Minds. https://www.ucl-changing-
(what’s that?) – and its value to risk oversight minds.org/webinar/banking-on-behaviour-
and risk management. The Risk Coalition. https:// al-insights/.
riskcoalition.org.uk/blog-posts/cognitive-di- Vaughan, D. (1999). The dark side of organizations:
versity-whats-that-and-its-value-to-risk- Mistake, misconduct and disaster. Annual Review
oversight-and-risk-management. of Sociology, 25, 271-305.
International Organization of Securities Wood, A. (2021, March 18). Behavioral science gains
Commissions (IOSCO) (2017). IOSCO task force traction as more banks seek to mitigate employ-
report on wholesale market conduct. IOSCO Task ee risk. Thomson Reuters. http://www.thomson-
Force FR07/2017. http://www.iosco.org/library/ reuters.com/en-us/posts/corporates/behavio-
pubdocs/pdf/IOSCOPD563.pdf (archived at ral-science-bank-risk/.
https://perma.cc/ N2GW-JCTR).
MacDonald, T. (2020). Leader character: In isola-
tion and for mental health. Ivey Institute. https://

Behavioral Economics Guide 2021 84

 Tired of Behavioral Economics? How to Prevent
the Hype Around Behavioral Economics From
Turning Into Disillusionment
Florian Bauer and Manuel Wätjen*
Vocatus

Applying the behavioral economics effects found in academic experiments to marketing is becoming more
and more popular. However, there is increasing evidence that copy-and-pasting academic effects does
not achieve the desired effects in real life. This article aims to show that this is not because customers are
becoming wise to nudges or that behavioral economics does not work at all, but because the application
of behavioral economics typically ignores the contextual aspects of the actual decision to be influenced.
Herein, we present a framework that considers these aspects and helps develop more effective behavioral
interventions in marketing, pricing, and sales.

Situation: Doubts and Disillusionment effects are used (S-M-L portfolios on every cor-
Replace Initial Enthusiasm ner; artificial scarcity (“Only 3 rooms left!”) are
not only implausible, but they have also become
Over the past few decades, behavioral economics annoying to many customers. Effects that are so
has not only revolutionized economic thinking, but crudely implemented in practice cause customers
it has also significantly changed business manage- to lose trust in providers (Shaw, 2019).
ment. The focus on the decision-making process • In B2B: Here is where the transferability of the
of customers and the associated greater attention behavioral economics effects has always been
to touchpoints in the customer journey make be- questioned more strongly, based on the assump-
havioral economics a great source of ideas on how tion that professional decision-makers should
to influence decision-making behavior systemat- act much more rationally than private customers
ically. Ultimately, there is additional potential for when making complex decisions.
increasing conversion and margins behind every
effect of predictable irrationality. It seems as if the This can give the impression that the findings
only thing to do is to transfer individual effects of behavioral economics are a dying form of hype
from the academic literature to practice, in order to (B2C) or that they are not even worth trying in the
increase company results significantly. first place (B2B).
As great as the enthusiasm was—and still is—for This skepticism is well-founded. However, this
taking up and implementing the findings of behav- is not because behavioral economics does not work
ioral economics in marketing, pricing, and sales, generally (B2B) or any more (B2C) but because we
doubts are growing that a simple transfer of aca- tend to make it too easy for ourselves when it comes
demic effects into practice does not always work to practical application: We may spot a funny ef-
(Smets, 2018): fect in a book or even a scientific article and then
try to apply it to our own marketing (Piper, 2020).
• In B2C: Obtrusiveness and the extent to which the This can be done swiftly, especially in e-commerce,
where traffic is often high, and A/B tests are quick-
* Corresponding author: [email protected]. ly implemented. Sometimes, this direct transfer

Behavioral Economics Guide 2021 85

Florian Bauer & Manuel Wätjen Tired of Behavioral Economics?

works and is rewarded with higher conversion rates, (Popper, 1963): Even if you see only white swans
sometimes nothing happens at all, and sometimes your entire life, that does not mean you should claim
it backfires and results in fewer sales than before. that all swans are white. But if you see a black swan,
Why do the effects work sometimes, but some- you can certainly say that not all swans are white.
times they do not do so? Is it really because the In the best case, this means that what we learn with
behavioral economics insights are not—or no every empirical observation or experiment is which
longer—valid, as customers have become accus- model of the world is wrong, but never which is
tomed to them, or is it because they have never been right. The purpose of experiments is not to develop
valid? a theory but to falsify it. Experimental results can
As we shall show in this paper, there are two be—at most—inspirations for developing a theory,
answers to this question: The nature of academic but they can never replace it.
empirical behavioral economics, based on which In this respect, behavioral economics has impres-
the widely cited effects were found, and the typ- sively succeeded in showing that people do not de-
ical approach of transferring scientific findings cide rationally. However, what behavioral econom-
into practice. We shall show that the combination ics has not yet achieved is to contrast the model of
of experience background, situation, and heuris- rational decision-making with an alternative model
tics (condensed in a typology of decision-making that can explain all the experimental findings with
strategies) helps resolve the issues of transferring as few assumptions as possible (‘Ockham’s Razor’).
academic insights into practice. Only such a model would allow marketing practi-
tioners to plan interventions, which, provided the
Background: The Goal of Empirical theory is valid, we could expect to have an effect in
Behavioral Economics the intended sense (e.g., to influence people’s deci-
sions in a particular direction).
Academic behavioral economics is not primari-
ly about helping practitioners find new marketing The Conceptual Consequence: ‘Homo
tricks. Instead, it was—and continues to be—di- Heuristicus’ as a New Paradigm
rected towards a very simple goal: It wants to refute
the model of rational decision-making that under- In the absence of an alternative empirically-based
lies neoclassical economic theory. model of human decision-making behavior, popu-
Academic research focusses on model falsification lar science and practitioners have quickly settled
and has an entirely different goal to practical ap- on ‘Homo Heuristicus’ (alias ‘Homer Simpson’;
plication. It is about disproving that people always Gigerenzer & Brighton, 2009) as a counter-model
decide rationally and always maximize their utility. to ‘Homo Economicus’ (alias ‘Mr. Spock’) or ‘Econ’
It is about showing that people are not always per- (Thaler, 2005). This model is bold and intuitive.
fectly informed or always have stable and intransi- However, this model is negative in the sense that it
tive preferences. This makes behavioral economics is not only designed to falsify ‘Homo Economicus’,
essentially a ‘negative’ endeavor that repeatedly but also that it primarily points out the inadequa-
shows us how people are not making decisions. Yet, cies of the human perception and decision-making
the effects, based on which this is shown, neither apparatus (take the terms ‘biases’ or ‘misbehaving’
embody the actual scientific message nor claim as an example; see Tversky & Kahneman, 1974,
generality through their selective proof. They are and Thaler, 2005, respectively). While this model
only a means to an end. This procedure is entirely is very easy to understand, it also complicates the
legitimate, albeit from a strictly academic point of acceptance of behavioral economics: The fact that
view. customers are biased and can misbehave may make
Experimental results—very many of them—do us smirk, but it is hard to convey, especially when
not show us what things are but at best what they are we talk about business customers and professional
not. Take Popper’s famous black swan, for example purchasers.

Behavioral Economics Guide 2021 86

Florian Bauer & Manuel Wätjen Tired of Behavioral Economics?

The Practical Consequence: Nudging as that has become no longer useful.

an Imperative, and Copy-and-Paste as a Nudging via copy-and-paste is a lot of fun and
Method suits the zeitgeist of corporate organization, thanks
to its agile approach. However, that alone is not a
Scientific behavioral economics experiments are guarantee of success. Successful interventions are
not—and were never intended to be—a positiv- based on valid models. This is where Einstein’s fa-
ist guide to practice. However, in the absence of a mous quote holds true: “Nothing is more practical
practical theory to apply behavioral economics to than a good theory.” With this in mind, we set out
real-life problems, the nudging hype has elevated in the following sections to find a consistent deci-
precisely this positivism religiously (even though sion-making model that will enable and facilitate
this may not have been the intention of the orig- practitioners to apply the findings to practice.
inal authors). Now, the key insight of behavioral
economics is no longer seen in the refutation of From Academic Behavioral Economics to
the rational decision-making model or other basic Practical Behavioral Marketing
assumptions of neoclassical economic theories. It
is rather seen in the selective result in itself, which From Biases to Heuristics
is transferred to practice with the expectation that
the experimentally elaborated effects found under Let us start with the following question: Why do
laboratory conditions will also show up in practice. people actually make predictable mistakes in their
The fact that many of the published effects are decisions? Why don’t we just learn to make the
contradictory to each other (because they suggest ‘right’ decisions?
conflicting recommendations for practice) makes Why we decide the way we do inevitably leads us
clear that this is the wrong approach. Take the to the bigger question of why we are the way we are.
question of how many options to offer a customer We are the way we are because the mechanisms of
to convince him to buy: Only one option to avoid the evolution have shaped us that way. Our entire con-
‘paradox of choice’? Two options in order to use one stitution has simply proven to be sufficiently well
as an ‘anchor’? Three options for the ‘Goldilocks’ adapted, given the environment in which we live.
effect? Behavioral economics has no answer to this This is also what is meant by the famous and often
question, as this was not the aim of the research misused phrase ‘survival of the fittest’. Put simply,
efforts. Each of these effects in itself contradicts the traits that survive best and are more likely to be
one or more basic assumptions of rational choice inherited are those that are best adapted to prevail-
theory and thus has an epistemic value from an ac- ing environmental conditions. This is not an active
ademic point of view—but it has no specific value selection or development towards an ‘ideal’ goal
from a practical point of view. This paradox leads to but a passive selection. Thus, those characteristics
the following notion: The more effects we find, the that were a hindrance to survival and inheritance in
more difficult it is to apply them, because the more the past simply ceased to exist.
likely it is that effects contradict each other. For example, we do not perceive light, volume,
This makes the fascination around nudging a and many other stimuli in absolute terms, such as
double-edged sword: It is nice to get more ideas for a physical measuring instrument, but in relative
implementation with every new scientific publica- terms: The brighter the light, the more additional
tion (and scientific journals explicitly ask for such light intensity it takes for us to notice a difference.
suggestions), but we should not forget that this While this is not optimal in the sense of ‘objectivi-
superficial derivation often harbors more risks than ty,’ it is extremely efficient, because it allows us to
opportunities. The most considerable risk of all is cope with a much wider range of brightness differ-
that based on a superficial and ultimately unsuc- ences. This is why our ancestors could still perceive
cessful transfer, the whole topic of behavioral eco- the saber-toothed tiger at dusk as well as in bright
nomics is dismissed either as unhelpful or as hype sunshine and run away in time. In this respect,

Behavioral Economics Guide 2021 87

Florian Bauer & Manuel Wätjen Tired of Behavioral Economics?

such biased brightness perception has evolved into capacities are always limited. Thus, people con-
a survival advantage and could be passed on to the stantly apply heuristics, to be able to make deci-
next generation, as the tiger did not eat our ances- sions at all.
tors. The contrast effect and the relativity effect The use of heuristics has nothing to do with in-
of behavioral economics can be traced back to this competency (which the terms ‘biases’ and ‘misbe-
perceptual heuristic. having’ suggest). It is rather a handy and efficient
Just as our bodies are not designed for any kind (and indeed the only possible) way of processing
of optimum, our decision-making apparatus is not information and making decisions, and it has prov-
made for utility maximization; rather, it functions en good enough in many situations. However, while
‘well enough’ in most, but not perfectly in all, situ- the same decision-making rule may lead to a util-
ations (‘satisficing’ instead of ‘optimizing’; Simon, ity-maximizing outcome in one situation, it may
1956). Decision-making rules that have proven to lead to a predictable decision error in another. The
be ‘good enough’ in certain situations are stored as ‘irrationality’ of a given heuristic, thus, is rooted
‘heuristics’, which can be thought of as conscious in the fact that it has proven to be ‘good enough’ in
or unconscious rules of thumb. They allow us to many situations, but in others, on the contrary, it
make decisions even with limited capacity, ability, has turned out to be inadequate.
time, and energy. The core message of behavioral Let us look at the following example of the scor-
economics is not that people always make irrational ing heuristic as illustrated in Figure 1 (Bauer, 2000).
and error-prone decisions but that their cognitive In many cases, it is good enough to rate offers ac-

Figure 1: Preferences for telco plans.

cording to their individual attributes on a binary in both choice situations, the zero-sum shift of 270
scale (better/worse) and sum up these individual DKK from device to installment costs for offer 2 in
judgments to form an overall judgment. However, choice 2 means that offer 1 ‘wins’ in two out of three
in some situations, it leads to a systematic misjudg- price elements. Consequently, offer 1 is perceived
ment: cheaper in choice 2 and preferred in the experiment,
Although both offers have equal (absolute) costs even though effective costs are still the same.

Behavioral Economics Guide 2021 88

Florian Bauer & Manuel Wätjen Tired of Behavioral Economics?

From Heuristics to Decision-Making Strategies However, when removing the print-only option
(that nobody wanted!), the decoy that makes the
To understand (and influence) how people make combination appear a good deal is missing, result-
decisions, i.e., to apply behavioral economics in ing in a shift of preferences to the cheaper e-pa-
practice, we need to understand when they use per-only option in choice B.
which heuristics (situation and experience back- However, marketing practice can often not rep-
ground), and how specific heuristics are activated licate this effect, because the situation is different.
(by the design of choice architecture). While one would make the same choice with regard
Let us look at the much-cited Decoy Effect (Ariely to these offers in the experiment (preference), in an
& Wallsten, 1995). actual purchase situation getting two products of
In choice 1, as illustrated in Figure 2, by an obvi- very different value for the same price might make
ously inferior option (print-only), people strongly people more skeptical and may keep them from
prefer the bundle option of both e-paper and print. buying (actual behavior).

Figure 2: Decoy Effect.

Think of the ‘Goldilocks’ effect (De Ridder, 2008) the fact that different heuristics combine to form
for another example. Those who have had experi- holistic decision-making strategies (e.g., reduc-
ences with data throttling or high roaming costs in ing uncertainty), that certain decision strategies
excessively cheap mobile plans, and being ripped off are more likely in certain decision situations (e.g.,
with overly expensive plans, will be more likely to purchasing a mobile phone plan compared to, for
opt for a mid-range plan in situations where there instance, grocery shopping), and that different sets
are several mobile plans from which to choose. Poor of heuristics (e.g. Goldilocks, Bandwagon) are acti-
experiences thus make the use of heuristics that vated in such situations.
focus on minimizing risk more likely. In situations For practical purposes, this suggests segment-
where uncertain customers perceive a high risk for ing customers according to their decision-making
wrong decisions, due to many decision options, strategies, to predict which behavioral economics
highlighting mid-range plans (design of the choice effects are best suited for which marketing task,
architecture) can activate or reinforce the Goldi- and to avoid the mistakes of an overly simple copy-
locks heuristic. and-paste approach.
Although not all heuristics are applied in all situ-
ations, people use a rather systematic approach to
making purchasing decisions. This is reflected in

Behavioral Economics Guide 2021 89

Florian Bauer & Manuel Wätjen Tired of Behavioral Economics?

From Decision-Making Strategies to Decision- • Using the GRIPS types has been shown to influ-
Making Typology ence decision-making behavior: In many projects
across different industries, we have demonstrated
The GRIPS typology (Bauer & Wätjen, 2018) is significant increases in conversion and margins
one way to segment customers according to their through the type-specific application of behavio-
decision-making processes, with a focus on the ral economics effects:
application of behavioral economics to marketing,
pricing, and selling: º Banking, branch office: Increasing the rate
Three aspects show the validity and practical rel- of scheduled consultation appointments by
evance of this decision-making typology: a factor of 3
º Energy, mailing: Reducing churn rate by 31%
• The GRIPS types (Figure 3) replace the negative º Insurance, branch office: Reducing average
‘Homo Heuristicus’ model with a positive model rebates by 44%
of how people really decide. º Print media, call center: Increasing conver-

Figure 3: GRIPS typology.

• The GRIPS types do indeed react differently to sion rates by 148%

behavioral economics effects (Figure 4), but dif- º Telco, call center: Increasing conversion
ferentiation by situation also resolves the contra-
dictions that arise from the overly simple copy- has no negative effect for the Bargain Hunter, because
and-paste approach. 1
multiple options do not complicate the decision for his
decision-making strategy. This contrasts with the Risk
1 For example, bargain hunting (or seeking to optimize Avoider, who does not want to maximize the advantag-
the transaction utility of a purchase decision) is an inher- es of a decision but does wish to minimize its disadvan-
ently consistent decision strategy. We can predict that the tages. The ‘decoy effect’ makes him less decisive because
‘decoy effect’ works well for a Bargain Hunter because it he is more skeptical, and the ‘paradox of choice’ affects
changes the decision context to give the option an obvious him because the numbers of possible wrong decisions in-
high transactional utility. And that ‘paradox of choice’ crease.

Behavioral Economics Guide 2021 90

Florian Bauer & Manuel Wätjen Tired of Behavioral Economics?

rates by 35% private customers are? Do the findings of behavio-

ral economics apply here at all?
From B2C to B2B
The answer is yes, because:
Let us return to our second initial question: Are
professional decision-makers more rational than • First, heuristics have evolved and are hardwired

Figure 4: Reaction of the GRIPS types to behavioral economics effects.

into our perceptual apparatus. An example in this • Third, the decision-making structure often cre-
regard is the contrast effect, according to which ates misaligned incentives and structural irra-
we perceive differences (e.g., light/dark) stronger tionalities. Incentivizing buyers with negotiated
than they actually are. Given two different pric- discounts, for example, does not necessarily lead
es, the fact that the smaller one is perceived to to the optimal long-term, cost-effective pur-
be lower than it actually is, and the larger one is chasing decision but to the decision in favor of the
perceived to be higher, is a perception that even supplier that gives the largest discount.
business customers cannot defy.
• Second, B2B is sometimes formally involved in If we abandon the negative model of ‘Homo Heu-
more important decisions with more expensive risticus’, alias ‘Homer Simpson’, and use a positive
consequences. Nonetheless, B2B decision-mak- model instead, for example the GRIPS typology, the
ers are rarely personally liable, and therefore they acceptance of the validity of behavioral economics
often do not have the same aspiration to make the in the B2B sector will be much higher.
‘right’ decision compared to private consumption
choices. And when they do, the ‘right’ decision is Summary and Implications
not necessarily the rational one but the one that
the decision-maker can best represent and sell As we have shown, the unreflective collection
internally. of ever-increasing behavioral economics effects

Behavioral Economics Guide 2021 91

Florian Bauer & Manuel Wätjen Tired of Behavioral Economics?

Figure 5: GRIPS in B2B.

harms the idea of applying behavioral economics gy and Economics at the Technical University of
in practice more than it helps: The negative, over- Darmstadt (TU Darmstadt), the Massachusetts
simplified ‘Homo Heuristicus’ model is a barrier Institute of Technology (MIT), and Harvard Uni-
to accepting the validity of behavioral economics, versity before going on to complete his doctorate
especially in areas like B2B. In addition, the ap- at TU Darmstadt. Having spent four years working
plication of nudging via copy-and-paste creates as a strategy consultant for Booz Allen Hamilton,
disappointments, as not every effect works in every he joined with two partners in 1999 to found the
situation, and it can even generate more and more consulting firm Vocatus AG. Bauer is a sought-af-
contradictions. ter expert on the topics of pricing psychology and
The good news is, the combination of experience, behavioral pricing, and he has authored numerous
situation, and heuristics solves the contradictions articles and books on pricing strategy and pricing
of the copy-and-paste approach, and it explains research. He is also an honorary professor at the
why which effects work in which situations. The Technical University of Munich.
decision typology GRIPS is (certainly not the only,
but probably the best practically proven) a positive Manuel Wätjen is a member of the management
model that predicts and influences decision-mak- team at Vocatus with a focus on applying behavioral
ing behavior, as it segments decision-making strat- economics to portfolio, product, and pricing strate-
egies (and the typical heuristics involved in these). gy in B2B and B2C. He studied Social Sciences at the
In sum, the key challenge for applying behavioral Ludwig Maximilian University in Munich and the
economics in practice is to understand which deci- University of Queensland in Brisbane, and worked
sion-making strategies are activated in which de- as a brand strategy consultant before joining Voca-
cision-making situations. What we can then do in tus in 2010.
marketing, pricing, and sales is to shape this deci-
sion context actively (choice architecture), in order References
to influence the likelihood of people using specific
heuristics. Ariely, D. & Wallsten, T. S. (1995). Seeking subjec-
tive dominance in multidimensional space: An
The Authors explanation of the asymmetric dominance ef-
fect. Organization Behavior and Human Decision
Professor Dr. Florian Bauer studied Psycholo- Process, 63(3), 223-232.

Behavioral Economics Guide 2021 92

Florian Bauer & Manuel Wätjen Tired of Behavioral Economics?

Bauer, F. (2000). Psychologie der Preisstruktur:

Entwicklung der Entscheidungspsychologischen
Preisstrukturgestaltung zur Erklärung und
Vorhersage nicht-normativer Einflüsse der
Preisstruktur auf die Kaufentscheidung. CS Press.
Bauer, F. & Wätjen, M. (2018). A positive typology of
irrational decision strategies. In A. Samson (Ed.),
The Behavioral Economics Guide 2018 (pp. 111-119).
https://www.behavioraleconomics.com/be-
guide/the-behavioral-economics-guide-2018/.
De Ridder, J. (2008). Goldilocks pricing for broad-
band: Options for pricing access to the national
broadband network. Telecommunications Journal
of Australia, 58(1), 12.1-12.13.
Gigerenzer, G. & Brighton, H. (2009). Homo heuris-
ticus: Why biased minds make better inferences.
Topics in Cognitive Science, 1, 107-143.
Jacowitz, K. & Kahneman, D. (1995). Measures of
anchoring in estimation tasks. Personality and
Social Psychology Bulletin, 21(11), 1161-1166.
Piper, K. (2020, February 26). Why we can’t
always be “nudged” into changing our
behavior. Vox. https://www.vox.com/
future-perfect/2020/2/26/21154466/re-
search-education-behavior-psychology-nudg-
ing.
Popper, K. (1962). Conjectures and refutations: The
growth of scientific knowledge. Routledge.
Shaw, S. (2019, June 12). Consumers are becoming
wise to your nudge. Behavioral Scientist. https://
behavioralscientist.org/consumers-are-be-
coming-wise-to-your-nudge/.
Simon, H. (1956). Rational choice and the structure
of the environment. Psychological Review, 63(2),
129-138.
Smets, K. (2018, July 24). There is more to behavioral
economics than biases and fallacies. Behavioral
Scientist. http://behavioralscientist.org/there-
is-more-to-behavioral-science-than-biases-
and-fallacies/.
Thaler, R. (2005). Misbehaving. The making of be-
havioral economics. W. W. Norton & Company.
Tversky, A. & Kahneman, D. (1974). Judgement
under uncertainty: Heuristics and biases. Science,
185, 1124-1131.

Behavioral Economics Guide 2021 93

 Crowdsourcing Nudges: Insights and
Experiences From Two Healthcare Systems
Amir Goren* Mitesh Patel
Geisinger University of Pennsylvania

Michelle N. Meyer and Christopher F. Chabris

Geisinger

Behavioral science-informed changes to choice architecture can help improve decisions and outcomes in
healthcare. Given the complexity of the healthcare setting, involving numerous tools shared by stakehold-
ers across a wide variety of specialties and roles, it is challenging to identify and prioritize problem areas
and viable solutions to nudge more optimal patient and provider behaviors. Crowdsourcing can be a power-
ful means of generating innovation and prioritizing problems and potential solutions. Herein, we describe
how the Geisinger Behavioral Insights Team and the Penn Medicine Nudge Unit implemented crowdsourc-
ing competitions to solicit nudge problems and solutions from employees in two healthcare systems. Both
nudge units found the competitions successful, not only in generating viable, novel project ideas, but also
in increasing awareness of their work and helping them engage with relevant stakeholders within their
respective organizations.

Introduction er Behavioral Insights Team (BIT) and the Penn

Medicine Nudge Unit (PMNU), have deployed and
Nudges, which are behavioral science-informed assessed nudge interventions across multiple clin-
changes to how choices are organized and present- ical areas and delivery modalities (e.g., Patel et al.,
ed to decision-makers, have proven helpful in im- 2018; Patel et al., 2016a; Patel et al., 2016b), includ-
proving decisions and outcomes across numerous ing, most recently, patient-directed text messages
domains, including healthcare (Thaler & Sunstein, to increase influenza vaccination at upcoming ap-
2008; Patel & Volpp, 2012; Loewenstein et al., 2007; pointments, which resulted in a 5% average increase
Halpern et al., 2007). They tend to be low-cost and in relative vaccination rates (Milkman et al., 2021).
preserve the freedom to choose, making them po- This exemplifies the potential usefulness of nudges.
tentially attractive tools for healthcare systems as Yet, healthcare is multifaceted and complex, with
they attempt to overcome suboptimal behaviors— innumerable areas where suboptimal behaviors
typically by patients and providers—to improve prevent the execution and delivery of healthcare
value of care, all the while facing pressures to min- practices at their full potential. Aside from some
imize wasteful spending, including on low-value guidance suggesting areas where the value of care
services (Berwick & Hackbarth, 2012; Porter, 2010). delivered to patients can be improved—e.g., the
Nudges avoid mandating or “shoving” people to- Choosing Wisely Campaign for a reduction in un-
ward specific choices in a complex environment necessary care (Cassel & Guest, 2012)—and state-
where a single option will often not be appropriate ments of “best practices” that, despite their pop-
for or acceptable to everyone. ularity and implied authority, can be surprisingly
Nudge units in healthcare, such as the Geising- uninformed by evidence, there is often little sys-
tematic guidance for nudge units to decide (1) how
* Corresponding author: [email protected]. to prioritize among problem areas, much less to

Behavioral Economics Guide 2021 94

Amir Goren et al. Crowdsourcing Nudges: Insights and Experiences From Two Healthcare Systems

identify (2) where feasible behavior-change oppor- the crowdsourcing tournaments shared the follow-
tunities exist and (3) what forms the corresponding ing goals: (1) They offered an opportunity to raise
interventions might take. awareness of the newly established nudge units
Crowdsourcing—the use of a large group of among frontline clinicians and staff, as well as sys-
self-nominated volunteer contributors to gather tem leadership (including surfacing potential col-
ideas, materials, or other information in response laborators); (2) they provided frontline clinicians
to a need or want—is a mechanism of collective in- and staff an opportunity to feel more engaged (and
telligence, which in this case refers to the phenom- hence have a voice) in how decisions were made re-
enon of groups performing a task better than ex- garding project prioritization; and (3) they helped
perts or the group’s own best-performing members identify potential opportunities for nudges to have
(Malone, 2018; Surowiecki, 2004; Galton, 1907). a meaningful impact on improving healthcare.
Crowdsourcing has been used to improve forecasts
of future events (e.g., Atanasov et al., 2016), discov- Setting Up a Nudge Crowdsourcing
er more efficient algorithms to solve computational Competition
problems (e.g., Boudreau & Lakhani, 2015), and to
unearth solutions to difficult engineering and sci- In any healthcare system, an important first step
entific challenges (Lakhani et al., 2013). There is in setting up a crowdsourcing effort is to get buy-in
also some evidence that it can bear fruit in health- from leadership and other relevant stakeholders,
care innovation (Terwiesch et al., 2013; Tucker et given the need for system-wide outreach to employ-
al., 2019; Ranard et al., 2014; Wang et al., 2020). ees. At Geisinger, the BIT convened a nudge crowd-
This article describes how two nudge units in- sourcing committee consisting of the BIT program
corporated crowdsourcing competitions within director and faculty co-directors, Geisinger’s chief
their respective healthcare systems. The Geising- innovation officer, the chairs of the Medicine and
er BIT, established in 2018 with a minimal team, Heart Institutes, and the associate vice president
sought to increase familiarity with different areas of Product Innovation. An important consideration
of the broader organization and to develop a wid- was to include leaders with sufficient clout and di-
er portfolio of projects than those initially identi- versity of perspectives that they could provide both
fied. Therefore, one of the BIT’s first priorities was actual and perceived institutional legitimacy for
to look to other stakeholders in the organization the competition, in addition to providing a critical
for suggestions. Following the lead of the PMNU, review of the submissions. The PMNU’s committee
which had undertaken a similar initiative in 2016, consisted of the PMNU director, Penn Medicine’s
the BIT developed and executed a crowdsourcing innovation director, UPHS’s chief innovation, med-
competition with the goal of learning what a di- ical, medical informatics, and medical information
verse set of healthcare workers spanning the en- officers, and other experts in behavioral science.
tire organization (and including clinical, research, Several key decisions revolve around the format
administrative, service, and various other roles) of a crowdsourcing competition, including but not
had to say about what interventions were needed. limited to the:
The PMNU had sourced its participants from all
members of the University of Pennsylvania Health • Competition. In the PMNU tournament, there were
System (UPHS), including physicians, nurses, staff, several rounds of a submission-based competi-
and students. The BIT also endeavored to estab- tion that ended with an in-person “Shark Tank”
lish a more extensive network of clinical and other pitch day, where the ten finalists presented their
stakeholders needed to form collaborative partner- ideas in an open forum with the PMNU advisory
ships to execute and eventually scale up successful board as judges. The BIT, given the busy sched-
projects. Therefore, crowdsourcing was anticipated ules and limited time of both healthcare workers
to be helpful in identifying the right people with and leaders on their committee, opted to judge
whom to connect. At both the BIT and the PMNU, only on the basis of the initial text submissions.

Behavioral Economics Guide 2021 95

Amir Goren et al. Crowdsourcing Nudges: Insights and Experiences From Two Healthcare Systems

More complex approaches, such as allowing indi- tee narrowed the pool of ideas to fewer than ten
viduals to augment, update, or revise their initial finalists. The BIT opted to have submissions fed
contributions after seeing what others have done directly to its team: (1) The four members of the
(Boudreau & Lakhani, 2015) may offer incremen- BIT categorized and independently rated the full
tal benefits, but here we discuss “one-stage” set of submissions on a 1–5 scale reflecting their
methods whereby individual ideas are gathered perceived relevance, feasibility, and importance;
and judged by the organizers of the competition. (2) the team agreed by consensus on only the
• Participants. One must decide whether the audi- most promising submissions to pass along to the
ence should be only internal to the organization— crowdsourcing committee; and (3) the committee
and whether that should include all employees or arrived at the final set of winners, based on a dis-
only those with certain roles—or also external, cussion of their feasibility, novelty over existing
allowing an even broader potential range of input. efforts, and importance to the system.
• Intake. Data collection for submissions can take
various forms, including an intake form built into Execution
a web page, a mobile application, or a survey. An
online survey tool is a straightforward way for Additional decisions revolve around the execution
participants to submit their ideas, and the sub- of the competition, including:
missions are then easily downloaded as a spread-
sheet of responses for categorization and analysis. • Timeline. The BIT and PMNU provided five weeks
• Incentives. Rewards for prolific contributors or and three weeks, respectively, for people to sub-
creators of the best ideas or solutions can spur mit ideas, in order to create some sense of ur-
increased participation and better-quality sub- gency while also allowing time to formulate and
missions. Rewards for participation itself can also submit suggestions. Timelines were also created
be helpful in contexts such as healthcare, where and communicated for winner selection and prize
employees may be busy and not in the habit of dissemination.
responding to such requests for their input. One • Intake design. It can be tempting to solicit a high
must also be aware of legal and regulatory con- degree of detail from participants, but given their
straints on prize types and limits. The BIT’s top presumed lack of time or familiarity with nudges,
three winners were given a choice from among it is important to follow nudge principles in de-
three fun and/or practical devices (e.g., a smart signing the intake (e.g., survey) itself, ensuring
speaker), while two runners-up and the remain- that it:
ing eight finalists received gift cards. The PMNU’s º Is easy to fill out, short, and engaging (to en-
finalist idea submitters could choose from among courage broad participation).
gift cards to local attractions and establishments º Is selective in terms of scope (to discourage
or a fitness tracker. Both the BIT and PMNU se- submissions that do not lend themselves to
lected at random, among participants with a valid nudge solutions).
submission, one who would receive a grand prize. º Encourages participants to think through
This was intended to further motivate participa- the problems they identify in such a way that
tion among those skeptical that they might win. submissions on average reflect true areas of
• Judging the winners. One must decide whether to need rather than just half-baked proposals
have winner selections determined by peer re- or irrelevant personal interests (i.e., not just
view (e.g., by other submitters or by employees a “suggestion box”); both the BIT and PMNU
in the broader system), clinical stakeholders, made it optional to propose solutions, in or-
the nudge unit, or some combination thereof. der to minimize burden on submitters and to
The PMNU opted for an approach wherein peers avoid missing out on important problem ar-
could comment on and rate the proposals of other eas where the nudge units could themselves
UPHS members anonymously. Then, the commit- suggest solutions.

Behavioral Economics Guide 2021 96

Amir Goren et al. Crowdsourcing Nudges: Insights and Experiences From Two Healthcare Systems

• Intake dissemination. Summary of Submissions

º Communication. Potential participants should
be educated briefly about nudges, given rele- At Geisinger, there were 291 unique problems
vant examples thereof, and provided an easy (with or without solutions) provided by 260 partic-
way to access the competition. Both the BIT ipants. At Penn Medicine, there were 225 submis-
and PMNU opted for email as the primary sions from 192 participants.
modality for distribution, sending currently BIT members, using the 1–5 rating scale, iden-
eligible recipients a description of the com- tified 35 contenders for “finalist” status, prior to
petition and a link to the competition website. bringing these to the broader crowdsourcing com-
º Advertising. Broad advertising, with repeated mittee. These contenders hailed from 27 depart-
messages over time, helps increase aware- ments and represented at least ten distinct prob-
ness of the competition and ensure input lem areas. The 13 winners/finalists represented 12
from diverse areas of the system. Emails per- departments, and their submissions included, in
sonally forwarded from leaders of various de- descending order from the three winners to the two
partments can provide additional legitimacy runners-up to the eight finalists:
and authority to the requests for input, as can
online presentations and question-and-an- 1. Increasing flu vaccinations by providing a $1
swer sessions for the organization. It is es- scratch-off lottery ticket incentive
pecially important to rely on official organ- 2. Reducing follow-up appointment no-shows by
ization-based sources for communications, giving patients an appointment card with a photo
given the prevalence of unsolicited and pos- and a short bio of the provider
sibly malicious communications in modern 3. Improving medication refills with text reminder
industrial settings. automated response options
• Categorizing the entries. Labeling and categorizing 4. Increasing Geisinger’s Medicare Advantage in-
entries into a manageable set is a helpful step to- surance plan enrollment by indicating potential
ward judging the entries, as it allows for the quick savings on healthcare invoices
identification of duplicates and inappropriate 5. Reducing unnecessary blood count testing by
submissions, as well as the easy ranking of entries changing default order search results
of the same type. This further enables a descrip- 6. Improving appointment rescheduling via auto-
tive analysis and reporting of the submissions, mated text response options when patients cancel
both winning and non-winning. During intake, 7. Increasing yearly well visit appointments by
both the BIT and PMNU asked participants to cat- inviting next-year scheduling at appointment
egorize their submissions (e.g., improving care checkout
coordination, decreasing a low-value service, 8. Increasing employee hand-washing by applying
or increasing a high-value service). Ultimately, brightly colored stickers to sanitizer dispensers
however, both teams used an inductive approach 9. Increasing online patient portal and mobile
to categorize all entries, based upon the actual application enrollment via default patient invi-
frequency and range of the response types re- tations upon electronic health record (EHR) en-
ceived. rollment
• Results dissemination. Finally, when the compe- 10. Increasing patient adherence to health main-
tition concludes, one must decide whether and tenance screenings via reminders enabling auto-
how to release the results, which can help inform mated ordering and scheduling
future proposals and determine ongoing engage- 11. Reducing inappropriate emergency room (ER)
ment with such efforts. visits via a triage mobile application
12. Reducing inappropriate ER visits and delays via
a digital dashboard showing ER and nearby urgent
care facility wait times

Behavioral Economics Guide 2021 97

Amir Goren et al. Crowdsourcing Nudges: Insights and Experiences From Two Healthcare Systems

13. Increasing timely physical therapy (PT) re- 4. Improving access to medications by defaulting
ferrals and reducing unnecessary opioid use via prescriptions to Penn outpatient pharmacies
primary care default referrals to PT rather than a 5. Improving adherence and reducing costs via au-
specialist. tomatic 90-day mail order supply for established
medication prescriptions
The PMNU competition included 321 ratings of 6. Increasing long-access reversible contraception
submissions provided by 57 people, with 24 people prescription and use in the immediate postpar-
providing a total of 66 comments. The submissions tum period
of the three winners and six runners-up were, in 7. Reducing unnecessary computerized tomogra-
descending order from the winners: phy scans for suspected pulmonary embolisms
by using the Wells Criteria
1. Reducing inappropriate opioid prescribing in the 8. Increasing pharmacist-managed inpatient anti-
emergency department by changing default set- coagulation monitoring to improve dosing con-
tings sistency across patients
2. Improving cardiac rehabilitation referral rates 9. Reducing appointment no-shows via mobile ap-
following inpatient cardiovascular events by us- plication scheduling reminders
ing default options
3. Reducing unnecessary imaging for radiation Beyond the winners, the BIT identified 20 rela-
therapy in palliative care via order set changes tively well-defined and prevalently used categories

Figure 1: Geisinger BIT crowdsourcing themes. Note: Presented are the percentages of unique submissions (n =
291), broken down by problem type.

that accounted for all submissions (see Figure 1). and efficiency were the most prevalent and ac-
Topics related to scheduling, care quality/coordi- counted for a majority of submissions (57.5%).The
nation of care, the alleviation of inappropriate ER PMNU identified several submission themes (see
use, employee satisfaction, patient satisfaction, Figure 2). Individual submissions could be reflected

Behavioral Economics Guide 2021 98

Amir Goren et al. Crowdsourcing Nudges: Insights and Experiences From Two Healthcare Systems

Figure 2: PMNU crowdsourcing themes. Note: Presented are the percentages of unique submissions (n = 225),
broken down by theme.

in more than one theme, so the percentages shown partments and stakeholders to see how and where
do not sum to 100. The most common theme was the proposed interventions could complement ex-
improving satisfaction, followed by better leverag- isting initiatives.
ing technology and optimizing transitions in care. The BIT reflected on occasional challenges to
Proposed nudge solutions (n = 204) included alerts implementation: (1) conflicts with similar inter-
and reminders (35.8%), developing ways to enable ventions already underway, (2) lack of availability
choice (27.0%), changing default settings (23.0%), of appropriate stakeholders for collaboration, and
using incentives or disincentives to motivate (3) the prohibitive cost and/or timing of imple-
behavior (7.8%), and restricting or eliminating menting relatively complex (e.g., technological)
choices altogether (6.4%). solutions, such as integrating mobile app use into
the EHR or making significant changes to EHR and/
Follow-up or operations programming. The BIT had to adjust
expectations and study designs to comport with
The Geisinger BIT notified the winners, sent out current system priorities and to focus on feasible
prizes, and worked with Marketing to create an in- novel approaches. They advocated for randomized
ternal online article describing the competition and controlled trials wherever possible, but occasion-
nudges, revealing the winners and prizes, and out- ally only a pre- vs post-intervention design was
lining the nine broad categories that the 13 winning feasible or a retrospective comparison against a
submissions represented. The BIT also worked with propensity-matched control group.
the crowdsourcing committee to prioritize among Still, several BIT projects were either directly
winning submissions which projects to undertake. or indirectly launched as a function of the crowd-
Several entries were investigated further, reaching sourced submissions and are currently well under-
out to submitters to see if they wanted to collabo- way or in the process of being implemented, thanks
rate, and communicating with the appropriate de- to connections forged with interested parties across

Behavioral Economics Guide 2021 99

Amir Goren et al. Crowdsourcing Nudges: Insights and Experiences From Two Healthcare Systems

Geisinger: e.g., Clinical Informatics, Pharmacy, Geisinger is structured as a more conventional

the Geisinger Health Plan, and others. The sub- healthcare system, albeit one that is also a health
missions included department and role informa- insurer for its members. These differences no doubt
tion that helped educate the team regarding what affected the range of participant roles—and hence
types of problems were encountered or tackled by types of submissions—in the two competitions.
what types of stakeholders, so the BIT knew where For example, the inclusion of Geisinger Health Plan
to turn to for help. Stakeholders also reached out employees likely allowed for more submissions
spontaneously to the team, having learned about focused on insurance-related issues and access to
the competition and the nudge unit running it. care. Other systems should consider the inclusive-
The PMNU’s original competition has resulted ness of their own participant pools with respect to
in published interventions: (1) doubling guide- the range of issues they would like to explore.
line-concordant prescribing from 21% to 43% by The BIT and PMNU both made the nudge sugges-
changing EHR default opioid prescribing quanti- tions optional, so the BIT was pleasantly surprised
ties (Delgado et al., 2018), (2) quintupling cardiac that only six participants (2.1%) had no solutions
rehabilitation referrals from 15% to 85% via a de- to offer, and similarly, the PMNU competition only
fault opt-out EHR decision pathway (Adusumalli had 21 participants (9.7%) fail to include a nudge
et al., 2021), and (3) halving unnecessary imaging intervention. On the other hand, it was not too sur-
from 68% to 32% during palliative radiotherapy for prising to see—in spite of the brief definitions and
adults with cancer via order set changes (Sharma et examples of nudges provided to participants—that
al., 2019). submissions also included solutions that went be-
yond the typical scope of a nudge and seemed more
Discussion appropriate for the “suggestion box”, e.g., in-
creased funding, added facilities, technological tool
We cannot quantify the precise return on invest- overhauls, and the development of new smartphone
ment of the crowdsourcing competition, intellec- apps. Finally, a high proportion of submissions were
tually (e.g., in terms of whether the competition aligned with key system priorities, perhaps reflect-
yielded a higher quantity or quality of interventions ing the impact of communications from leadership
than other potential approaches), logistically, or and general awareness of common issues of con-
financially (e.g., in terms of whether the cost of cern across healthcare settings.
the prizes and the time spent were offset by sub-
sequent gains). Subjectively, however, we conclude Conclusion
that crowdsourcing was a great benefit to the nudge
units, helping identify (1) innovative nudge solu- Two nudge units, namely, the Behavioral Insights
tions, (2) important problem areas in healthcare Team at Geisinger and the Penn Medicine Nudge
that can benefit from nudge solutions, and (3) con- Unit, both found crowdsourcing to be a valuable way
nections to important potential partner stakehold- for their institutions to reveal important problems
ers, as well as (4) increasing unsolicited subsequent and novel nudge interventions that might solve
outreach from others in the organizations who have them. We recommend a structured competitive
become more familiar with the nudge units thanks approach, sampling from the frontlines of care, to
to the exposure. The submissions, while represent- other healthcare systems and institutions looking
ing a relatively small percentage of total potential to accomplish the same. Several important consid-
participants, were nonetheless rich and diverse, erations are detailed above, but for the effort to be
thereby providing plenty of projects to consider. At successful, we emphasize in particular the need for
the PMNU, crowdsourcing has become an annual buy-in from system leadership and stakeholders
part of the process to engage frontline clinicians who will play a critical role in dissemination and in
and staff. the final selection and prioritization of projects.
Penn Medicine is an academic health center, while

Behavioral Economics Guide 2021 100

Amir Goren et al. Crowdsourcing Nudges: Insights and Experiences From Two Healthcare Systems

The Authors using the electronic health record.

Amir Goren, PhD, is Program Director of the Be- Christopher Chabris, PhD, is a Professor at
havioral Insights Team (BIT) at Geisinger, design- Geisinger, where he co-directs the Behavioral and
ing and evaluating interventions to improve patient Decision Sciences Program, and is also faculty
outcomes and reduce system costs. He received his co-director of the Behavioral Insights Team. His
PhD in Psychology from Princeton University, co- research focuses on attention, intelligence (indi-
authoring a Science article with Alexander Todorov vidual, collective, and social), behavior genetics,
on the prediction of election outcomes from face- and decision-making. He received his Ph.D. in Psy-
based inferences of competence and serving as chology and A.B. in Computer Science from Harvard
teaching assistant for Kahneman and Shafir’s JDM University, and he is a Fellow of the Association for
course. He received a BS in electrical engineering Psychological Science. Chris is the co-author of the
from the University of Maryland, College Park. Amir New York Times bestseller The Invisible Gorilla: How
conducted postdoctoral work at NIH, studying work Our Intuitions Deceive Us, which has been published
complexity and intellectual functioning over time in 20 languages to date. He shared the 2004 Ig Nobel
in diverse populations, and the Yale Rudd Center for Prize in Psychology (awarded for “achievements
Food Policy and Obesity, testing communications that first make people laugh, and then make them
to shift public perceptions of unhealthy food mar- think”), given for the scientific experiment that
keting to children. Before joining the BIT, Amir was inspired the book. Chris has spoken to audienc-
a Senior Director at Kantar Health, consulting and es at major conferences and businesses, includ-
publishing health outcomes research incorporating ing PopTech, Google, Credit Suisse, and Procter &
patient-reported outcome measures to assess pa- Gamble, and his work has been published in leading
tient-centered needs and experiences. journals, including Science, Nature, PNAS, Psycho-
logical Science, Perception, and Cognitive Science.
Mitesh S. Patel, MD, MBA, is Director of the Penn
Medicine Nudge Unit, the world’s first behavioral Michelle Meyer, PhD, JD, is an Assistant Professor
design team embedded within a health system. He at Geisinger, where she co-directs the Behavioral
is the Ralph Muller Presidential Associate Profes- and Decision Sciences Program and is also faculty
sor of Medicine and Health Care Management at co-director of the Behavioral Insights Team. She
the Perelman School of Medicine and The Wharton is also Associate Director for Research Ethics at
School at the University of Pennsylvania. Dr. Patel is Geisinger, chairs Geisinger’s IRB Leadership Com-
on the leadership team at the Penn Medicine Center mittee, and directs its Research Ethics Advice and
for Health Care Innovation, Associate Director at Consulting Service. In her research, she investi-
the Center for Health Incentives and Behavioral gates judgments and decision-making related to
Economics, a staff physician at the Crescenz VA research, innovation, and healthcare, especially
Medical Center in Philadelphia, and a Senior Fellow ethically- or legally-relevant judgments and de-
at the Leonard Davis Institute of Health Economics. cisions such as the acceptability of nudges, AI, and
His research focuses on combining insights from randomized controlled trials. Her writing has ap-
behavioral economics with scalable technology peared in leading journals of bioethics (American
platforms to improve health and healthcare. He has Journal of Bioethics, Hastings Center Report, Kennedy
led more than 25 clinical trials in partnership with Institute of Ethics Journal), law (Harvard Law Review,
health systems, insurers, employers, and commu- Administrative Law Review), medicine (New England
nity organizations that tested ways to design nudg- Journal of Medicine, JAMA Network Open) and science
es, incentives, and gamification, in order to change (Science, Nature, PNAS), as well as in popular media
clinician and patient behavior. This work includes outlets (New York Times, Slate, Wired, Los Angeles
digital health interventions using wearable devices Times, and Forbes.com). She has served on numer-
and smartphones, and health system interventions ous boards and commissions, including National

Behavioral Economics Guide 2021 101

Amir Goren et al. Crowdsourcing Nudges: Insights and Experiences From Two Healthcare Systems

Academies study committees and working groups, Berwick, D. M., & Hackbarth, A. D. (2012). Eliminating
an American Psychological Association blue ribbon waste in US health care. JAMA, 307(14), 1513-1516.
commission, and the editorial board of Advances in Cassel, C. K., & Guest, J. A. (2012). Choosing wise-
Methods and Practices in Psychological Science. She ly: Helping physicians and patients make smart
earned a Ph.D. in religious studies, with a focus on decisions about their care. JAMA, 307(17), 1801-
applied ethics, from the University of Virginia and 1802.
a J.D. from Harvard Law School, where she was an Delgado, M. K., Shofer, F. S., Patel, M. S., Halpern,
editor of the Harvard Law Review. Following law S., Edwards, C., Meisel, Z. F., & Perrone, J. (2018).
school, she clerked on the U.S. Court of Appeals for Association between electronic medical record
the Eleventh Circuit. She graduated summa cum implementation of default opioid prescription
laude from Dartmouth College. quantities and prescribing behavior in two emer-
gency departments. Journal of General Internal
Acknowledgments Medicine, 33(4), 409-411.
Galton, F. (1907). Vox populi. Nature, 75, 450-451.
The Geisinger Behavioral Insights Team acknowl- Halpern, S. D., Ubel, P. A., & Asch, D. A. (2007).
edges Henri Carlo Santos and Maheen Shermoham- Harnessing the power of default options to
med for their critical review, categorization, and improve health care. New England Journal of
evaluation of crowdsourcing contenders, and the Medicine, 357(13), 1340-1344.
crowdsourcing committee (Karen Murphy, Edward Lakhani, K. R., Lifshitz-Assaf, H., & Tushman, M.
J. Hartle, Alfred S. Casale, and Rebecca A. Stametz) L. (2013). Open innovation and organization-
for help with evaluation and selection of winners. al boundaries: Task decomposition, knowl-
The Penn Medicine Nudge Unit thanks the following edge distribution and the locus of innovation.
steering committee members for reviewing crowd- In A. Grandori (Ed.), Handbook of Economic
sourcing contenders: PJ Brennan, Bill Hanson, Su- Organization (pp. 355-382). Edward Elgar.
san Day, Roy Rosin, Judd Kessler, Kevin Volpp, and Loewenstein, G., Brennan, T., & Volpp, K. G. (2007).
David Asch. Asymmetric paternalism to improve health
behaviors. Journal of the American Medical
References Association, 298(20), 2415-2417.
Malone, T. W. (2018). Superminds: The surprising
Adusumalli, S., Jolly, E., Chokshi, N. P., Gitelman, Y., power of people and computers thinking together.
Rareshide, C. A. L., Kolansky, D. M., & Patel, M. Little, Brown.
S. (2021). Referral rates for cardiac rehabilitation Milkman, K. L., Patel, M. S., Gandhi, L., Graci,
among eligible inpatients after implementation H. N., Gromet, D. M., Ho, H., Kay, J. S., Lee,
of a default opt-out decision pathway in the T. W., Akinola, M., Beshears, J., Bogard, J. E.,
electronic medical record. JAMA Network Open, Buttenheim, A., Chabris, C. F., Chapman, G. B.,
4(1), e2033472. https://doi.org/10.1001/jama- Choi, J. J., Dai, H., Fox, C. R., Goren, A., Hilchey,
networkopen.2020.33472. M. D., … Duckworth, A. L. (2021). A megastudy
Atanasov, P., Rescober, P., Stone, E., Swift, S. A., of text-based nudges encouraging patients to
Servan-Schreiber, E., Tetlock, P., Ungar, L., get vaccinated at an upcoming doctor’s ap-
& Mellers, B. (2016). Distilling the wisdom of pointment. Proceedings of the National Academy
crowds: Prediction markets versus prediction of Sciences, 118(20), e2101165118. https://doi.
polls. Management Science, 63(3), 691-706. org/10.1073/pnas.2101165118.
Boudreau, K. J., & Lakhani, K. R. (2015). “Open” dis- Patel, M. S., Day, S. C., Halpern, S. D., Hanson, C.
closure of innovations, incentives and follow-on W., Martinez, J. R., Honeywell, S., & Volpp, K. G.
reuse: Theory on processes of cumulative inno- (2016). Generic medication prescription rates
vation and a field experiment in computational after health system-wide redesign of default op-
biology. Research Policy, 44(1), 4-19. tions within the electronic health record. JAMA

Behavioral Economics Guide 2021 102

Amir Goren et al. Crowdsourcing Nudges: Insights and Experiences From Two Healthcare Systems

Internal Medicine, 176(6), 847-848. happiness. Yale University Press.

Patel, M. S., & Volpp, K. G. (2012). Leveraging in- Tucker, J. D., Day, S., Tang, W., & Bayus, B. (2019).
sights from behavioral economics to increase the Crowdsourcing in medical research: Concepts
value of health-care service provision. Journal of and applications. PeerJ, 7, e6762. https://doi.
General Internal Medicine, 27(11), 1544-1547. org/10.7717/peerj.6762.
Patel, M. S., Volpp, K. G., & Asch, D. A. (2018). Nudge Wang, C., Han, L., Stein, G., Day, S., Bien-Gund, C.,
units to improve the delivery of health care. New Mathews, A., Ong, J. J., Zhao, P.-Z., Wei, S.-F.,
England Journal of Medicine, 378(3), 214-216. Walker, J., Chou, R., Lee, A., Chen, A., Bayus, B., &
Patel, M. S., Volpp, K. G., Small, D. S., Wynn, C., Zhu, Tucker, J. D. (2020). Crowdsourcing in health and
J., Yang, L., Honeywell, S., & Day, S. C. (2016). medical research: A systematic review. Infectious
Using active choice within the electronic health Diseases of Poverty, 9(1). https://doi.org/10.1186/
record to increase physician ordering and pa- s40249-020-0622-9.
tient completion of high-value cancer screening
tests. Healthcare, 4(4), 340-345.
Porter, M. E. (2010). What is value in health care?
New England Journal of Medicine, 363(26), 2477-
2481.
Ranard, B. L., Ha, Y. P., Meisel, Z. F., Asch, D. A., Hill,
S. S., Becker, L. B., Seymour, A. K., & Merchant,
R. M. (2014). Crowdsourcing—Harnessing
the masses to advance health and medicine: A
systematic review. Journal of General Internal
Medicine, 29(1), 187-203.
Sharma, S., Guttmann, D., Small, D. S., Rareshide,
C. A. L., Jones, J., Patel, M. S., & Bekelman, J. E.
(2019). Effect of introducing a default order in
the electronic medical record on unnecessary
daily imaging during palliative radiotherapy
for adults with cancer: A stepped-wedge cluster
randomized clinical trial. JAMA Oncology, 5(8),
1220-1222.
Sharma, S., Guttmann, D., Small, D. S., Rareshide,
C. A. L., Jones, J., Patel, M. S., Bekelman, J. E.
(2019). Effect of introducing a default order in
the electronic medical record on unnecessary
daily imaging during palliative radiotherapy
for adults with cancer: A stepped-wedge cluster
randomized clinical trial. JAMA Oncology,5(8),
1220-1222.
Surowiecki, J. (2004). The wisdom of crowds.
Doubleday.
Terwiesch, C., Mehta, S. J., & Volpp, K. G. (2013).
Innovating in health delivery: The Penn Medicine
innovation tournament. Healthcare (Amsterdam,
Netherlands), 1(1-2), 37-41.
Thaler, R. H., & Sunstein, C. R. (2008). Nudge:
Improving decisions about health, wealth, and

Behavioral Economics Guide 2021 103

 Applying Behavioural Economics to Improve
Hearing Aid Conversations Between First-Time
Clients and Clinicians: The Development and
Evaluation of Interventions to Increase
Informed Choice

Megan Gilliver,* Elizabeth F. Beach and Brent Edwards

National Acoustic Laboratories

For the first-time audiology client, considering which hearing aid to choose can be complex. Clinicians,
too, struggle with balancing how to approach conversations and support clients. Default, rather than in-
formed, decision-making means clients can miss out on beneficial experiences provided by premium-level
hearing aids. This project aimed to tackle the problem by identifying and addressing cognitive biases that
may be present during hearing aid discussions. Interventions developed to target these biases (including
anchoring, confirmation bias, self-fulfilling prophecy and cognitive overload) were trialled and evaluated
over a three-month period. Clinicians responded favourably to the intervention, and the results showed,
compared to business-as-usual centres, an increase in the proportion of interested clients accessing pre-
mium-level hearing aids.

The final common pathway for the application of a hearing appointment is an important first step.
nearly every advance in medicine is human be- When a hearing loss is diagnosed, access to hearing
havior. aids relies on the clinician accurately conveying the
—Patel et al. (2018, p. 214). relevant information in an accessible way, and the
client deciding to proceed with a hearing aid fitting.
As for other areas of health, human behaviour However, even at this point, having made the nec-
plays a critical role in hearing healthcare. Significant essary decisions to seek help and acquire a hearing
scientific advances have improved our understand- aid, there is yet another decision to be made – which
ing of the hearing system and the technologies used hearing aid to choose? It is this choice, and the con-
for hearing assessment and rehabilitation, and yet text in which it occurs, that forms the focus of this
long-term hearing health outcomes rely largely on paper.
the choices made by individuals. For the first-time
audiology client,1 the decision to book and attend Hearing Aid Choices

1 Throughout this paper, we use the term client when re- Hearing aid choices are presented, considered and
ferring to audiology patients, clients and consumers. We made within the therapeutic environment of the
use clinician when referring to audiologists and other pro- client-clinician dynamic. Many clinicians report
fessionals providing hearing health services. We use hear- some level of difficulty in deciding how and when
ing aids and devices interchangeably. information about hearing aids should be shared
with clients. They cite clients’ clinical test results
* Corresponding author: [email protected]. and preferences as key influencers for how they

Behavioral Economics Guide 2021 104

Megan Gilliver et al. Applying BE to Improve Hearing Aid Conversations Between First-Time Clients & Clinicians

tailor the presentation of options (Boisvert et al., all HSP-eligible clients. Most providers also offer
2017). Whilst test results are reasonably objective, a range of premium-level hearing aids. These are
preferences are not. Pre-existing beliefs and un- only partially subsidised, and therefore the client
conscious biases are likely to be shaping both client is required to make a contribution to cover the re-
preferences and also the clinician’s ability to accu- maining cost.
rately perceive and respond to those preferences. The choice of either a base- or a premium-level
Previous work has noted the influence of potential hearing aid can be influenced by a range of client
cognitive biases on clinicians’ decision-making. and clinical factors. A recent review found that, on
, but no peer-reviewed studies have examined in- average, approximately one-third of clients select-
terventions that aim to reduce their impact (Galvin ed premium-level products over base-level hear-
et al., 2020), and very little research has examined ing aids. However, across the 20 largest providers
the effect of client biases. The current research uses in Australia, this figure varies from 10% to 70%
a new approach in hearing health research – the (PricewaterhouseCoopers Australia, 2017), sug-
development and evaluation of interventions, built gesting that the processes, materials and approach
on BE principles and applied to the problem of how taken by service providers has a strong influence on
best to support clients to make an informed choice clients’ decision-making.
about which hearing aid is the most appropriate for
them. The Problem – Are Clients Making an
Hearing aids vary in both form and function, and Informed Choice?
most companies offer a range of devices, from base-
to premium-level products containing the most The research was initiated following a request
recent technology available, based on continuous from a large hearing healthcare service provider to
research and development. This increasingly so- examine their HSP-eligible clients’ behaviour when
phisticated technology is aimed at providing more accessing hearing healthcare for the first time, and
comfortable, clear and easy listening experiences, to identify the key influences on decision-making.
and technological advances to hearing aids have Only approximately 20% of their first-time clients
been associated with increasing user satisfaction were accessing premium-level hearing aids, which
(Bishop & Eby, 2010; Picou, 2020). Clients are en- is considerably lower than the proportion of experi-
couraged to choose hearing aids best suited for their enced/returning clients (with uptakes closer to the
needs, considering not only experiences offered by industry average).
different technology (e.g., sound quality, connec- Most HSP clients will not be re-eligible for sub-
tivity, rechargeability, performance in noise), but sidised hearing aids for a number of years, adding
also shape, size and – ultimately – price. to the impact of their choices. The sponsor was
concerned that compromised decision-making
A Special Context: The Australian Hearing was influencing new clients to default to “free”
Services Program base-level choices and miss out on the benefits
associated with premium-level hearing aids. They
In Australia, many clients make their hearing aid wished to identify and ameliorate cognitive biases
choice within the context of a government-fund- (in both clinicians and clients) acting as barriers to
ed Hearing Services Program (HSP). Administered
through registered hearing healthcare providers,
the program provides subsidised hearing services 2
We will continue to use the term base-level to describe
and hearing aids to eligible clients. Providers usu- the fully-subsidised hearing aids, as they are lower-fea-
ally offer a range of different devices for clients to tured devices. However, as they must meet minimum
choose from. All offerings must include at least one quality standards set by the HSP, some may have ad-
device that is fully subsidised by the HSP , which
2
ditional features to some other commercially available
ensures access to a nocost hearing aid option for base-level devices.

Behavioral Economics Guide 2021 105

Megan Gilliver et al. Applying BE to Improve Hearing Aid Conversations Between First-Time Clients & Clinicians

informed choice, so as to better support their first- were modified to include items for first-time cli-
time HSP clients to choose the right hearing aid for ents (eligible for hearing aids). Twenty-five clients
their needs. completed these additional items asking about
their experience of the appointment, including
Phase I: Identifying Roadblocks to conversations about hearing aid choice and their
Informed Choice beliefs about base- vs. premium-level hearing aids.
The client survey included some items that closely
Method mirrored items in the clinician surveys, thereby en-
abling comparisons of the perceptions held by the
Data collection two groups.
Before attending their first appointment,
Appointment observations. Members of the re- semi-structured interviews were conducted with
search team observed first-time clients’ hearing 17 “pre-clients” (who had not yet attended an ap-
appointments, to understand clinicians’ approach- pointment but were present at a hearing screening
es to discussing hearing aid choices, clients’ re- activity). The interviews focused on understanding
sponses and the overall dynamic between the two clients’ knowledge and beliefs about hearing aids
parties. when presenting at their first appointments.
Clinicians. Unstructured interviews were held
with experienced senior-level clinicians and clin- Results
ical trainers to understand better the context in
which hearing aid choices are made. Information Analysis of the surveys, interviews and focus
gathered from these conversations informed the groups revealed three key factors that were ad-
development of online surveys and focus groups. versely affecting client hearing aid decision-mak-
Surveys asked questions about clients’ uptake of ing:
premium-level hearing aids and perceived benefits,
factors influencing client choices and the current 1. Client beliefs around price and value: The impact
appointment process. Eighty-four clinical staff of anchoring and zero-price effects
responded to the survey, with 70% completing all
items. The cost of hearing aids was one of the most fre-
Two focus groups (one metropolitan and one re- quently cited concerns by clients and clinicians.
gional) were held with 16 on-the-ground clinicians. Whilst available base-level devices are effectively
Groups discussed how hearing aid choices were zero-cost for HSP-eligible clients (and frequently
presented to clients, conversation pain points, and described as “free” by marketers and clinicians),
opportunities for improvement. Clinicians also par- the client contribution for partially subsidised
ticipated in “persona-creation” activities in which devices can reach over $5000 for premium-level
they were asked to consider the characteristics hearing aids. Appointment observations revealed
of two fictional clients – one who would typically that many clients were unprepared for these pric-
pursue premium-level hearing aids, and another es, openly showing surprise or seeking clarification
who would not do so. This provided an opportunity when presented with this information. Clinicians
to collect clinicians’ professional insights into fac- reported that these reactions created a barrier to
tors influencing clients’ decision-making, whilst further conversations. Clients (and clinicians) also
simultaneously providing the research team with described base-level hearing aids in a qualitatively
insights regarding factors influencing clinicians’ different way to premium-level devices, using the
choices of when and how to have these conversa- perceived low-risk zero client contribution as a sig-
tions. nificant benefit.
Clients. Existing online client surveys (distributed Interviews with pre-clients further highlighted
to all clients the day following their appointment) the low levels of awareness about pricing held prior

Behavioral Economics Guide 2021 106

Megan Gilliver et al. Applying BE to Improve Hearing Aid Conversations Between First-Time Clients & Clinicians

to appointments. When asked to estimate the worth 2. The impact of clinicians’ beliefs on behaviour:
of a pair of base-level hearing aids under the HSP, Confirmation bias leads to a self-fulfilling
their responses ranged from $60 to $200. prophecy
Roadblock 1: These data make it clear that first-
time clients’ beliefs about hearing aid costs were Appointment observations revealed high variabil-
inaccurately anchored. As a result, the client contri- ity between clinicians in the way hearing aid choic-
bution required for premium-level devices was per- es were presented to clients. Most approached the
ceived as disproportionately high, leading to neg- conversation about premium-level products with
ative emotions and hindering further discussion. hesitancy. Clinicians agreed they often spent rela-
Furthermore, the zero-price effect of base-level tively little time discussing choices. In particular,
hearing aids created inertia in clients that was par- they would often change the focus of the conversa-
ticularly resistant to attempts to encourage them to tion or not pursue the discussion with clients they
consider the higher-priced premium-level options. considered resistant. Clinicians interpreted clients’
questions about cost as a sign of concern and an in-
dicator of their likely disinterest in premium-level
devices. Some clinicians were observed to actively
deflect questions from clients about premium-level

Figure 1: Clinicians’ responses to the question asking, “What percentage of your eligible clients do you think...?”

hearing aids, instead responding with reassurances feedback in focus groups, where they expressed
about the efficacy of the available base-level devic- frustration about their clients’ unwillingness to
es. discuss premium-level hearing aids and lamented
This behaviour was in contrast to the clinicians’ the benefits that the clients were foregoing as a re-

Behavioral Economics Guide 2021 107

Megan Gilliver et al. Applying BE to Improve Hearing Aid Conversations Between First-Time Clients & Clinicians

sult. When surveyed, clinicians were positive about um-level hearing aids are ultimately not choosing
the value of premium-level hearing aids (Figure 1), them.
believing that many of their clients would benefit A potential explanation may be related to a dif-
from them. However, they believed that less than ferent gap – captured by the relative difference in
half of clients were interested in premium-level reported beliefs for A & B. Focus group clinicians
hearing aids – and that less than a quarter would openly shared their hesitancy in discussing pre-
choose them. mium-level hearing aids with clients perceived as
The difference between B and C in Figure 1 above disinterested – behaviour that was also observed
(18%) potentially estimates the size of the inten- in appointments. Thus, the number of clients likely
tion-action gap at the heart of this work. It shows to receive information about premium-level hear-
that almost half of clients whom clinicians believe ing aids is only a proportion (potentially two out of
to be interested in accessing the benefits of premi- three) of those who clinicians believe stand to ben-

Figure 2: Clinician and client beliefs about premium- vs. base-level products.

efit. ever, clients’ responses to corresponding survey

When asked about the benefits of premium-level items present an alternate story. Just like their cli-
hearing aids relative to base-level devices, clini- nician counterparts, they, too, held positive beliefs
cians’ responses showed they strongly believed in about the benefits of premium-level hearing aids.
the advantages of the former but considered their However, they reported perceiving clinicians as be-
clients were less convinced of their superiority (see ing the party less convinced about the superiority of
Figure 2). This aligns with other clinician feedback, these options.
in that they generally believed clients to be sceptical Roadblock 2: These data showed evidence of a
about the value of premium-level options. How- self-fulfilling prophecy resulting from confirma-

Behavioral Economics Guide 2021 108

Megan Gilliver et al. Applying BE to Improve Hearing Aid Conversations Between First-Time Clients & Clinicians

tion bias. Clinicians believe clients are disinterest- tions around pricing. Clinicians’ responses involve
ed in, or unable to afford, premium-level hearing focusing attention on reassuring clients about the
aids, leading them to interpret clients’ questions benefits of base-level options in an attempt to re-
as a signal of their inability or hesitancy to consider assure them by downplaying any limitations of
these options. This is more pronounced when cli- pursuing base- compared to premium-level op-
ents’ responses include surprise or negative emo- tions. Clients subsequently perceive clinicians to

Figure 3: Elements of the hearing appointment and associated pain points.

be equivocal about the benefits of premium-level Clinicians reported wanting to remain neutral
hearing aids, which reduces their likelihood to pur- when presenting information, to maintain clients’
sue them. Their decision further confirms the cli- autonomy, for instance, ‘I don’t want to come across
nicians’ original biased interpretation around client as pushy or salesy’. In surveys, clinicians reported
disinterest, thereby perpetuating the cycle. believing that clients overwhelmingly preferred
to receive information and make their own deci-
3. Presentation of choice: A cognitive overload sion rather than follow advice from the audiologist
(Figure 4). In contrast, clients’ survey responses
Throughout the appointment, clients are required to a similar item showed that less than a quarter
to engage in tasks with high emotional and/or cog- strongly identified as wanting to make independent
nitive load (examples shown in Figure 3). Clients decisions, with the majority having a strong prefer-
expend considerable cognitive energy throughout ence for clinical advice.
the appointment, recalling salient hearing expe- Although clinicians believed that clients preferred
riences and managing emotions whilst processing to assess information about different hearing aids
instructions and new information. There are low for themselves, their survey responses indicated
cognitive resources available to process the infor- concerns about clients’ ability to do so (Figure 5).
mation presented in order to choose the best hear- Clinicians acknowledged the large volume of infor-
ing aid for their needs. mation to be processed and its complexity, noting
In observations, choice presentation was incon- their clients often struggled, e.g., ‘So many aid op-
sistent between clinicians and between clients. tions, style options, platforms, features and accessories

Behavioral Economics Guide 2021 109

Megan Gilliver et al. Applying BE to Improve Hearing Aid Conversations Between First-Time Clients & Clinicians

to consider. If it is overwhelming for audiologists, it can occurs at a time of maximal decision fatigue
only be more so for clients’. • The costs associated with hearing aids and the
Roadblock 3: These data reveal that the context of relatively short timeframe of appointments con-
hearing aid decision-making contains many factors tributes to decision-task difficulty
associated with choice overload: • The choice-set is complex – there are multiple
levels of devices, many with their own set of op-
• Multiple cognitive demands throughout the ap- tions.
pointment mean that the choice of hearing aid

Figure 4: Distribution of responses (clinicians vs. client) regarding beliefs about a preference for autonomy and
professional advice when making hearing aid decisions.

Faced with this overload, and in fear of making the hearing aid pricing: to reduce the impact of inertia
wrong choice, it is perhaps unsurprising that clients caused by the zero-price effect.3
turn to what they (and, often, their clinicians) see 2. Restructuring the presentation of hearing aid
as a safer “default” choice, i.e. the fully subsidised choices (particularly within appointment discus-
base-level hearing aid. sions) by developing new processes and accompa-
nying tools: to reduce choice overload for clients
Phase II: Intervention (whilst maintaining autonomy) and to counter
clinician behaviours that lead to a self-fulling
Method prophecy.
3. Provision of information to clients prior to ap-
Procedures pointments: to help them prepare for hearing
aid conversations and reduce cognitive overload
Interventions were developed to address these during the appointment. Also designed to provide
three roadblocks and increase clients’ ability to an accessible introduction to hearing aid choic-
make informed choices about hearing aids.
Four broad changes were made to client appoint- 3
NB: Pricing changes also included adjustments to ad-
ments: dress additional identified barriers not detailed here.
Although they are tangential to the stated aim, they are
1. Adjustment to the presentation of premium-level mentioned here for completeness.

Behavioral Economics Guide 2021 110

Megan Gilliver et al. Applying BE to Improve Hearing Aid Conversations Between First-Time Clients & Clinicians

es (using the revised pricing and presentation Study design

structures outlined in 1 and 2) and more realistic
anchors for hearing aid costs. Interventions were trialled over a three-month
4. Clinician training: to increase clinicians’ period to ensure sufficient data were collected. Due
awareness of, and ability to address, the impact to sponsor constraints about the availability of cen-
of confirmation bias and information overload tres and clinicians, only two experimental groups
influencing decision-making. were possible within the available geographical

Figure 5: Overall, how confident do you think your clients are when differentiating between the benefits and
features of different hearing aids? (10=very confident).

area during this time. As a result, the decision non-participating hearing healthcare centres.
was made to evaluate the effects of all changes Clinicians in the full-implementation group were
(‘full-implementation’; 1-4) against a group im- surveyed about the intervention’s impact, and we
plementing the “pricing-only” intervention (1). also monitored the sponsor’s regular client satis-
Both groups had equivalent numbers of first-time faction survey responses for both groups.
HSP-eligible clients proceeding with a hearing aid
fitting. Results
To assess the success of the interventions in each
group, we measured the change in percentage up- Device uptake
take of premium-level hearing aids compared to
baseline data from the nine months prior to the The change in uptake of premium-level hearing
intervention trial. We compared data from the pric- aids was an increase of only 1% for both the control
ing-only group vs. the full-implementation group group and the pricing-only group. In contrast, the
and compared both of these to a control group of full-implementation group recorded an increase of

Behavioral Economics Guide 2021 111

Megan Gilliver et al. Applying BE to Improve Hearing Aid Conversations Between First-Time Clients & Clinicians

Uptake Rates (%)

Increase from baseline
Baseline Trial period

Control 21 22 5%

Pricing-only 15 16 7%

Full-implementation 17 34 100%

Table 1: Percentage of first-time HSP-eligible clients selecting premium-level hearing aids at baseline and
throughout intervention trial.

17% – effectively doubling the proportion of clients be the main barrier to clients choosing a premi-
selecting premium-level devices (see Table 1). um-level hearing aid, our results showed that pric-
ing changes alone had virtually no impact on their
Clinician and client feedback behaviour. This finding supports previous work
showing that although cost may influence decisions
Clinicians showed high levels of engagement to acquire hearing aids in the first place, it does not
throughout the trial period, and there was a 60% appear to be the key driver when choosing between
completion rate for the survey. Clinicians in the different options (Ramachandran et al., 2011). Our
full-implementation group rated interventions 2 results suggest that hearing aid choices are more
and 3 as most important for improving informed heavily influenced by the context in which they are
choice for clients. All clinicians reported that during presented, including the interaction between client
the trial, they approached hearing aid discussions and clinician.
with their clients differently, with all but one (who As health professionals, audiologists are mo-
recorded a neutral response) reporting that they tivated by a desire to provide client-centred care
considered the changes had made a positive impact. grounded in a strong evidence base. The current
Clients’ responses to the sponsor’s satisfaction study found that for many clinicians, this approach
survey remained positive throughout the trial pe- means they see their role as a neutral provider of
riod. Data analysis showed that respondents’ sat- information to clients, who should be left to make
isfaction ratings were at ceiling prior to the com- their own decision. The problematic nature of this
mencement of the trial, and they remained there approach has been noted previously: ‘This behav-
throughout. iour… appears to shift decision-making control
and responsibility to the client… and does not ac-
Discussion knowledge the importance of providing decisional
support to patients throughout the decision-mak-
Hearing aid users generally keep their hearing ing process’ (Boisvert et al., 2017, p. 9). The clients
aids for several years, which underlines the impor- surveyed in the current study also agree, with many
tance of first-time clients making the best choice, expressing a desire for “more” from the health
right from the start. This project identified some of professional whose assistance they have sought.
the critical client and clinician biases that operate
in the context of hearing aid decision-making and Limitations
designed a suite of interventions to address biases
and reduce the proportion of clients opting for the The main limitation of this project was that mul-
“default” choice. tiple interventions could not be trialled concurrent-
Although pricing was considered by many to ly. As a result, the outcomes reflect the impact of the

Behavioral Economics Guide 2021 112

Megan Gilliver et al. Applying BE to Improve Hearing Aid Conversations Between First-Time Clients & Clinicians

entire package of interventions, and it is not pos- hearing-related behaviours, including leisure noise
sible to determine the relative effectiveness of in- exposure and its impact on hearing health, the de-
dividual elements. Further work would be required velopment of strategies for encouraging safe lis-
to determine how implementing a smaller set, or tening for staff and patrons and motivating young
individual elements, would affect the results. adults to protect their hearing. She also worked on
projects related to improving the delivery of hearing
Conclusion health services and streamlining decision-making
processes for clients with hearing loss, using be-
A key part of this work involved identifying how havioural insights.
cognitive biases influence not only the behaviour
of individual clients and clinicians, but also the cli- Brent Edwards, Ph.D., is Director of the National
ent-clinician dynamic. Clinicians, like many health Acoustic Laboratories (NAL), where he is current-
professionals, are constantly processing informa- ly leading new innovation initiatives that focus on
tion to decide how to proceed “in the moment.” transforming hearing healthcare. For over 22 years,
Like clients, they, too, are susceptible to misper- he headed research at major hearing aid companies
ceptions and missteps as a result of biased percep- and Silicon Valley start-ups that have developed
tions and beliefs. Understanding and addressing the innovative technologies and clinical tools used
client-clinician dynamic is important for improv- worldwide. Dr. Edwards founded and ran the Star-
ing the quality of conversations and related deci- key Hearing Research Center in Berkeley, Califor-
sion-making. This is true not only for hearing, but nia, which was a leading site for research in hearing
also for many healthcare decisions. Designing and impairment and cognition. Dr. Edwards is a Fellow
implementing processes that target specific client of the Acoustical Society of America and an Adjunct
behaviours and also address the underlying needs Professor at Macquarie University.
of clinicians and clients are likely to be successful
in bringing about lasting change in clinical practice Acknowledgments
and long-term benefits for health outcomes.
The authors would like to thank Taegan Young
The Authors (NAL) for her expert advice on all things audiolo-
gy-related throughout this work, and her contribu-
Megan Gilliver, Ph.D., is a Research Scientist in tions to intervention development and implemen-
NAL’s Behavioural Sciences team. Since joining NAL tation. We also acknowledge the contribution of the
in 2007, she has worked on a wide variety of hearing sponsor, particularly their clinicians and clients
loss prevention and hearing rehabilitation projects. upon whose time and input this work was based.
With a background in psychology, Megan predom-
inantly spends her time considering those at the References
heart of the research, the humans – on both sides of
the otoscope. Megan’s current research focuses on Bishop, C. E., & Eby, T. L. (2010). The current status
understanding why people do (and do not!) engage of audiologic rehabilitation for profound unilat-
with hearing behaviours. She is interested in how to eral sensorineural hearing loss. Laryngoscope,
apply research from health promotion and behav- 120(3), 552-556.
ioural economics to improve the hearing health of Boisvert, I., Clemesha, J., Lundmark, E., Crome, E.,
individuals and the broader community. Barr, C., & McMahon, C. M. (2017). Decision-
making in audiology: Balancing evidence-based
Elizabeth Beach, Ph.D., is the former Head of the practice and patient-centered care. Trends in
Behavioural Sciences Department at the Nation- Hearing, 21, 1-14.
al Acoustic Laboratories, Australia. Her research Galvin, K. L., Featherston, R. J., Downie, L. E., Vogel,
covered a range of issues relevant to hearing and A. P., Hamilton, B., Granger, C., & Shlonsky, A.

Behavioral Economics Guide 2021 113

Megan Gilliver et al. Applying BE to Improve Hearing Aid Conversations Between First-Time Clients & Clinicians

(2020). A systematic review of interventions

to reduce the effects of cognitive biases in the
decision-making of audiologists. Journal of the
American Academy of Audiology, 31(2), 158-167.
Patel, M. S., Volpp, K. G., & Asch, D. A. (2018). Nudge
units to improve the delivery of health care. New
England Journal of Medicine, 378(3), 214-216.
Picou, E. M. (2020). MarkeTrak 10 (MT10) survey
results demonstrate high satisfaction with and
benefits from hearing aids. Seminars in Hearing,
41(1), 21-36.
PricewaterhouseCoopers Australia. (2017). Review
of services and technology supply in the Hearing
Services Program.
Ramachandran, V., Stach, B. A., & Becker, E. (2011).
Reducing hearing aid cost does not influence
device acquisition for milder hearing loss, but
eliminating it does. Hearing Journal, 64(5), 10-18.

Behavioral Economics Guide 2021 114

 BI Org: Fostering a Behavioural Mindset
at the OECD
Faisal Naru and Kate Laffan*
OECD

The OECD is hard at work fostering a behavioural mindset across the organisation. In the current paper, we
detail the 4 Ps behind our efforts: the principles, practices, proof and phases. In doing so, we aim to provide
a blueprint for other organisations looking to promote management excellence through the application of
behavioural insights.

Introduction oped and rolled out behaviourally informed work-

streams in several corporate areas, from recruit-
The past decade has seen exponential growth in ment to cybersecurity, to responsibilities of staff
the application of behavioural insights to public and beyond. Some of these projects have emerged
policy challenges. To date, both academic research from applying a behavioural lens to examining
and the world of practice have largely focused their ways to improve the delivery of services within the
efforts on issues around policy implementation and organisation. Others have come about through re-
on targeting individual, and in particular consumer, flecting on evidence emerging from the behaviour-
behaviour change. In other words, work has been al science literature and the evolving needs of the
directed at the “micro-level.” As the field matures, organisation’s stakeholders.
increasing attention is being paid to the potential of The application of behavioural insights is well-
behavioural insights to speak to meso-level issues, aligned with the OECD’s scientific, data-driven and
including asking the question: how can behaviour- human-centred approach to understanding and
al insights contribute to organisational change addressing the operational challenges that arise
(Gavetti, 2012; Stingl, & Geraldi, 2017; Shepard, in the running of the organisation. Their use also
2017; OECD, 2017; OECD, 2020)? demonstrates the OECD’s commitment to ‘walk the
This question has been front of mind at the OECD. talk’ and lead in implementing good practices that
Josée Touchette – the OECD’s Executive Director – it recommends others adopt (OECD, 2017). With this
is committed to innovating operations to meet the article, we seek to provide a blueprint for other in-
needs of the organisation’s stakeholders, now and stitutions looking to draw on behavioural insights
in the future. She considers behavioural insights an to bring about organisational change.
invaluable part of the organisation’s toolkit with In what follows, we detail the 4 Ps behind our
which to achieve this goal. effort to foster a behavioural mindset at the OECD,
namely principles, practices, proof and phases.
The use of behavioural insights married with data These four elements capture the current thinking
visualisation and strategic thinking is essential in and behavioural work going on within the OECD’S
modern management. Executive Directorate, as well as plans going
-Josée Touchette forward.

Over the past several years, the OECD has devel- 1. Principles

* Corresponding author: [email protected] Setting out clear guiding principles ensures

Behavioral Economics Guide 2021 115

Faisal Naru & Kate Laffan Fostering a Behavioural Mindset at the OECD

not only that opportunities to apply behavioural example, regulations or financial structures – and
insights are systematically identified, but also that behavioural insights can inform the design and
projects are ultimately aligned with the organisa- implementation of these instruments, too (see
tion’s goals. The behavioural work within the OECD Carbon Pricing in the Proof section).
is guided by the following principles:
2. Practices
• Recognise that people are human. We apply a be-
havioural lens to the challenges that we face. This While many of the features of the OECD’s behav-
involves identifying behavioural barriers to de- iourally informed work vary significantly, depend-
sired outcomes, as well as potential solutions. This ing on factors such as the area of focus, the time-
approach stands in contrast with process-driven line for delivery and the scope for intervention and
approaches that do not properly account for indi- evaluation, many practices remain constant. These
viduals’ motivations, both intrinsic and extrinsic, practices represent suggestions for others looking
or their behavioural biases and tendencies. to embed behavioural insights into their organisa-
• Be cognisant of our own limitations and co-pro- tion.
duce knowledge. We are mindful of our own biases
and limited information. We do not fool ourselves • Identify the important behavioural and structural
that we are objective or that we automatically components of the issues that arise. The solution
know what is driving an issue or how best to solve to many operational challenges lies in behaviour
it. Instead, we seek to co-produce knowledge change. At the same time, it is important to recog-
with our stakeholders, adopting a 360° approach nise that individuals operate within constraints,
in which people at all levels of the organisation including those relating to time, resources and
have the opportunity to feed in their experiences, know-how. Often, structural constraints need to
opinions and suggestions. As a team, we encour- be addressed to make behaviour change possible.
age a healthy challenge function that recognises • Seek out quality data and use data visualisation
and addresses assumptions. tools to understand and communicate operation-
• Align work with existing organisational princi- al challenges. Operational data, such as on the
ples and goals. Our behaviourally informed work number of emails sent or flights booked etc., is
is aligned with the organisation’s ethos. This in- often a valuable but underexplored resource that
cludes upholding ethical standards, safeguarding can provide insights.
organisational integrity and promoting social, • Look to insights from state-of-the-art litera-
economic and environmental sustainability with- ture in behavioural science and the opinions of
in the organisation, among other things. It is also experts in the field to inform work. There exists
important to demonstrate organisational em- a rich body of behavioural science literature that
pathy, which the behavioural and social sciences can help inform organisational practices, and
can help instil through the use of appropriate lan- academics are often keen to partner with organ-
guage, messages and timely actions. These high- isations to carry out translational work, as well as
er-order goals not only help shape the issues we new research.
look at through a behavioural lens, but also act as • Demonstrate and communicate impact. Robust
the litmus test for the solutions we propose. evaluation strategies help identify what works, as
• Expand beyond traditional “nudging” when well as what needs more work. Clear behavioural-
considering how to shape behaviour. Nudges ly informed communication strategies help build
undoubtedly have a role to play in encouraging awareness of the work and ensure transparency
behaviour change within organisations, including among stakeholders. Importantly, insights from
at the OECD (e.g. see Cybersecurity in the Proof behavioural science can act as valuable inputs
section). In many cases, however, organisations into decision-making, even when evaluation is
will have harder policy instruments in place – for not possible.

Behavioral Economics Guide 2021 116

Faisal Naru & Kate Laffan Fostering a Behavioural Mindset at the OECD

• Adopt an iterative approach. Organisations typi- consisted of 1) a standard message that flagged the
cally operate in dynamic environments and have prevalence of phishing attacks; 2) a message that
to balance the needs of multiple stakeholders. It made salient the correct courses of action by high-
is important to improve on and adapt behavioural lighting them in red at the beginning of the email; 3)
efforts continually as new information emerges a message that emphasised the recipient as having
and contexts change, always keeping the organ- been personally identified as a target for phishing
isation’s overarching goals in mind. attacks and 4) a message presenting a metaphor
involving credit card details, to emphasise the risk
3. Proof of phishing attacks.
The impact of the messages on both clicking and
Within the OECD, behavioural insights have fed reporting was evaluated, using employee responses
into efforts to understand and shape both organi- to fake phishing attacks that came from the digital
sational policies and practices and the behaviour of security team. The results indicate that the personal
our employees while at work. and risk-based messages increased reporting by
The proof, as they say, is in the pudding. In what 150% on average compared to the control group,
follows, we present a snapshot of four examples with the risk being the more effective of the two
of how behavioural insights have been applied at strategies. The evaluation also identified 12% of
the OECD. These examples relate to cybersecurity, staff as ‘risky clickers’, in that they both click and
delegations of authority, hiring and carbon pricing. do not report. This group will be the focus of further
Importantly, these examples do not represent one- work on this topic.
off projects but rather ongoing streams of work.
Delegations of Authority
Cybersecurity
Delegations of authority are formal means for
Like all large organisations, the OECD is subject officials within the OECD who have been entrusted
to phishing attacks. On average, it experiences 300 with the responsibility of a budget by the Secre-
attacks a month. While the organisation employs tary-General to share this responsibility with other
state-of-the-art technical cybersecurity measures, officials. These delegations apply to a wide range of
OECD staff remain the first line of defence against tasks, including authorising travel for guests, ac-
data breaches. To optimise cybersecurity, employ- cepting delivery for payment and approving budget
ees should avoid opening and clicking on phishing commitments.
attachments and links, and also report the attacks In 2018, the OECD introduced an e-Delegation
internally to the digital security team. tool for use across all of the directorates and pro-
The OECD’S executive directorate has explored a grammes. It is an automated, paperless, transpar-
range of behaviourally informed communications ent process with organisation-wide standard text
to promote these behaviours. For example, in an and templates, thus providing a clear audit trail.
email-based field experiment, OECD staff received Entering a delegation of authority into the system
communications from the digital security team triggers an electronic request for acceptance by the
which indicated the correct actions to take when selected staff member. It is vital to the organisa-
they suspect they have been targeted by a phishing tion that staff members in receipt of these requests
attack. They were randomly allocated either to re- carefully engage with the details of the DoA and in
ceive one of four messages or be part of a control particular are cognisant of the responsibility they
group that did not receive any communications on are taking on when accepting the DoA and subse-
the topic. quently while at work.
The treatment messages were informed by pre- Research from the behavioural sciences empha-
vious work on risk communication (Sopory, 2017) sises the benefits of personalising communications,
and email interventions (Service et al., 2014). They as well as making important information salient

Behavioral Economics Guide 2021 117

Faisal Naru & Kate Laffan Fostering a Behavioural Mindset at the OECD

(OECD, 2017). Other work indicates that gamifica- examining the influence of the physical and profes-
tion can improve engagement in learning activities. sional appearance of candidates on hiring decisions
In this regard, multiple-choice quizzes have been and evaluating the potential of virtual reality train-
shown to enhance learning, engagement and en- ing, which mimics OECD interviewing conditions,
joyment (Cheong et al., 2013). to support inclusive hiring practices (Mobius &
Drawing on these behavioural insights, the Rosenblat, 2006).
OECD’S Executive Directorate has developed an in-
teractive and visual tutorial explaining the nature of Carbon Pricing
the DoAs and the responsibilities involved, as well as
a quiz that assesses learning and provides feedback The OECD implemented an internal carbon price
on staff members’ understanding. Staff members in 2013. Under this initiative, directorates are
who are assigned to DoAs receive personal invites charged a flat rate of 30€ per tonne of CO2 emitted
to engage with these materials, to help them under- by air travel on missions. The funds from the initi-
stand and be aware of what they are agreeing to. ative are used to purchase carbon offsets and fund
projects related to improving the organisation’s
Hiring environmental performance. With the help of this
initiative, the OECD achieved carbon neutrality in
Existing research in the behavioural sciences has 2020. However, mission flights still account for a
identified important sources of bias in hiring, in- large share of the organisation’s pre-offset carbon
cluding in interview processes, with implications footprint, so there is room to further enhance the
for the diversity of hired candidates (Bohnet, 2016). sustainability of the OECD’s travel behaviour.
In 2020, the OECD’S Executive Directorate spear- While this initiative involves a hard policy in-
headed an initiative to identify the best way to as- strument aimed at reducing flying behaviour, there
sess candidates at the interview stage, in order to are lessons from behavioural science that can speak
ensure fairness within the process. to both how this price is structured and how it is
In line with this goal, the OECD’S Executive Di- communicated to staff. For example, research from
rectorate carried out a field experiment to explore the internal carbon pricing system at Yale Univer-
the relative impact of conducting video vs. phone vs. sity identified that a redistributive carbon price, in
transcript interviews. This exploration was devel- which revenues from the scheme were redistrib-
oped based on the insight from behavioural science uted from poor-performing to well-performing
that blinding recruiters to candidates’ demographic buildings, was the most effective structure among
characteristics can work to debias hiring processes a range of different strategies (Gillingham et al.,
(Goldin & Rouse, 2000). The primary outcome of 2017). Other work indicates that the way carbon
interest in the field experiment was the scores given prices are communicated matters, too. Tactics such
to each candidate across the different media types, as normative messaging, indicating what the rev-
and the analysis explored how these scores varied enue from the taxes will be used for, and inducing
based on the candidates’ gender, nationality and personal responsibility for carbon emissions all
language skills. aid in increasing support for and the effectiveness
The results indicate that female candidates scored of the policy (Hurlstone et al., 2014; Bristow et al.,
similarly to men across all three mediums (video, 2010; Ünal, Steg & Granskaya, 2019).
audio, transcript). At least at the OECD, therefore, Having reviewed the existing behavioural science
we have evidence to suggest that the interview for- literature on carbon pricing, as well as patterns
mat does not appear to promote the hiring of one around pre-pandemic flying behaviour from our
gender over another. This “null” finding emphasis- operational data, the OECD’s Executive Directorate
es the importance of testing behavioural insights in is in the process of exploring a selection of these
context. insights – in the context of the OECD’s carbon price
Ongoing projects which build on this work are – in both qualitative interviews and contingent

Behavioral Economics Guide 2021 118

Faisal Naru & Kate Laffan Fostering a Behavioural Mindset at the OECD

choice experiments with key stakeholders. Looking ing internally in the office of the Executive Director
forward, these exercises will feed into the organisa- but has drawn on the skills of communications and
tion’s carbon pricing policy. data visualisation experts and involved interested
and willing staff members from across the organ-
4. Phases isation. Involving people with a range of comple-
mentary skills, and interested partners, has allowed
The workstreams highlighted in the Proof section us to apply behavioural insights successfully to
represent the first phase in a long-term strategy to strategically important areas, as well as effectively
promote and mainstream a behavioural mindset communicate the value-added of this approach. It
across the organisation. The work carried out to also helped create behavioural champions, and in
date has been led by behavioural scientists work- doing so it has begun to embed behavioural think-

Figure 1: Fostering a behavioural mindset at the OECD: A phased approach.

ing across different areas of the organisation. improving decision clarity, especially in an ev-
In phase two of the mainstreaming strategy, the er-complex, dynamic and uncertain environment.
Executive Director’s office will act as an incubator, Clinical consciousness of one’s own decisions, as
playing a supporting role for the bottom-up be- well as clarity in the understanding of those deci-
havioural workstreams that are emerging across sions by stakeholders, especially those affected, is
the organisation. This work largely stems from a critical part of management excellence. In Phase
those employees who contributed to the Phase 1 3, we will leverage BI to promote decision clarity
workstreams, many of whom now regularly apply a throughout the organisation. This will include, for
behavioural lens to understanding their own oper- example, designing the choice criteria for strategic
ational challenges and have gained the confidence decisions and the processes by which those deci-
to lead on a workstream of their own. Key tasks for sions are made, as well as clearly communicating
the team in the Executive Director’s office in Phase the decisions that are being made, by whom and
2 include supporting these people to identify stra- how.
tegically important research questions and design
rigorous evaluations, as well as helping, where nec- Conclusion
essary, to interpret the results.
To leverage BI further in the promotion of opera- Through adhering to our principles and engaging
tional excellence, Phase 3 will see the OECD’s Exec- in the core practices, we have developed a series
utive Directorate’s behavioural work move further of behavioural projects at the OECD, as well as a
upstream, asking: How can behavioural insights be roadmap for future work. Plenty of questions and
used to help support strategic decision-making in opportunities remain, but we are well on the way to
the organisation? The projects involved in Phase 1 fostering a behavioural mindset inside the organi-
and Phase 2 largely focus on using behavioural in- sation. We would encourage other organisations to
sights to provide clarity on the problem, communi- consider doing likewise and to help lead the way.
cation and policy implementation. Phase 3 is about

Behavioral Economics Guide 2021 119

Faisal Naru & Kate Laffan Fostering a Behavioural Mindset at the OECD

Figure 2: Our working definition of decision clarity.

The Authors Guo, Nick Johnstone, Eva Knop, Katrina Koser,

Trish Lavery, Makoto Miyasako, Guillermo Mo-
Faisal Naru heads the office of the Executive Di- rales, Kevin O’Brien, Sandra Ragg, Dan Roulstone,
rector and the behavioural insights work in the Dimitra Staikou, Victoria Romero Tejedor, Ana José
Executive Directorate. Previously, he founded the Varela and Josée Touchette.
OECD’s work on behavioural insights and is re- Particular recognition goes to Jun Nakagawa and
sponsible for a number of OECD publications in this Francesca Papa for providing the behavioural ex-
area. He serves on a number of international com- pertise and implementation during the foundation-
mittees and advises governments and public bodies al steps of this journey and experimentation.
on strategy, governance and implementation.
References
Dr Kate Laffan is a Marie Skłodowska-Curie Fel-
low in the School of Economics and the Geary In- Bohnet, I. (2016, April 18). How to take the bias out
stitute for Public Policy at University College Dublin of interviews. Harvard Business Review. https://
and a Behavioural Science Fellow in the Executive hbr.org/2016/04/how-to-take-the-bias-out-
Directorate of the OECD. In her work, Kate develops of-interviews/.
and applies behavioural insights to promote sus- Bristow, A. L., Wardman, M., Zanni, A. M., &
tainable lifestyles and organisational policies. Chintakayala, P. K. (2010). Public acceptability
of personal carbon trading and carbon tax.
Acknowledgements Ecological Economics, 69(9), 1824-1837.
Cheong, C., Cheong, F., & Filippou, J. (2013). Quick
The opinions expressed and arguments employed quiz: A gamified approach for enhancing learn-
herein do not necessarily reflect the official views ing. PACIS 2013 Proceedings, 206. https://aisel.
of the member countries of the OECD. The authors aisnet.org/pacis2013/206.
would like to acknowledge the cutting-edge work Gavetti, G. (2012). Perspective—Toward a behavio-
and valuable inputs across multi-disciplinary ral theory of strategy. Organization Science, 23(1),
teams that contributed to the work in this paper/ 267-285.
article and in forming a behavioural mindset at Gillingham, K., Carattini, S., & Esty, D. (2017).
the OECD. In addition to the authors, these include Lessons from first campus carbon-pricing
Ingrid Barnsley, María José Cuesta, Anna De Saint scheme. Nature News, 551(7678), 27-29.
André, Leonie Decrinis, Audrey Depeige, Junyang Goldin, C., & Rouse, C. (2000). Orchestrating im-

Behavioral Economics Guide 2021 120

Faisal Naru & Kate Laffan Fostering a Behavioural Mindset at the OECD

partiality: The impact of” blind” auditions on

female musicians. American Economic Review,
90(4), 715-741.
Hurlstone, M. J., Lewandowsky, S., Newell, B. R., &
Sewell, B. (2014). The effect of framing and nor-
mative messages in building support for climate
policies. PloS One, 9(12), e114335. https://doi.
org/10.1371/journal.pone.0114335.
Mobius, M. M., & Rosenblat, T. S. (2006). Why beau-
ty matters. American Economic Review, 96(1),
222-235.
OECD. (2017). Behavioural insights and public policy:
Lessons from around the world. OECD Publishing.
https://doi.org/10.1787/9789264270480-en.
OECD. (2020). Behavioural insights and organisations:
Fostering safety culture. OECD Publishing. https://
doi.org/10.1787/e6ef217d-en.
Service, O., Hallsworth, M., Halpern, D., Algate, F.,
Gallagher, R., Nguyen, S., Ruda, S. & Sanders, M.
(2014). EAST: Four simple ways to apply behaviour-
al insights. https://www.bi.team/publications/
east-four-simple-ways-to-apply-behaviour-
al-insights/.
Shepard, D. (2017). Applying behavioural insights
to organisations: Global case studies. OECD and
European Commission.
Sopory, P. (2017). Metaphor in health and risk
communication. Oxford Research Encyclopedia of
Communication. https://doi.org/10.1093/acre-
fore/9780190228613.013.304.
Stingl, V., & Geraldi, J. (2017). Errors, lies and mis-
understandings: Systematic review on behav-
ioural decision making in projects. International
Journal of Project Management, 35(2), 121-135.
Ünal, A. B., Steg, L., & Granskaya, J. (2019). “To
support or not to support, that is the ques-
tion.” Testing the VBN theory in predicting
support for car use reduction policies in Russia.
Transportation Research Part A: Policy and Practice,
119, 73-81.

Behavioral Economics Guide 2021 121

 Complex Behavioral Challenges Require Multi-
Faceted Behavioral Solutions: Driving Change in
Sustainable Agriculture
Philipe M. Bujold* and Erik Thulin
Center for Behavior & the Environment, Rare

Behavioral science has been criticized for developing solutions that are only tractable for easy challenges,
and even then, generally have small impacts. However, this is not a limitation of the behavioral science
approach per se but rather of the common narrow focus on nudges. Instead, we propose that intensive,
multi-layered behavioral science solutions are necessary for addressing what might otherwise be seen as
intractable behavioral shifts. To demonstrate this point, we showcase the development of the Lands for Life
program in Colombia, designed to increase both the environmental and the financial sustainability of rural
farming. This challenge is of particular interest, due to both its significance and difficulty, as the target
behavior is nothing short of community-wide shifts in farmers’ fundamental livelihoods. Built from the
ground up to incorporate behavioral insights, the program draws on the theory behind the key phenomena
of social proof to overcome ambiguity aversion, as well as social pressure to drive pro-social behavior. In
this article, we present a case study analysis of how this multi-stage intervention incorporates these and
other behavioral insights in a three-phase process which, instead of focusing on the individual, focuses on
community-wide changes that deliver transformative results more effectively than traditional nudges.

Introduction limitations of our traditional approach and allow

for the tackling of more multifaceted problems
The behavioral sciences have been increasingly (Nesterak, 2020; Sanders et al., 2018; Spencer, 2018;
criticized for targeting the “low-hanging-fruits” Van Der Linden, 2018).
of behavior change while achieving only marginal We propose that the next phase of behavioral sci-
effects (Spencer, 2018). As a result of this focus, be- ence-driven programming will break away from
havioral science has been slow to tackle problems the singular focus on choice architecture through
that require more than marginal shifts in behavior, marginal nudges, moving instead towards the in-
ignoring some of the most pressing, wicked, or tensive, multi-faceted, and behaviorally-informed
simply just complex problems of our time. solutions necessary for addressing what might
But why is this the case, and how can the field otherwise be seen as intractable behavioral shifts.
move past targeting these “low-hanging fruits” Yet, simply calling for such a shift does not make
to address more systemically complex challenges it so. Nor does it offer a blueprint for what this
such as rising inequality, healthcare, and even cli- approach will look like. In this article, we present
mate change? While identifying clear and reliable the Lands for Life (L4L) program, designed by the
marginal shifts were perhaps necessary in the early environmental behavior change non-profit Rare,
days of applied behavioral science to demonstrate which strategically applies staged and layered be-
the applicability of key strategies, the field has now havioral change interventions by using a variety
matured sufficiently to be able to learn from the of behavioral levers to go beyond norm messaging
and into the realm of norm shifts. Through this case
* Corresponding author: [email protected] study, we show how a precise and organized set of

Behavioral Economics Guide 2021 122

Philipe Bujold & Erik Thulin Complex Behavioral Challenges Require Multi-Faceted Behavioral Solutions

strategies can be combined to help a ‘social proof’ high return on investment (but see De Ridder et al.,
program snowball into one that leverages a growing 2020). Take, for example, Allcott and Rogers finding
constituency of sustainable farmers to create ‘social that, while energy-use social comparisons reduced
pressure’ on others to do the same—ultimately cre- energy consumption by less than 2%, it was so in-
ating a self-enforcing norm in which farmers, their expensive to administer that it remains one of the
communities, and the climate stand to collectively most cost-effective strategies available for utilities
benefit. to reduce demand (2014).
These cost-efficient and non-coercive nudges
The Limits of One-Size-Fits-All Nudges ought to be applauded for their meaningful effect
on the low-hanging fruits of behavior change, but
According to Thaler and Sunstein, nudging is what about those cases where a 1.4% shift in be-
“choice architecture that alters people’s behavior in havior is simply not enough to achieve the nudger’s
a predictable way without forbidding any options or goal? Considering that many nudges aim to address
significantly changing their economic incentives” energy consumption, pro-environmental behav-
(2009, p.6), thereby crafting a choice environment iors, or the provision of healthcare, such effects
to encourage behavior change without altering the might prove insufficient (Hummel & Maedche,
value of—or access to—the options available. 2019); and while alternatives are often more costly,
The idea that behavior can be cheaply and mean- they can also be more effective. For example, a 2019
ingfully influenced by simple changes to a choice’s meta-analysis of 84 different behavior-change so-
“architecture” has proven extremely popular. It cial marketing campaigns highlighted how, while
would be no exaggeration to say that many have notably more expensive per adoption than nudging,
come to view and treat nudging as a silver bullet they are also more effective—shifting behaviors by
whereby reframing, reminding, and comparing are an average of 18.1 percentage points (Green et al.,
key factors in solving all our behavioral qualms. 2019).
Nevertheless, the explosion of nudge units across While nudges are often perceived as low-cost
the world has been accompanied by an explosion alternatives to more intensive strategies such as
of evidence highlighting the limits of relying on community-based social marketing, they nonethe-
small changes in choice architecture to create large, less all fall under the larger umbrella of behavior
long-lasting, and generally applicable change. change. As such, we see the bridging and layering of
these techniques as a new frontier to achieve more
The Relative Effect of Nudges on Behavior effective and efficient large-scale shifts in behavior.

While individual nudges tend to be cost-effective Maintaining Behavior Change in the Long Term
(Benartzi et al., 2017), it is worth noting that their
average effect on behavior can often be overestimat- Another area of concern on the effectiveness of
ed. A 2020 meta-analysis of nudge-based interven- individual nudges has been how their impact holds
tions found that the average nudge in the academic up over time. The few studies that measure their
literature rendered a shift of only 8.7 percentage long-run results find that the effects of nudges—
points. But even that is incredibly overstated, due whether repeated or not—tend to decay signifi-
to publication bias. Applying the same methods to cantly over time (for examples, see Ferraro & Price,
analyze the pre-registered analysis of hundreds of 2013; Damgaard et al., 2018; Hallsworth et al., 2016;
studies run by a nudge unit that were not subject to Venema et al., 2018; and the review in Brandon et
publication bias, they found an average effect of a al., 2017).
mere 1.4 percentage points. This does not mean, however, that no nudges see
That nudges have relatively small effects should their effects persist. Rather, it highlights how those
in no way preclude their use: They can often be that achieve continuous effects do so through rec-
deployed at low cost and on a large scale, giving a ognizing a fundamental behavioral insight: Human

Behavioral Economics Guide 2021 123

Philipe Bujold & Erik Thulin Complex Behavioral Challenges Require Multi-Faceted Behavioral Solutions

behavior is inherently social (Brandon et al., 2017). Beyond Individual Nudges

Despite much of applied behavioral science focus-
ing on the individual, people naturally observe, While many view the above limitations as inher-
internalize, and act according to the norms set by ent to the behavioral sciences, they arise because of
their peers. If a nudge results in people perceiving the historically narrow focus on choice architecture
that a given behavior is widespread and/or is the and nudging. When general practitioners, politi-
acceptable social norm, its effects are more likely cians, and members of the public think behavioral
to be long-lasting, because they shift entire groups science, they think peer comparisons, reminders,
to new equilibria (Chudek & Henrich, 2011). Beyond and framing—techniques that i) tend only to mar-
the relatively limited toolkit of norms-messaging, ginally shift behavior, ii) often do not self-reinforce
then, practitioners can bolster their interventions over time, and iii) treat all individuals the same—
by actively engaging with established social sys- and so are left wholly unequipped to deal with com-
tems or by generating entirely new social norms plex problems.
(Mols et al., 2015; Van Der Linden, 2018). Only so much change can be driven by “shooting
at low-hanging fruit” with nudges (Hansen, 2018,
People Are Diverse, and so Are Their Reactions to p.195). But academics and practitioners agree that
Nudges behavioral insights provide a much wider and rich-
er toolkit than nudging alone (Ewert, 2020). In this
A final critical criticism of the behavioral scienc- article, we showcase the L4L program, which ap-
es’ focus on nudging revolves around the reliance plies a sequential set of behavioral interventions to
on one-size-fits-all interventions deployed homo- a challenge that has been seen as so wicked that it
geneously across heterogeneous populations. has been largely left unaddressed: Shifting small-
Nudges deployed without a thorough under- scale producers away from an overreliance on agro-
standing of the population and the diverse interests chemical and towards a suite of more sustainable,
at play can reduce their effectiveness or—in a few less carbon-emitting practices.
cases—actually backfire for specific subpopula-
tions (for reviews, see Osman et al., 2020; and Sun- Lands for Life: Layering Interventions to
stein, 2017). Boomerang effects, for example, have Maximize Change
been documented in a variety of cases (Bacon &
Krpan, 2018; Byrne et al., 2018; Richter et al., 2018) In Colombia, 87% of farmers operate on less than
where an intervention causes a segment to reverse, two hectares of land. Together, they feed most of the
or move away, from the target behavior. People are country, and yet very few receive the support needed
different, and their reactions to nudges differ, too. to help bridge the 30-50% yield gap between their
This standard buckshot approach can be contrast- production and that of their larger, industrial-scale
ed with developing personalized or micro-targeted counterparts. As best practices change, smallholder
interventions that adapt or are selectively deployed farmers are left in the dark, and so many continue
to a subpopulation given their particular character- to farm using the practices with which they are fa-
istics (Beshears et al., 2015; Gravert & Kurz, 2019; miliar, even though these have led (and continue to
Metcalf et al., 2019a, 2019b). Alternatively, and as lead) to declining soil health, water pollution, and
is the case with the L4L project featured here, prac- the progressive undermining of agricultural pro-
titioners can combine and stagger complementary ductivity, especially as the climate along the Andes
strategies so that their effects ripple across popula- changes.
tion segments and ultimately snowball into some- Though this may seem like a particularly complex
thing much bigger (Spencer, 2018). behavioral problem, getting farmers to adopt cli-
mate-friendly practices remains just that: A prob-
lem of behavior. However, it is also a problem that
sole reliance on nudging—whether in the form of

Behavioral Economics Guide 2021 124

Philipe Bujold & Erik Thulin Complex Behavioral Challenges Require Multi-Faceted Behavioral Solutions

framing, comparing, or reminding—would fail to Bollinger, 2017; McDonald et al., 2020; Pickering et
address meaningfully. For one, the marginal effects al., 2019), slowly weaving individual behavioral in-
of a nudge spread over thousands of farmsteads terventions into a cohesive norm-change solution.
would have little impact at a landscape level. For Aided by local technical assistants, our program
positive environmental impacts to accrue in a given focuses on the adoption of three climate-friendly
region, a meaningful and localized proportion of behaviors: Fertilizing according to individual need,
producers needs to adopt climate-friendly practices. irrigating according to individual need, and enrich-
Similarly, because farmers are familiar with inten- ing soil with nutrient-rich, carbon-storing com-
sive (rather than sustainable) production, any effort post—a short but impactful collection of practices
to change behavior will be met by a strong negative that provide tangible benefits for farmers on an
norm: Where a nudge may emphasize that ‘less-is- individual level (less fertilization costs, increased
more’, their own experience says ‘more-is-more’. climatic resilience, etc.) as well as collectively (sta-
Past efforts in the region have found this normative bilization of the water supply, the return of pol-
barrier particularly hard to overcome, resulting in linators, a reduction in pests and diseases, etc.).
lower and shorter adoption rates across the farm- Nonetheless, where many would assume that the
ing population. Lastly, even on a relatively localized benefits of the practices speak for themselves, L4L
landscape, farmers are a diverse group of people, rests on one overarching insight: ‘When in doubt,
in that they belong to different social groups, have farmers stick to what they know’—and in the case
different levels of education, and demonstrate dif- of sustainable farming, there is a lot of doubt.
ferent levels of resistance to change. Consequently, For example, while incorporating compost into
to achieve the levels of adoption needed to influ- one’s fertilization plan significantly counteracts
ence the environment meaningfully, Rare needed soil depletion, many farmers remain unconvinced—
to design a behaviorally-informed program that i) and who can blame them? Agricultural production
adapted to farmers’ differing needs and attitudes, is financially precarious, which makes most farm-
ii) generated new, positive norms around sustain- ers reluctant to experiment with methods with
ability in farming communities, and iii) succeeded which they are unfamiliar. Additionally, because
in getting a localized-but-ever-increasing number they live season-to-season, we find that farmers
of farmers to adopt climate-friendly practices. In do not necessarily notice and associate the pattern
short, we needed to fully rethink the conventional of decline in their productivity to anything related
approach to agricultural behavior-change, doing to soil depletion (a direct effect of overfertilization,
so by leveraging the entire toolkit provided by the overirrigation, and a lack of compost). Confirmation
behavioral sciences. bias is ubiquitous, whereby farmers tend to over-
look information or experiences in favor of those
Tackling Barriers to Adoption Through Behavioral that support their viewpoint; very often, in fact,
Science they simply blame climate change for their woes.
There is little trust in information coming from
Rare’s L4L program saw light in 2019 as a response outside their communities—a legacy of Colombia’s
to the failings of past behavior-change efforts that 40-year internal conflict, not to mention a legacy
relied mainly on payment services and/or on educa- of what feels like many broken promises from past
tion that encouraged shifts in behavior. Specifically, NGO activities that culminated in no real change for
though many such programs had been successful at farmers. Combining these factors with our human
recruiting and training small cohorts of curious or tendency to avoid uncertainty, we can see why Co-
innovative farmers, they had often failed—in the lombian farmers prefer to stick with the devil they
Colombian context—to expand their reach because know, but where risk aversion is usually to blame,
of deeply-rooted community norms. To counter here the situation is steeped in ambiguity. This dis-
this issue, L4L builds on the success of other holis- tinction is important, because while humans tend
tic behavior-change programs (e.g., Gillingham & to both be risk- and ambiguity-averse, the type of

Behavioral Economics Guide 2021 125

Philipe Bujold & Erik Thulin Complex Behavioral Challenges Require Multi-Faceted Behavioral Solutions

uncertainty one faces changes how decisions can sistant (Mid-Resistance Farmers, MRFs), and some
be swayed. While people are generally risk-averse will resist anything that is remotely ambiguous to
when thinking about outcomes in terms of gains, them (High-Resistance Farmers, HRFs). But rather
for instance, framing choices in terms of potential than defining quantitative farmer groups in ad-
losses leads most of us to become risk-seeking. vance, it is the time at which they interact with the
This is where a simple nudge might be particularly program—essentially the mechanism by which we
effective. Unfortunately, when tackling ambiguity, induce behavioral shifts—that informs how best to
practitioners’ options are more limited: Farmers target them.
prefer certainty over risk and risk over ambiguity, Farmers that fall along the lower end of the re-
and the more ambiguity-averse a farmer is, the sistance-to-ambiguity axis, for example, are likely
more resistant they are to trying and adopting new to require but minimal exposure to a new, more
practices (e.g., Crentsil et al., 2020; Ross et al., 2012; cost-effective practice before wanting to adopt
Warnick et al., 2011). it. LRFs, then, are exemplified by those producers
Our analysis revealed that—as for many of us— that come forth and engage with open-recruitment
farmers’ production decisions relied on their ex- NGO programs. For them, L4L needs to ensure that
perience with various techniques to gauge their adoption is made simple and easy to maintain,
reliability. In other words, outcomes that farmers thereby alleviating any ambiguity about the prac-
have themselves experienced feel more tangible, tices with timely reminders and personalized advice,
less ambiguous. Another important factor is what and encouraging record-keeping, to make their ben-
they see others around them doing. This is why efits more tangible.
norms are so powerful, in that people tend to follow MRFs, on the other hand, are those farmers that
the behaviors they see others engaging in; they also require more convincing, in that while they recog-
tend to rely on the decisions of others when unsure nize that the new practices may be better, it is seeing
about what to do (Venema et al., 2020). If everyone others around them succeed that encourages them
is doing one thing, the reasonable assumption is to participate. Consequently, this group is particu-
that it works. Thus, when presented with new, sus- larly influenced by what they see around them (by
tainable production techniques, farmers need only the descriptive norm in their community) and by
look around to feel confident in their decision not to the social proof of success that their peers provide.
adopt them. On the flip side, the more people they Farmers are more likely to adopt a new behavior
see shifting their practices, the more likely they will if they know that other farmers also have (Cole &
be to follow suit. Farmers will also more likely adopt Fernando, 2016; Genius et al., 2014; Kuhfuss et al.,
new practices if there is a sense that others in their 2016; Vu et al., 2020) and recommend the behavior
community expect them to do so. (Fafchamps et al., 2020; Villamayor-Tomas et al.,
2019). Using these levers once early social proof has
Layering Evidence to Alleviate Biases been established should maximize MRF adoption.
Finally, for some farmers, more evidence just
No single intervention can provide enough confi- does not suffice. They are so averse to ambiguity (or
dence in the practices for all farmers to adopt. The change) that the benefits of adopting new practic-
people L4L works with are varied, and so is the degree es—to them—do not outweigh the costs. Supple-
to which they rely on social norms and personal ex- mental gains or losses are needed to sway behavior,
perience to make sense of their choices. Some farm- either by pairing the novel practices with additional
ers demonstrate little aversion towards ambiguous benefits or by imposing a cost to maintaining pre-
outcomes; others are more apprehensive. L4L was vious ones. However, these incentives need not be
designed with this in mind, to meet farmers wher- material. With sufficient support from the rest of
ever they fall on the ‘resistance-to-ambiguity’ axis: the community, injunctive norms can drive social
Some farmers will demonstrate little (Low-Resist- sanctions sufficient to support compliance. In this
ance Farmers, LRFs), some will be moderately re- way, the normative expectations of the rest of the

Behavioral Economics Guide 2021 126

Philipe Bujold & Erik Thulin Complex Behavioral Challenges Require Multi-Faceted Behavioral Solutions

community, thinking that they should be adopt- of farmers who have adopted sustainable farming
ing the novel practices because this is what is best practices in the community. Here, the target is
for the community, can shift the balance and sway MRFs, whereby anyone convinced by the ongo-
these most resistant farmers (Bicchieri, 2016). ing campaign can join the program to i) receive
training on how to build an agroclimatic station,
Bringing Theory With Practice: The Blueprint in order to signal their commitment, ii) attend
peer-led workshops and centralized technical
The above theoretical framework is what guides assistance clinics, and iii) gain access to our two-
and allows L4L to go beyond a one-size-fits-all way SMS platform. Throughout this process, L4L
strategy, and to ultimately leverage farmers’ dif- monitors the perception farmers have about the
ferences in a way that reinforces—rather than prevalence of sustainable practices in their com-
impedes—the effectiveness of the program. The munity (i.e., the descriptive norm).
result is a three-phased effort where each phase • Phase 2: Generates a community-wide under-
delivers its own set of nudges, social marketing, standing that everyone (but particularly farmers)
and community-building activities to slowly build benefits from all farmers farming sustainably.
community confidence. First, invest heavily in an Once ongoing monitoring reveals a significant
early cohort of innovadores (farmer innovators), shift in the community’s descriptive norm, the
second, leverage the social proof they provide the program moves into its final phase. Social mar-
community to set a change of norm in motion, and keting now revolves around highlighting the col-
finally third, leverage the ever-growing number of lective benefits of widespread adoption and how a
innovadores and their families to reinforce that the growing proportion of the community expect you
community now expects farmers to farm sustain- to be farming sustainably, too. Community events
ably (see Figure 1). The farmer segments described (traditional plays, songs, school activities) high-
in the previous section are thus more mechanistic light the positive externalities of farming sus-
than descriptive in nature, with L4L participants tainably and continue to reinforce the good work
self-segmenting based on their confidence in the of LRFs and MRFs. Those HRFs who do not make
practices—or lack thereof. the shift will eventually have to contend with rep-
utational losses. This is also where L4L exits the
• Phase 0: Generates tangible evidence of adop- community, once proof has turned to pressure,
tion ease and relevant, local success. L4L offers and the new status quo reaches a point of self-en-
individualized training to LRFs that approach forcing equilibrium.
the program and provides them with timely,
practice-specific reminders over a two-way SMS Conclusion
platform, to ease and reinforce the adoption of
our three practices. Each innovador also receives The chorus is growing in calls for the behavioral
an agroclimatic station—a marker placed on sciences to answer to more than the “low-hang-
their plot that publicly signals their participation ing-fruits” of the initial nudge revolution (Nest-
in the program. At the end of their training, the erak, 2020; Sanders et al., 2018; Spencer, 2018;
graduating innovadores are celebrated in a public Van Der Linden, 2018). Addressing these complex
recognition event in which they receive accolades and wicked problems requires tackling the many
from a prestigious member of their community (a shortcomings of traditional nudges, including their
mayor, a priest, a local celebrity, etc.). relatively small effects, their lack of persistence, as
• Phase 1: Publicly showcases the increasing num- well as their often one-size-fits-all approach. L4L
ber of farmers that are adopting and benefiting meets this challenge by joining a handful of holistic
from sustainable practices. Following the public behavior change programs that move beyond the
recognition event, L4L launches a social market- field’s narrow focus on singular nudges, to achieve
ing campaign designed to highlight the successes behavior change using a holistic evidence-based

Behavioral Economics Guide 2021 127

Philipe Bujold & Erik Thulin Complex Behavioral Challenges Require Multi-Faceted Behavioral Solutions

Figure 1: The L4L program blueprint. Taking a multi-layered, three-phased approach to encouraging sustainable
farming practices, L4L reaches farmers when they are most likely to make the shift. Its three phases each target
different farmer subgroups, but each phase builds on the one preceding it—taking advantage of successive
behavior change outcomes to achieve an impact greater than the sum of its parts. Put simply, phase 0 makes
climate-smart agriculture easy, phase 1 makes it observable, and phase 2 makes it expected.

approach (e.g., Gillingham & Bollinger, 2017; Pick- of behavior-change levers that the behavioral
ering et al., 2019). sciences offer. As this case study shows, the toolkit
To have a meaningful environmental impact—in that the behavioral sciences have made available is
most cases—it is simply not enough to shift behav- much wider than nudges and choice architecture. All
ior a mere couple of percentage points spread out we need now is a bit more creativity in how we apply
across the population. Rather, program designers the more intensive, multi-staged, and fine-tuned
and policymakers have aimed to engender wide- tools to ensure larger, long-lasting, and far-reach-
spread collective action, a call that the behavioral ing outcomes for our behavior-change programs.
sciences need to answer. To drive collective action,
we need a collection of behavior-change interven- The Authors
tions. We need strategies that are up to the task of
shifting social systems as a whole and that drive Philipe Bujold is a behavioral scientist at Rare’s
community-wide change. Center for Behavior & the Environment, where he
Rather than the silver bullet that nudges are often translates the latest insights from the behavioral
sold as being, practitioners would do well to re- sciences into application-ready behavior-change
member that nudges are only part of a wider toolkit interventions. He holds a Ph.D. in Behavioral Neuro-

Behavioral Economics Guide 2021 128

Philipe Bujold & Erik Thulin Complex Behavioral Challenges Require Multi-Faceted Behavioral Solutions

science. His past research focused on the neurobio- technology adoption: Evidence from aquacul-
logical underpinnings of primate decision-making ture in Ghana. Journal of Economic Behavior &
and the role evolution played in biasing our choices. Organization, 179, 46-68.
Damgaard, M. T., Gravert, C., & Villalobos, C. (2018).
Erik Thulin is the behavioral science lead at Rare’s Forgetful or reluctant? Evidence on reminder re-
Center for Behavior & the Environment. He leads sponse and donor behavior from panel data. In K.
the center’s research agenda and collaborations. He Scharf & M. Tonin (Eds.), The economics of phi-
holds a Ph.D in Psychology, previously focusing on lanthropy: Donations and fundraising. MIT Press.
the shifting and measurement of social norms in De Ridder, D., Feitsma, J., Van Den Hoven, M.,
public health interventions. Kroese, F., Schillemans, T., Verweij, M., Venema,
T., Vugts, A., & De Vet, E. (2020). Simple nudges
Acknowledgments that are not so easy. Behavioural Public Policy, 36.
https://doi.org/10.1017/bpp.2020.36/.
Thank you to The MIT Applied Cooperation Team, Ewert, B. (2020). Moving beyond the obsession with
Erez Yoeli, and the Rare Lands for Life team for their nudging individual behaviour: Towards a broad-
support in developing this concept. This work was er understanding of Behavioural Public Policy.
supported in part by a gift from Itaú Unibanco to the Public Policy and Administration, 35(3), 337-360.
MIT Sloan Sustainability Initiative. Fafchamps, M., Islam, A., Malek, M. A., & Pakrashi, D.
(2020). Can referral improve targeting? Evidence
References from an agricultural training experiment. Journal
of Development Economics, 144, 102436. https://
Allcott, H., & Rogers, T. (2014). The short-run and doi.org/10.1016/j.jdeveco.2019.102436.
long-run effects of behavioral interventions: Ferraro, P. J., & Price, M. K. (2013). Using nonpecu-
Experimental evidence from energy conserva- niary strategies to influence behavior: Evidence
tion. American Economic Review, 104(10), 3003- from a large-scale field experiment. The Review
3037. of Economics and Statistics, 95(1), 64-73.
Bacon, L., & Krpan, D. (2018). (Not) Eating for the Genius, M., Koundouri, P., Nauges, C., & Tzouvelekas,
environment: The impact of restaurant menu V. (2014). Information transmission in irriga-
design on vegetarian food choice. Appetite, 125, tion technology adoption and diffusion: Social
190-200. learning, extension services, and spatial effects.
Beshears, J., Dai, H., Milkman, K. L., & Benartzi, S. American Journal of Agricultural Economics, 96(1),
(2015, November). Framing the future: The risks 328-344.
of pre-commitment nudges and potential of fresh Gillingham, K., & Bollinger, B. (2017). Solarize your
start messaging. Paper presented at the 36th community: An evidence-based guide for accel-
Annual Conference of the Society for Judgment erating the adoption of residential solar. https://
and Decision Making, Chicago, IL. cbey.yale.edu/research/solarize-your-commu-
Byrne, D. P., Nauze, A. L., & Martin, L. A. (2018). nity-an-evidence-based-guide-for-accelerat-
Tell me something I don’t already know: ing-the-adoption-of.
Informedness and the impact of information Gravert, C., & Kurz, V. (2019). Nudging à la carte:
programs. The Review of Economics and Statistics, A field experiment on climate-friendly food
100(3), 510-527. choice. Behavioural Public Policy, 11. https://doi.
Cole, S., & Fernando, A. N. (2016). ‘Mobile’izing org/10.1017/bpp.2019.11.
agricultural advice: Technology adoption, dif- Green, K. M., Crawford, B. A., Williamson, K. A., &
fusion and sustainability. SSRN. https://doi. DeWan, A. A. (2019). A meta-analysis of social
org/10.2139/ssrn.2179008. marketing campaigns to improve global con-
Crentsil, C., Gschwandtner, A., & Wahhaj, Z. (2020). servation outcomes. Social Marketing Quarterly,
The effects of risk and ambiguity aversion on 25(1), 69-87.

Behavioral Economics Guide 2021 129

Philipe Bujold & Erik Thulin Complex Behavioral Challenges Require Multi-Faceted Behavioral Solutions

Hallsworth, M., Chadborn, T., Sallis, A., Sanders, M., Sciences, 24(12), 969-980.
Berry, D., Greaves, F., Clements, L., & Davies, S. C. Pickering, J., McIntosh, T., Moore, S., Priwitzer,
(2016). Provision of social norm feedback to high S., Haanterä, K., Preston, G., & Hong, J. (2019,
prescribers of antibiotics in general practice: A May). Project Cane Changer: Using behavioural
pragmatic national randomised controlled trial. science to create practice change. Paper presented
The Lancet, 387(10029), 1743-1752. at the Proceedings of the 2019 Conference of the
Hansen, P. G. (2018). What are we forgetting? Australian Society of Sugar Cane Technologists,
Behavioural Public Policy, 2(2), 190-197. Toowoomba, Queensland, Australia.
Hummel, D., & Maedche, A. (2019). How effective Richter, I., Thøgersen, J., & Klöckner, C. (2018).
is nudging? A quantitative review on the effect A social norms intervention going wrong:
sizes and limits of empirical nudging studies. Boomerang effects from descriptive norms in-
Journal of Behavioral and Experimental Economics, formation. Sustainability, 10, 2848. https://doi.
80, 47-58. org/10.3390/su10082848.
Kuhfuss, L., Préget, R., Thoyer, S., & Hanley, N. Ross, N., Santos, P., & Capon, T. (2012, August). Risk,
(2016). Nudging farmers to enrol land into ambiguity and the adoption of new technologies:
agri-environmental schemes: The role of a col- Experimental evidence from a developing economy.
lective bonus. European Review of Agricultural Paper presented at the International Association
Economics, 43(4), 609-636. of Agricultural Economists (IAAE) Conference,
McDonald, G., Wilson, M., Veríssimo, D., Twohey, Foz do Iguaçu, Brazil.
R., Clemence, M., Apistar, D., Box, S., Butler, P., Sanders, M., Snijders, V., & Hallsworth, M. (2018).
Cadiz, F. C., Campbell, S. J., Cox, C., Effron, M., Behavioural science and policy: Where are we
Gaines, S., Jakub, R., Mancao, R. H., Rojas, P. T., now and where are we going? Behavioural Public
Tirona, R. S., & Vianna, G. (2020). Catalyzing Policy, 2(2), 144-167.
sustainable fisheries management through Spencer, N. (2018). Complexity as an opportuni-
behavior change interventions. Conservation ty and challenge for behavioural public policy.
Biology, 34(5), 1176-1189. Behavioural Public Policy, 2(2), 227-234.
Metcalf, A. L., Angle, J. W., Phelan, C. N., Muth, B. Sunstein, C. R. (2017). Nudges that fail. Behavioural
A., & Finley, J. C. (2019a). More “bank” for the Public Policy, 1(1), 4-25.
buck: Microtargeting and normative appeals Thaler, R. H., & Sunstein, C. R. (2009). Nudge:
to increase social marketing efficiency. Social Improving decisions about health, wealth, and
Marketing Quarterly, 25(1) 26-39 happiness. Penguin.
Metcalf, A. L., Phelan, C. N., Pallai, C., Norton, Van Der Linden, S. (2018). The future of behavioral
M., Yuhas, B., Finley, J. C., & Muth, A. (2019b). insights: On the importance of socially situated
Microtargeting for conservation. Conservation nudges. Behavioural Public Policy, 2(2), 207-217.
Biology, 33(5), 1141-1150. Venema, T. A. G., Kroese, F. M., Benjamins, J. S., &
Mols, F., Haslam, S. A., Jetten, J., & Steffens, N. K. de Ridder, D. T. D. (2020). When in doubt, follow
(2015). Why a nudge is not enough: A social iden- the crowd? Responsiveness to social proof nudg-
tity critique of governance by stealth. European es in the absence of clear preferences. Frontiers
Journal of Political Research, 54(1), 81-98. in Psychology, 11, 1385. https://doi.org/10.3389/
Nesterak, E. (2020, January 20). Imagining the fpsyg.2020.01385.
next decade of behavioral science. Behavioral Venema, T. A. G., Kroese, F. M., & Ridder, D. T. D. D.
Scientist. https://behavioralscientist.org/imag- (2018). I’m still standing: A longitudinal study
ining-the-next-decade-future-of-behavio- on the effect of a default nudge. Psychology &
ral-science/. Health, 33(5), 669-681.
Osman, M., McLachlan, S., Fenton, N., Neil, M., Villamayor-Tomas, S., Sagebiel, J., & Olschewski,
Löfstedt, R., & Meder, B. (2020). Learning from R. (2019). Bringing the neighbors in: A choice
behavioural changes that fail. Trends in Cognitive experiment on the influence of coordination and

Behavioral Economics Guide 2021 130

Philipe Bujold & Erik Thulin Complex Behavioral Challenges Require Multi-Faceted Behavioral Solutions

social norms on farmers’ willingness to accept

agro-environmental schemes across Europe.
Land Use Policy, 84, 200-215.
Vu, H. T., Tran, D., Goto, D., & Kawata, K. (2020).
Does experience sharing affect farmers’ pro-en-
vironmental behavior? A randomized controlled
trial in Vietnam. World Development, 136, 105062.
https://doi.org/10.1016/j.worlddev.2020.105062.
Warnick, J. C. E., Escobal, J., & Laszlo, S. C. (2011).
Ambiguity aversion and portfolio choice in
small-scale Peruvian farming. The B.E. Journal
of Economic Analysis & Policy, 11(1). https://doi.
org/10.2202/1935-1682.2331.

Behavioral Economics Guide 2021 131

 Gamification in Behavioral Science: An Engaging
Prospect for Online Research
Joshua Henk Balsters* and Jo Evershed
Gorilla Experiment Builder

The Covid-19 pandemic has forced most behavioral science researchers to transition from in-person and
lab-based testing to online research. The sudden explosion in the quantity of online research studies has
inevitably affected both participant recruitment and engagement. In this increasingly crowded market
space, how can researchers maximize data quality and recruitment? In the last decade, gamification has
been increasingly employed as a tool to promote recruitment, engagement, and learning in a range of fields
(i.e., marketing, education, health and productivity tools, and fundamental research). Here, we review the
concept of gamification in behavioral science research, outline some of the different ways it is used, and
discuss examples where gamification has been employed to great effect in online research.

What Is Gamification? game elements (e.g., points, badges, leader boards,

etc.). The authors also suggest a distinction between
The gamification trend is believed to have started extrinsic and intrinsic game elements. An extrinsic
around 2010; however, the concept of embedding game element can be perceived clearly and objec-
tasks within a game to increase learning and moti- tively by the user, whereas an intrinsic element is
vation has been around for hundreds of years (Zich- presented more subtly so that the user is unaware of
ermann & Cunningham, 2011). A game is a system perceiving it (Toda et al., 2019).
in which players engage in a fictional challenge, In their taxonomy, Toda et al. (2019) described 21
defined by rules, interactivity, and feedback, which game elements and their synonyms. These elements
results in a quantifiable outcome often provoking were validated by two surveys with experts in the
an emotional reaction (Kankanhalli et al., 2012). field of gamification in education. This resulted in a
Given this definition, we can define gamification as taxonomy of five gamification dimensions: Perfor-
the use of these game elements and techniques in mance/measurement, Social, Ecological, Personal,
non-game contexts (Deterding et al., 2011; Kankan- and Fictional (see Figure 1). Whilst this taxonomy
halli et al., 2012). is highly useful for anyone wishing to develop a
gamified task, it is worth remembering that the
A Taxonomy of Gamification taxonomy proposed by Toda et al. (2019) was based
solely on expert opinion, not on actual users.
There are many potential ways to gamify a be-
havioral science task, each of which elicits differing Performance/measurement
levels of motivation and requires different levels
of input from the experimenter. In order to help Performance/measurement elements are the
researchers decide which game elements are more most commonly used in gamification. They include
appropriate in a particular context, Toda et al. (2019) rewarding performance by using points, levels,
eloquently summarize the different approaches to and achievements/badges. There are hundreds of
gamification, using a clear taxonomy of different examples where performance measurements have
been used to turn different tasks into games, such
* Corresponding author: [email protected] as Datacamp (gamifying computer programming

Behavioral Economics Guide 2021 132

Joshua Balsters & Jo Evershed Gamification in Behavioral Science: An Engaging Prospect for Online Research

education), Habitica (gamifying habit forming), will describe the HIVE,1 a fun and powerful game for
and Peloton (gamifying fitness), to name but a few. exploring conformity and diversity across multiple
All of these examples use extrinsic features to mo- individuals.
tivate participants and provide feedback.
Ecological
Social
The ecological dimension relates to the environ-
Another common feature in gamification is to ment implemented in gamification. This dimension
compare learning and performance with other us- includes elements such as chance (manipulating the
ers, i.e., the social dimension. The most common probability of winning or the size of the prize), time
way to do this is through leader boards, either mak- pressure, or rarity of prizes (e.g., availability cer-
ing users work hard to catch up with friends or work tain Pokémon). Decades of research in the fields of
hard to stay on top and maintain their reputation. neuroeconomics and value-based decision-making
Depending on the tasks, there are plenty of oppor- have provided robust neurocomputational models
tunities for social interaction either via competition that describe these behaviors (Rangel et al., 2008).
or cooperation. Towards the end of this article, we Crucially, the emerging field of computational

Figure 1: The taxonomy of gamification proposed by Toda et al. (2019). Image from Toda et al. (2019).

psychiatry is linking these mechanisms to psycho- the case studies section, we will discuss FunMaths,2
logical and psychiatric phenotypes (Montague et which has used gamification to help children that
al., 2012). Therefore, gamification could become a struggle with arithmetic and understanding num-
useful tool for clinical diagnostics and treatment. In ber relations.

1 https://gorilla.sc/the-hive/. 2
https://gorilla.sc/funmaths/.

Behavioral Economics Guide 2021 133

Joshua Balsters & Jo Evershed Gamification in Behavioral Science: An Engaging Prospect for Online Research

Personal suggested that engaging and rewarding participants

through gamification can help solve these problems
The personal dimension is related to the user of and therefore increase data quality by increasing
the environment; for example, is the game repeti- attention and motivation.
tive, or does it stay fresh and novel? Is it simple, or Bailey et al. (2015) investigated the impact of
do the tasks challenge the user? Are there pleasing gamification on survey responses. Here, the au-
sensations for the user, such as vibrant colors or thors refer to ‘soft gamification’, where traditional
sounds, to improve their experience? These ele- survey responses were replaced with more inter-
ments are implicitly rewarding to the user. esting tools like dragging and selecting images. We
consider this an increase in the personal dimension
Fictional in Toda et al.’s (2019) taxonomy. Bailey et al. (2015)
found that gamification led to richer responses (a
Lastly, the fictional dimension aims to link the significant increase in the numbers of words used
user experience with context through narrative in responding) and participants were engaged for
and storytelling elements. In these games, the user longer. This is just one of many examples that show
is unaware they are learning a skill or performing the benefits of gamifying research.
a task. A number of educational board games have Looysten et al. (2017) conducted a systematic
been developed over the years, using the fictional review of online studies, employing gamification
dimension (for example playing CBT), and more to investigate the effects of game-based environ-
recently, apps are being developed to extract be- ments on online research. To do so, they looked at
haviors through naturalistic game play (Sea Quest different measures of engagement, e.g., the amount
Hero) and even treat behavioral conditions like of time spent on the program and numbers of visits.
ADHD (Neuroracer). We will later discuss Treasure Taken together, their results suggest that gamifi-
Collector, a game for children that takes a basic
3
cation increases engagement in online programs
psychological paradigm and turns it into an exciting and enhances other outcomes, such as learning and
adventure! health behavior. However, the authors also suggest
that the impact of gamification features reduces
Does Gamification Improve Behavioral over time as the novelty of points, levelling up, and
Science? badges wears off. Looyestyn et al. (2017) cite the
example of the gamified app Four-square, which
In normal settings, experimental tasks and ques- was hugely successful upon release but failed to
tionnaires may be tedious for participants. They retain users after 6-12 months. I am sure we can
often use simplistic stimuli, are repetitive, and all think of other examples of websites and apps
provide no feedback on personal performance or that were hugely popular at first but failed to retain
the performance of others (scoring poorly on four customers. This suggests that only utilizing perfor-
of the five gamification dimensions mentioned mance/measurement elements in gamification will
above). In any experiment or survey, participants’ lead to initial spikes but fail to retain customers/
attention gets lower over time, thereby increasing participants. This may not be a problem for most
the error rate. This is amplified in remote and on- behavioral science experiments where long-term
line testing, where there are likely to be any number retention is not required; however, we suggest it is
of distractions that are not present in the lab. When worth considering when developing longitudinal/
the experimenter is not present, participants are multi-session games.
more likely to drop out if they become bored—in- The systematic review by Looyestyn et al. (2017)
creasingly so when experiments require more than provides the strongest evidence to date that gami-
one session (Palan & Schitter, 2018). It has been fication significantly increases participant engage-
ment. However, it is also worth highlighting some
2 https://gorilla.sc/treasure-collector/. limitations. First, the positive effect of gamification

Behavioral Economics Guide 2021 134

Joshua Balsters & Jo Evershed Gamification in Behavioral Science: An Engaging Prospect for Online Research

was not found for all measures of engagement, Gamification Case Studies
which sheds some doubt on the generalizability of
these results. Second, although they began with Games That Improve Learning: Funmaths Gamifies
1,017 online studies, only 15 studies remained for Arithmetic Skills for Children
analysis after the exclusion criteria were applied.
This small sample size, and large heterogeneity Dyscalculia is a developmental condition that
in terms of population, methods, and outcomes, affects the ability to acquire arithmetical skills,
meant these studies were not directly compara- i.e., dyslexia for numbers. Individuals with dy-
ble, and thus it was not possible to conduct a me- scalculia lack an intuitive grasp of numbers and
ta-analysis. To overcome this limitation, future their relations. Reports suggest that around 5-7%
research should provide more standardized testing, of children may have developmental dyscalculia
measures, and analysis methods in online research. (similar prevalence to developmental dyslexia), and
There is some promising work in this direction. In a it is estimated that low numeracy skills cost the UK
recent study, Chierchia et al. (2019) provided a bat- £2.4billion annually (Butterworth et al., 2011). But-
tery of novel ability tests to investigate non-verbal terworth et al. (2011) further propose that bringing
abstract reasoning. The battery was validated on the lowest 19.4% of Americans to the minimum level
adolescents and adults who performed matrix rea- of numeracy would lead to a 0.74% increase in GDP
soning by identifying relationships between shapes. growth. There are clear economic and social bene-
While non-verbal ability tests are usually protected fits to improving arithmetical skills in the general
by copyright, Chierchia et al. (2019) made their bat- population. Professor Diana Laurillard (Professor of
tery open access for academic research.
4
Learning with Digital Technologies at UCL Institute
of Education) developed a series of math games to
train math skills in dyscalculic children through
simple manipulations of objects.

Figure 2: NumberBeads: Use a knife to split beads to match the target.

In one of the games (NumberBeads), children learn

about addition, subtraction, and the number line by
3 https://app.gorilla.sc/openmaterials/36164. combining and segmenting strings of beads. Figure

Behavioral Economics Guide 2021 135

Joshua Balsters & Jo Evershed Gamification in Behavioral Science: An Engaging Prospect for Online Research

2 shows an example where the target is a chain of ally constructing the game themselves.” She also
two beads. In this instance, a knife is being used to suggests that similar games could be developed to
slice the larger chain of beads into chains of two train language skills in people with dyslexia.5
beads. Participants can combine and cut these up as Games such as these provide a high-quality ed-
they wish, and when this is done correctly the chain ucational resource at a very low cost. According
disappears in a pleasing puff of success! The game to Prof. Laurillard, these games have tremendous
continually adapts to the player’s ability, building value, because they provide individualised and en-
up their knowledge of the number line, fluency of joyable mathematics tuition to students both with
the number line, and understanding of numerals and without dyscalculia. One player said, “I’d play
(utilizing performance and personal gamification it all day,” while a teacher said “I was absolutely
dimensions). In an interview, Prof. Laurillard calls
it a “constructionist” game, as children “are actu-

Figure 3: Treasure Collector: Executive function training tasks for children.

astounded by the work they were doing with this. Students clearly enjoy playing these games, which
They were clearly seeing things in a different way.” continually stretch and extend them (personal di-
mension of gamification taxonomy).
4
You can read the full interview here: https://gorilla.sc/
gamified-research-series-constructionist-gaming/.

Behavioral Economics Guide 2021 136

Joshua Balsters & Jo Evershed Gamification in Behavioral Science: An Engaging Prospect for Online Research

Games to Increase Motivation: Treasure Collector quality adaptive training for the research project.
Gamifies Executive Function Training in Children According to Prof Bishop, while you can typical-
ly get an adult to do around 100 trials of a boring
We previously mentioned the challenges of at- adaptive task, with kids, after three or four trials,
trition for online longitudinal research (Palan & they’ll say, “Is there much more of this?”.7 This is
Schitter, 2018). Professor Nikolaus Steinbeis, based bad news if you want them to train every day for 8
at UCL, wanted children (7-10yrs) to train 10 min- weeks! And yet, with the Treasure Collector Games,
utes a day for 8 weeks, in order to improve executive Prof Steinbeis had students train for 10 minutes on
function. Such a task would have been impossible the game, four times a week for 8 weeks. Overall,
without gamification. participants completed around 4,000 trials in to-
Children would be training on the Go/No-Go task, tal—and reported that they still enjoyed the game.
which tests attention and response inhibition by It is clear that without gamification, this study
asking participants to respond to certain stimuli would have been impossible, and so by employing
as fast as possible (Go trials) versus withholding a gamification, a range of developmental research
response to other stimuli (No-Go trials).6 To keep questions become possible.
participants engaged, the classic Go/No-Go task
was embedded into a larger narrative of being an Games That Answer New Research Questions – the
explorer (an example of the fictional dimension Hive and Multiplayer Games
of gamification). Figure 3 shows how participants
chose their own avatar, which was integrated into The Hive8 is a research platform for studying how
the story. The Go/No-Go task was then reskinned people think, feel, and behave together in groups
in a variety of situations, including when to dig for (Bazazi et al., 2019; Neville et al., 2020). It works as
treasure, or when to steal gold from a dragon (Fig- an app that people can access with their smartphone.
ure 3), or when to drive straight or swerve to avoid After logging in, the Hive environment displays a
ice on the road. The narrative elements and varied dot that can be dragged around. The coordinate of
game play increased compliance and helped deliver each dot is recorded, thus allowing experimenters

Figure 4: Schematic of an experiment set up and minimal group assignment. Image from Neville et al. (2020).

to analyze trajectories and rest periods in a similar they perform different tasks while monitoring oth-
way to experiments utilizing eye or mouse track- er individuals’ decisions, represented by the move-
ing. Each participant sees their own dot, and other ments of the other dots (see Figure 4).
participants’, moving on the central display. Then,
6
http://cauldron.sc/clients#hive, developed with Profes-
5
See an example here: https://research.sc/partici- sor Daniel Richardson at UCL.
pant/login/dynamic/907B0C84-A7AD-4E1C-8F32- 7
see https://gorilla.sc/gamified-research-series-inves-
093AEFB03039. tigating-language-development/.

Behavioral Economics Guide 2021 137

Joshua Balsters & Jo Evershed Gamification in Behavioral Science: An Engaging Prospect for Online Research

One of the studies involving the Hive investigates A fascinating series of studies by Manohar et al.
the link between mimicry and self-categorization, (2015) investigated the effect of extrinsic reward on
and it attempts to answer the following question: the speed-accuracy trade-off, which is supposed
Do we always do what others do and, if not, what are to be a fundamental law—as we move faster, we
the factors that influence our decisions in a group? become less accurate. However, monetary incen-
(Neville et al., 2020). The experiment has been con- tives break this law, and participants became both
ducted at multiple public events such as the Science faster AND more accurate. This is just one of many
Museum in London with groups between four and examples showing that extrinsic reinforcers change
12 people. Participants are assigned to two groups behavior. That said, external reinforcers are often
i.e., the red and the blue dot. Then, they play a series present in traditional lab-based behavioral eco-
of games involving moving their dots and looking at nomics studies. Total score bars are commonplace
the choices of the other participants (belonging to in value-based decision-making studies, yet re-
both their own group and to the other one). Overall, searchers in the field still argue about the extent to
the results show that participants are influenced by which this biases behavior in line with assumptions
the movements of the confederate dots that are the of prospect theory (Kahneman & Tversky, 1979).
same colors as their own. The authors conclude that Ryan and Deci (2000) distinguish between two
mimicry is affected by in-group/out-group knowl- different forms of motivation: Intrinsic and extrin-
edge, i.e., knowledge of whether people belong to sic. Intrinsic motivation relates to the individual’s
the same category as us. The Hive project allows one satisfaction in performing an activity in and of
to study a fundamental question, namely: Do people itself, while extrinsic motivation occurs when the
take decisions differently when they think as indi- activity is performed to obtain another and tangible
viduals or as a crowd? Instead of paying participants outcome, e.g., money as a reward. This dichotomy
for performing a long and boring experiment in a is mirrored by the implicit/explicit dimensions of
lab, the Hive allows researchers to investigate this gamification noted in Toda et al.’s (2019) taxon-
issue everywhere, using people’s smartphones, and omy. It is therefore likely that employing implicit
without any additional costs whilst maintaining the vs explicit dimensions of gamification will affect
precision of lab-based testing. intrinsic or extrinsic motivation in unique ways.
Mekler et al. (2017) investigated the effects of
Will Gamification Influence My Findings? individual game elements on intrinsic and extrin-
sic motivation in an image annotation task. Mekler
Our literature review suggests that gamification is et al. (2017) found that gamification significantly
effective at increasing participant engagement and improved extrinsic factors like performance, espe-
retention, and even increasing data quality in both cially when using leader boards and points, but not
qualitative and quantitative experiments. However, intrinsic motivation or competence (the perceived
Bailey et al. (2015) reported a concern that apply- extent of one’s own actions as the cause of desired
ing gaming mechanics to questions can change the consequences in one’s environment). This point
character of the answers and lead to qualitatively was also raised by Looyestyn et al. (2017), who not-
different responses. Therefore, to what extent does ed that the positive effects of gamification seemed
gamification change behavior? to lessen over time: The performance/measure-
A common concern for researchers is wheth- ment dimension of gamification is only effective in
er gamification will fundamentally change the the short run. Looyestyn et al. (2017) suggest that
outcomes of task behavior or surveys being ad- in order to be successful in the long term, gamified
ministered. Will gamifying my research mean the applications should focus on intrinsic, instead of
findings are no longer valid? We can boil these extrinsic, motivation, i.e., focus on the personal
questions down to ‘do external rewards and moti- and fictional dimensions of gamification. For future
vators change behavior?’ The answer to this final applications, it is crucial to design game environ-
question is certainly ‘yes’. ments that enhance users’ intrinsic motivation to

Behavioral Economics Guide 2021 138

Joshua Balsters & Jo Evershed Gamification in Behavioral Science: An Engaging Prospect for Online Research

keep them engaged over time, potentially moving have a running total score, but that’s probably not
more towards games instead of gamification (see what motives any of us to read. This approach to
below). gamification ignores the fact that the book itself is
Lastly, we wish to suggest the possibility that intrinsically motivating, or to put it another way, a
differences between traditional tasks and games good book doesn’t need gamifying. The objective of
may not be such a bad thing for gamified research. a game is pleasure or to learn a new skill, and there-
In these situations, we mostly consider lab-based fore the motivation to play it is often intrinsic (i.e.,
testing to be the ‘ground-truth’ in psychology and the personal and fictional dimensions). This intrin-
behavioral science. However, lab conditions and sic/extrinsic distinction in motivation changes the
tasks can actually be quite artificial. Psychological way behaviors are learned and reinforced.
tasks are often reduced to their most basic elements Most research tasks are designed to test a very
so that scientists can make accurate inferences specific question, and as such they will only have a
about the factors that influence behavior. Howev- limited number of response options that can easily
er, it is often the case that lab-based findings are be categorized as correct or incorrect. However,
not effective at predicting behaviors outside the lab games typically have a larger range of responses,
(Kingstone et al., 2008; Shamay-Tsoory & Mendel- which can lead to improvisation. They also offer
sohn, 2019). Therefore, even if you do find differ- the player the opportunity to explore a world, and
ent results between paradigms run in the lab and learning is often implicit and directed by the player
gamified versions of tasks, that does not mean that rather than by the experimenter. Compared to gam-
your game-based findings are inherently wrong ification, games often employ a more construction-
or less valid. We do not yet know which of these is ist approach that leads to discovery learning. The
closer to the ‘ground truth’. It may be that games, FunMaths game is one such example, as participants
which are often more natural and more intrinsically can achieve their goals in several different ways and
motivating, are in fact more relevant to real-world are given the opportunity to explore different op-
decision-making. tions. This is different to Treasure Collector, which
uses game elements such as a story narrative to
Gamification vs Games in Research increase motivation and engagement for a single,
simple task (Go/No-Go task). Thus, one can argue
There is a subtle, but important, distinction to that Treasure Collector is an example of gamifica-
be made between using gamification and games tion of the Go/No-Go task, whereas FunMaths is
in research. Whilst gamification refers to adding an educational game designed to improve learning.
game elements to existing tasks, it is also possi- However, when gamification is done well, it should
ble to create research games instead of gamifying be near impossible to distinguish it from a game.
existing research paradigms. Research games will When using games to investigate naturalistic
be intrinsically motivating (and thus, hopefully, behaviors, researchers must contend with a wider
maintain engagement over time) and allow for the array of behaviors—statistically, we could refer to
exploration of more naturalistic behaviors. this as a larger parameter space. Each decision is not
Typically, the objective of gamification is to in- made in isolation, and choices are likely to interact
crease motivation and engagement. This is often with one another, creating large, multi-factorial
achieved by using extrinsic motivators such as designs. Rich datasets like this are perfect for ma-
points, badges, and leader boards (i.e., the perfor- chine learning algorithms, which can help identify
mance/measurement dimension), but what is the which combinations of behaviors best predict out-
point of points? We can imagine gamifying reading comes. However, generating meaningful inferences
by stating that each page is a point, thus motivat- from potentially enormous matrices of behavior
ing someone to read more each day, as it is worth combinations requires an even larger number of
more points. Helpfully, books already have points datapoints, i.e., lots of participants. With tradition-
printed on each page (the page numbers) so you al lab-based or online testing this would increase

Behavioral Economics Guide 2021 139

Joshua Balsters & Jo Evershed Gamification in Behavioral Science: An Engaging Prospect for Online Research

participant costs hugely. However, we have already al Holloway) over the last 15+ years, investigating a
highlighted that games can be intrinsically moti- range of topics, including neuroimaging methods,
vating and genuinely enjoyable, thus significantly neuroeconomics and social decision-making, rule-
reducing participants’ fees (maybe even removing based learning and automation, cognitive decline
them all together). For each experiment, there will in ageing, and motivation in autism spectrum con-
be a breakeven point where, if you want more than a ditions. He holds a BSc in Psychology (2005) and a
certain number of participants, it becomes cheaper PhD in Cognitive Neuroscience (2009) from Royal
to invest in developing an exciting game compared Holloway University of London. He currently has
to using traditional behavioral science paradigms 47 peer-reviewed publications, has won over £1.5
and paying participants for their time. In the case million in grant funding, and is regularly an invited
of Sea Hero Quest (www.seaheroquest.com), they speaker at international conferences and top uni-
are reported to have recorded data from 4.3 mil- versities around the world.
lion players, who have played for a total of over 117
years. Collecting 117 years’ worth of data via a re- Jo Evershed is the host of BeOnline, a conference
cruitment service such as Prolific (117 participants, all about online behavioral research, and the crea-
525,600 minutes each at £7.50 per hour) would tor of Gorilla Experiment Builder. She holds a BSc in
cost over £10.7 million. As tools for making games Psychology from UCL and a BSc in Combined Stud-
become cheaper and more accessible, and the need ies (Economics and Business) from Oxford Brookes
for larger samples gets stronger (i.e., reproduci- and is an Innovate UK Women in Innovation Award
bility), games are going to be an important aspect Winner. Jo’s mission is to provide behavioral scien-
of scaling up experimental, social, behavioral, and tists with the tools and confidence to successfully
economic research. take their research online.

Conclusions References

Herein, we have reviewed the role of gamification Bazazi, S., von Zimmermann, J., Bahrami, B., &
in behavioral science. We have endeavored to define Richardson, D. (2019). Self-serving incentives
gamification and outline the different elements that impair collective decisions by increasing con-
can be considered when creating behavioral science formity. PLOS ONE, 14(11), e0224725. https://doi.
games. We have also provided examples of three org/10.1371/journal.pone.0224725.
different behavioral science games.9 We propose Butterworth, B., Varma, S., & Laurillard, D (2011).
that gamification will increase engagement and re- Dyscalculia: From brain to education. Science,
tention in online behavioral science studies. How- 332(6033), 1049-1053.
ever, one must consider whether this will in some Chierchia, G., Fuhrmann, D., Knoll, L. J., Pi-Sunyer,
way affect the data being collected. Anecdotal evi- B. P., Sakhardande, A. L., & Blakemore,
dence from researchers and participants suggests S.-J. (2019). The matrix reasoning item bank
that the benefits of employing gamification and (MaRs-IB): Novel, open-access abstract rea-
game-based learning far outweigh these concerns. soning items for adolescents and adults. Royal
Society Open Science, 6(10), 190232. https://doi.
The Authors org/10.1098/rsos.190232.
Deterding, S., Dixon, D., Khaled, R., & Nacke, L.
Joshua Balsters, PhD, has worked in both industry (2011). From game design elements to gameful-
(GlaxoSmithKline, Nielsen, Cauldron Science) and ness: Defining ‘gamification’. Proceedings of the
academia (Trinity College Dublin, ETH Zürich, Roy- 15th International Academic MindTrek Conference:
Envisioning Future Media Environments (MindTrek
Videos of these games and more can be found here:
9
’11), 9-15.
https://gorilla.sc/success/consulting/. Kahneman, D., & Tversky, A. (1979). Prospect theory:

Behavioral Economics Guide 2021 140

Joshua Balsters & Jo Evershed Gamification in Behavioral Science: An Engaging Prospect for Online Research

An analysis of decision under risk. Econometrica, Gamification by design: Implementing game me-
47, 263-291. chanics in web and mobile apps. O’Reilly Media,
Kankanhalli, A., Taher, M., Cavusoglu, H., & Kim, Inc.
S. H. (2012). Gamification: A new paradigm for
online user engagement. Proceedings of the Thirty
Third International Conference on Information
Systems, 3573-3582.
Kingstone, A., Smilek, D., and Eastwood, J. D. (2008).
Cognitive ethology: A new approach for studying
human cognition. British Journal of Psychology,
99(3), 317-340.
Looyestyn, J., Kernot, J., Boshoff, K., Ryan, J., Edney,
S., & Maher, C. (2017). Does gamification increase
engagement with online programs? A systematic
review. PLOS ONE, 12(3), e0173403. https://
doi.org/10.1371/journal.pone.0173403.
Mekler, E. D., Brühlmann, F., Tuch, A. N., & Opwis,
K. (2017). Towards understanding the effects
of individual gamification elements on intrin-
sic motivation and performance. Computers in
Human Behavior, 71, 525-553.
Montague, P. R., Dolan, R. J., Friston, K. J., & Dayan,
P. (2012). Computational psychiatry. Trends in
Cognitive Sciences, 16(1), 72-80.
Neville, F. G., Drury, J., Reicher, S. D., Choudhury,
S., Stott, C., Ball, R., & Richardson, D. C. (2020).
Self-categorization as a basis of behavioural
mimicry: Experiments in The Hive. PLOS ONE,
15(10), e0241227.
Palan, S. & Schitter, C. (2018). Prolific.ac: A subject
pool for online experiments. Journal of Behavioral
and Experimental Finance, 17, 22-27.
Rangel, A., Camerer, C., & Montague, P. R. (2008).
A framework for studying the neurobiology of
value-based decision making. Nature Reviews
Neuroscience, 9(7), 545-556.
Shamay-Tsoory, S. G., & Mendelsohn, A. (2019).
Real-life neuroscience: an ecological approach
to brain and behavior research. Perspectives on
Psychological Science, 14, 841-859.
Toda, A. M., Klock, A. C. T., Oliveira, W., Palomino,
P. T., Rodrigues, L., Shi, L., Bittencourt, I.,
Gasparini, I., Isotani, S., & Cristea, A. I. (2019).
Analysing gamification elements in educational
environments using an existing gamification
taxonomy. Smart Learning Environments, 6(1), 16.
Zichermann, G., & Cunningham, C. (2011).

Behavioral Economics Guide 2021 141

 No Contract, No Problem: The Value of Flexibility
in the Subscription Economy
Benny Cheung,* Abbie Letherby and Alice Pearce
Dectech

Across the world, companies are adopting the subscription model to increase sales and create more pre-
dictable revenue streams. Given this rise in popularity, we conducted an experiment with a sample size
of around 1,400 UK citizens, to investigate subscription services. Our aim was to explore how brands can
better acquire and retain subscription customers. We looked at three subscription types (Content, Service,
Product) across six different industries (Food, Delivery, Health & Beauty, Music, Fitness, TV & Film) and
measured their current appeal and avenues for improvement. Our results highlight key acquisition drivers,
including free trials, discount periods and bundles, the success of which is founded in understanding con-
sumer behavioural biases. We also explore the impact of different benefits on churn and how to improve
retention. The research was conducted prior to the pandemic, so while base levels will have been affected by
lockdowns, the insights are still applicable.

Introduction trials reduce customer lifetime value. This report

looks at free trials and whether alternative tactics,
The subscriptions model is being adopted by like initial price discounting, are more valuable.
brands across the world. Netflix and Amazon have The subscription model is not right for every busi-
shown how lucrative Content and Service subscrip- ness. Research (Dodds, 2019) finds venture capital
tions can be, and Hello Fresh is a prime example of deals for physical subscription boxes fell by more
the increasingly prominent brands looking to shake than half between 2015 and 2018. So why are di-
up the product market. rect-to-consumer (D2C) product subscriptions not
Traditional businesses are also pivoting towards enjoying the success many predicted? We explore
subscriptions, uprooting their previously success- possible reasons and look at the more successful
ful business models. In 2010, for instance, Microsoft Content and Service offerings, to unpick what dis-
unveiled its Office 365 package, offering monthly or tinguishes the varying degrees of success of each of
annual payments. This represents an extraordinary the three main subscription types.
step change from its one-time licence model, which
has operated since 1990. Furthermore, the decision Chapter One: Market Landscape
appears to have been vindicated, as Office 365 grew
27% year-on-year in Q4 2019 (Wilhelm, 2020). Talk of a “subscription economy” is not overblown.
Behavioural science would suggest that the flex- More than four in five UK adults hold at least one type
ibility for risk-averse customers to cancel their of subscription. But the youngest kind – Product – is
subscription at any time has helped with customer struggling to prosper.
acquisition, but can it help retention, too? Having surveyed the subscription landscape, we
Elsewhere, Netflix’s decision to end its offer of classified subscriptions into three distinct types:
free trials may spark a trend in the subscription Content, Service and Product (Table 1).
industry. Presumably, Netflix has found that free

* Corresponding author: [email protected].

Behavioral Economics Guide 2021 142

Benny Cheung et al. The Value of Flexibility in the Subscription Economy

Subscription
Definition Example Industry Example Providers
Type

Hello Fresh
Provides customers with a physical Food
Mindful Chef
Product product(s), usually delivered to their door
BirchBox
at regular intervals Health & Beauty
BeautyPie

Amazon Prime
Provides customers with unlimited use Delivery
Tesco Delivery
Service of a specific service for the duration for
PureGym
which they pay Fitness
Fitness First

Netflix
Provides customer with unlimited access TV & Film
Amazon Prime Video
Content to premium online content for the dura-
Spotify
tion for which they pay Music
Apple Music

Table 1: Definition of subscription types and example providers.

The Subscription Market Landscape most striking finding is the low acquisition volume
of Product subscriptions. The 1% growth in health &
Sizing the market clearly shows that the subscrip- beauty and food recipe box subscriptions is half that
tion hype is well substantiated. Our research finds of Content and Service types.
83% of UK adults hold at least one type of subscrip- This runs counter to expectations, since a younger
tion, the most prevalent of which are Service and industry will typically grow at a faster rate. We posit
Content subscriptions (Figure 1). that one of the reasons why D2C propositions are
The most popular combination of subscription not experiencing strong acquisition growth is due
types is Service and Content, held by almost 17m to a lack of intrinsic demand. For instance, brands
consumers (32%). Meanwhile, one in five UK adults that offer grooming goods, which can be bought at a
hold at least one of each type of subscription. To- range of other outlets, do not offer anything unique.
gether, these two combinations are held by more Another reason is simply the frequency of D2C
than half (52%) of the population. deliveries. Customers with Product subscriptions
Overall, Product subscriptions (13.6m) are not as typically receive their goods once a month, which
widespread as Content (30.4m) or Service (36.7m), tends to coincide with the payment cycle. Regular
which is partly because Product subscriptions are a deliveries may prompt customers to review their
comparatively newer offering. subscription and ultimately cut their expenditure,
particularly if the frequency is too great and cus-
Products Performing Poorly tomers don’t require another delivery within that
time frame. Depending on the value to customers,
Looking at the percentage of UK adults who sign this might trigger them to unsubscribe.
up to a subscription over 12 months – known as Furthermore, for D2C subscriptions, consumer
“acquisition growth” – is a useful lens through guilt around paying for goods that they do not use
which to assess the subscription landscape. When is a problem. Whether motivated by environmen-
looking at the acquisition growth of six different tal concerns or financial burden, it tends to result
industries within the three subscription types, our in higher churn. Finally, there is a distinct lack of

Behavioral Economics Guide 2021 143

Benny Cheung et al. The Value of Flexibility in the Subscription Economy

choice when it comes to goods available via D2C leader, Amazon Prime, offers a rolling contract. As
subscriptions. As a nascent industry, this is not of October 2018, 26% of UK adults were estimated
surprising, but it is hampering potential acquisition to be Amazon Prime members (Mintel Press Team,
growth. 2019), which is an increase of roughly 10 percent-
age points since 2017 (Annicelli, 2017).1 We expect
No Contract, No Problem in addition to Prime’s fast delivery and unrivalled
product offering, the no-contract subscription is
Remarkably, almost half (47%) of the UK adult attracting new customers.
population has a TV & film subscription. One in-
strumental factor behind the popularity of these Chapter Two: Acquisition Drivers
subscriptions is that they usually offer a fixed con-
tract, thereby improving retention rates. However, Subscription perks are now commonplace. Graze, the
our research indicates that disrupter brands like snack box company, is well known for its free box of-
Netflix and Amazon Video, which offer rolling con- fer. Referral and loyalty gifts are also tried and tested
tracts, contribute disproportionally to the success techniques to entice new customers. We suggest several
of Content and Service subscriptions (see Chapters other ways that brands can improve their acquisition
Three and Five). rate.
Whilst there is not much difference between ac- Our randomised controlled trial (RCT) experiment
quisition rates across the six industries, we think it tested some of the most common benefits that
is telling that delivery service subscriptions have the brands offer potential customers, including initial
highest acquisition growth (3%), when the market discounting, loyalty points and the ability to pause

Figure 1: Market size and percentage of UK adult subscribers by subscription type.

the subscription. We find that none of the benefits March 2017 and compared this figure to the Mintel claim
has a significant impact on a consumer’s propen- that 26% of the UK adult population was subscribed to
sity to purchase a subscription beyond reducing the Amazon Prime by October 2018. Based on the latest ONS
price. adult population estimates, 8 million adults represent-
ed roughly 16% of the UK adult population in 2017. This
1
Note: the 10 percentage point rise is based on our calcula- equates to approximately a 10 percentage point difference
tion, which used the 8m UK Amazon Prime subscribers in between March 2017 and October 2018.

Behavioral Economics Guide 2021 144

Benny Cheung et al. The Value of Flexibility in the Subscription Economy

Industry-Specific Subscription Hooks for every 100 consumers who see the offer.Mean-
while, consumers are 12% more likely to sign up
When looking at each industry individually, we to a TV & film subscription if offered a no-contract
found two significant acquisition drivers for super- deal. This approach has helped the likes of Netflix
market delivery and TV & film subscriptions, i.e. a and Amazon Prime Video become two of the most
free trial period increases the likelihood of a cus- popular on-demand video providers in the UK. Both
tomer subscribing to a supermarket delivery service platforms offer a rolling monthly contract, which
by 8%. This translates into 10 additional customers has seen them accrue almost 10m and 8m UK cus-

Figure 2: Net growth of subscribers in six industries.

tomers, respectively (Ampere Analysis Ltd, 2019). tomer base, but this low-balling may also affect
customer retention, should brands increase their
Extending Discount Periods price later. Therefore, brands should run their own
price modelling, to assess the likely outcomes of
Beyond reducing the price of their subscription, this approach.
brands can attract new customers by extending the
discount period. This would involve keeping the The ‘Freemium’ Model
subscription price low in the first year and increas-
ing the cost in subsequent years, in order to catch Another way to entice subscribers is by offering a
up on customer lifetime value. Consumers are tem- basic level of service for free with an option to up-
poral discounters and value price reductions early grade to a ‘premium’ offering for a subscription fee
on. – a model that has worked successfully for Spotify.
However, this may lead to a race to the bottom, The attractiveness of experiencing a small taste
whereby brands have to compete for price-con- of a brand’s offering with little commitment is
scious consumers. Consequently, brands that offer central to the success of the ‘Freemium’ model.
the lowest price are more likely to grow their cus- Psychologically, it works in a similar way to try-be-

Behavioral Economics Guide 2021 145

Benny Cheung et al. The Value of Flexibility in the Subscription Economy

Figure 3: Free trial vs. no contract as an acquisition lever for subscriptions. Note: Results shown for industries with
statistically significant differences at a 95% confidence level.

fore-you-buy promotions. The notion of free is also ing up to Virgin Broadband. Magazines, especially
consistently overvalued by consumers, who place paper-based ones in an era of digital-first, have
far greater value in a change in price from just 1p also discounted subscriptions to tempt consumers
to 0p than the equivalent change of 2p to 1p (Ari- into signing up.
ely & Shampan’er, 2006). For brands, ‘Freemium’
subscriptions help them reach more consumers, Bundles of Value
which is the first challenge in trying to acquire new
subscribers, but they also increase the likelihood Nevertheless, some companies may be able to go
that subscribers will upgrade their subscription at a further than this and bundle multiple subscriptions
cost. Crucially, brands need to balance the appeal of together. In the telecoms industry, this bundling
the free service with that of the premium, to ensure is referred to as “quad-play”, whereby consum-
sufficient volumes convert to a paid subscription. ers sign up to landline, broadband, TV and mobile
services through one subscription. One of the main
Perception Is Reality attractions of this proposition for customers is that
they have fewer suppliers to deal with and may
Contrary to the perceived wisdom behind the stand to benefit financially from a multi-package
numerous subscriptions that offer sign-up bene- deal.
fits, our experiment does not find that this popular Not only can this proposition help brands win new
marketing tactic is effective at acquiring customers. customers, but for firms with sufficient resources
However, we expect brands that offer extra sub- it can act as a retention driver, too. In the case of
scription benefits in the first year will likely improve quad-play, a customer who is considering leaving
consumer price perception, potentially attracting would have to reckon with losing four useful servic-
new customers. es – and potentially have to replace one subscrip-
A number of brands employ this method, includ- tion with four separate ones.
ing Virgin Media, which offer discounted SIM-only
plans as a benefit for consumers considering sign-

Behavioral Economics Guide 2021 146

Benny Cheung et al. The Value of Flexibility in the Subscription Economy

The Importance of Consumer Psychology customers. Done well, the right framing can signal
to consumers the superior quality of a brand’s of-
Furthermore, brands that incorporate consid- fering.
erations about consumer psychology into their
subscription propositions will most likely improve Chapter Three: Churn Landscape
their acquisition rates. Our suggestions to introduce
extended subscription benefits, as well as bundle What do intrinsic demand, frequency of contact and
propositions where possible, affects people’s risk negative feedback have in common? They all threaten
calculation (Samson, 2017). Consequently, con- the viability of direct-to-customer Product subscrip-
sumers are more likely to sign up to a subscription. tions. Despite the fanfare for D2C propositions, our
Framing propositions in these ways can attract new research shows an alarmingly high churn rate.

Figure 4: Annual churn of subscription types across six industries.

Failing to Deliver efficiency, our experiment indicates that customers

are not convinced.
The customer retention rates experienced by the
more established industries we tested, such as TV & Keeping It Exclusive
film and delivery service subscriptions, fare much
better than new industries such as health & beauty. Only 13% of TV & film subscribers cancel their
In fact, the churn landscape paints a bleak picture subscription after 12 months, which is the lowest
for Product subscriptions. Over a 12-month peri- annual churn we found across the six industries.
od, we find the highest churn rates across the six One of the reasons why the churn rate for Content
industries are for health & beauty and food recipe subscriptions is so much lower than that of D2C
boxes (31% and 38%, respectively). propositions is because of a lack of intrinsic de-
Whilst D2C subscriptions are meant to offer cus- mand.
tomers a more personalised service and greater Consumers are attracted to the exclusive content

Behavioral Economics Guide 2021 147

Benny Cheung et al. The Value of Flexibility in the Subscription Economy

provided by platforms such as Netflix and Prime en. Whilst the subscription may be forgotten about
Video. Netflix’s Original shows are incredibly pop- between deliveries, the frequency and physical de-
ular – its “Stranger Things” production alone was livery of D2C propositions reminds people of their
watched by 64m people worldwide (Mendelson, subscription on a regular basis. This may prompt
2019). Similarly, some products are only available them to review and cancel their subscription.
to buy on Amazon, which helps keep the churn rate In contrast, there are fewer touchpoints with
as low as 19% for delivery service subscriptions. Content and Service subscriptions, and, impor-
The inherent demand that brands like these have tantly, these interactions do not coincide with the
baked into their business models does not apply to payment cycle. This helps keep the churn rate low.
many D2C propositions. Instead, a great deal of the Touchpoint frequency does not necessarily affect
same products can be bought elsewhere, without the likelihood of cancelling Content or Service sub-
the need for a subscription, which consequently scriptions: high frequency interaction implies the
reduces consumer demand for the product. Brands subscription is valuable to the consumer and so
should therefore carefully consider whether their reduces the chance of churn, while low frequency
product is optimised for D2C subscriptions. interaction may see the consumer forget about the
subscription and so, without any reminder to can-
Avoiding Weak Spots cel, be less likely to churn.
Additionally, consumers can utilise their Content
Another reason for the much higher churn rates and Service subscriptions at any point during the
of Product subscriptions is the regular touchpoints payment cycle. Consequently, consumers’ inter-
with consumers on delivery. Customers usually actions with these subscriptions are more posi-
receive their D2C product after a payment is tak- tive. Even with a Service subscription like Amazon

Figure 5: Churn rates across all industries.

Prime, where a physical product is delivered, it is between the customer and their subscription.Deliv-
the service, not the physical good, that is being paid ering Negative Feedback
for. Therefore, the ease and convenience of such Consumers’ interactions with physical products
Service subscriptions provide positive interactions offer a further reason as to why Product subscrip-

Behavioral Economics Guide 2021 148

Benny Cheung et al. The Value of Flexibility in the Subscription Economy

tions experience high churn rates. Consumers who greater variety means that they have more opportu-
are not persuaded by the value of their subscription nities to match their experience of the subscription
receive regular reminders to judge the quality of with their preferences.
the same good. Without continuous reassessments,
human inertia would likely see consumers continu- De-Risking Churn Rates
ing the subscription.
For example, a customer who signs up to a men’s Strikingly, our analysis of the behavioural survey
grooming product may realise after a few deliveries finds that the rate of subscription cancellations at
that the quality of the product is not good value. Un- a given time – the churn rate – tends to run above
surprisingly, consumers with similar experiences average every 12 months for all industries we tested
are more likely to cancel their Product subscription. (see Figure 5). With many subscriptions lasting 12
Importantly, whilst consumers also frequently months, brands that automatically enrol customers
interact with Content and Service subscriptions, the onto a new subscription may improve their cus-

Figure 6: Churn rates for subscriptions that include an initial discount vs. free trial across all industries.

tomer retention rates. people to cancel their subscription, such as cost,

quality of service and whether a better alternative
Chapter Four: Retention Levers was available. We found that an initial discount to
the subscription price and not locking customers
The two main factors that reduce the rate of churn into a contract have the greatest effect on reducing
are an initial discount and, somewhat paradoxical- churn.
ly, no-contract subscriptions. Our research finds that
greater flexibility encourages more customers to stay The Best Things in Life Aren’t Free
with brands.
Our behavioural survey was designed so that we The findings reveal an initial discount offer is the
could determine the factors most likely to cause most powerful lever for customer retention, halving

Behavioral Economics Guide 2021 149

Benny Cheung et al. The Value of Flexibility in the Subscription Economy

annual churn from 23% to 12% across all subscrip- ly to cancel a subscription with a free trial than one
tion types in our research. What is also noteworthy with a discount in the first month.
is the significant impact of a discount period on
health & beauty and food recipe box subscriptions Loyal Customers Love Freedom
(-8% and -4%, respectively), both of which expe-
rience high churn rates. A key insight to emerge from our survey is the
Conversely, a free trial period leads to much draw of no-contract subscriptions, which allow
greater churn during the first few months of the customers to cancel their contract at any time,
subscription. Consumers are three times more like- contradicting commercial wisdom on contracts and

Figure 7: Annual churn for contract vs. no-contract subscriptions. Note: Results shown for industries with
statistically significant differences at a 95% confidence level.

consumer behaviour. A no-contract offer also influences consumers’

In contrast with longer-term contracts, we found perceptions of the product, service or content on
that giving customers the flexibility to cancel their offer. By offering a flexible subscription, the brand
subscription at any time significantly reduces can- signals to the consumer its superior quality. More
cellation rates. Across six industries, we find a 2% broadly, while subscriptions do often improve re-
decrease in customer churn over 12 months for such tention, our research suggests that a rolling month-
subscriptions. ly contract is a better subscription model that will
While the unpredictability of no-contracts may benefit companies in the long term.
sound counterintuitive from a business perspective,
rolling contracts actually avoid the regular remind- Recommendations
ers that define contract subscriptions, whereby
each payment period acts as a trigger for customers 1. Roll out no-contract subscriptions for better
to reconsider their subscription. Clearly, this re- retention. By offering no-contract subscriptions,
consideration can increase customer churn. brands will see lower churn rates. Some sectors,

Behavioral Economics Guide 2021 150

Benny Cheung et al. The Value of Flexibility in the Subscription Economy

such as TV & film subscriptions, can also benefit tistically modelling whether the addition of certain
from better customer acquisition rates. benefits influenced purchase intentions.
2. Make trial periods cheaper rather than free. Sub-
scriptions with an initial discount will typically Behavioural Survey
lead to better customer retention than a free tri-
al. Unless the free trial yields enough acquisition Participants were also required to answer a short
uplift, the discount approach offers better unit behavioural survey about the subscription prod-
economics. ucts they currently have or have had in the last 12
3. Look before you jump with D2C Product sub- months, which benefits were included and whether
scriptions. With higher churn and harder acqui- they have cancelled any subscriptions, when and for
sition, Product subscription providers should be what reason.
sure of their business case before launch. Down- This information was used to determine the per-
stream cross-sales or a halo impact on traditional centage of UK adults who hold multiple subscription
pay-as-you-go retail sales may offer some miti- types, one type of subscription or no subscription.
gation of poor standalone economics. We also sized the subscription market at an indus-
4. Know your target audience. Effective targeting try level and compared it to the annual customer
will substantially lower customer acquisition acquisition rates across the six industries explored
costs. Brands should own the problem of direct herein. Using churn data, we created the hazard
marketing optimisation, rather than delegating rates for each industry.
this to a digital agency. Crucially, this targeting
should incorporate forecast churn, and hence Survival Analysis
customer lifetime value, rather than simply sales
and margin. Survival analysis was conducted to understand
customer retention across different subscription
Detailed Methodology products and the impact of different benefits on re-
tention. A broader range of industries was included,
Behaviourlab Paradigm in order to increase the reliability of the results, but
with the key six industries explored in detail.
Behaviourlab is our bespoke online test platform The parameter of estimation was the number of
that utilises randomised controlled trials to answer months the subscription was held. The impacts of
key commercial questions. The method follows customer characteristics and proposition benefits
modern academic standards of eliciting consumer were modelled, to predict the type of customer and
preferences. elements of a proposition that significantly improve
This research involved taking participants retention.
through a realistic simulation of purchasing a
product through a retailer’s website. Each partici- The Authors
pant was required to make three purchases for each
of the three subscription types selected at random Benny Cheung is a Director at Dectech. He com-
from the six industries explored. pleted a PhD and a two-year research fellowship in
The propositions were all unbranded so that the behavioural science at the University of Cambridge
impact of different acquisition levers could be meas- before joining Dectech in 2005. His areas of com-
ured cleanly, without the influence of branding. The mercial expertise include retail, utility, financial
subscription price was varied at random around the services, telecom, ecommerce and advertising. He
current market price. Six different sign-up benefits also heads the firm’s Brand and Fraud Prevention
were also tested, with either one randomly shown Practices, overseeing all related client accounts and
or none. Participants were required to indicate their internal R&D initiatives.
likelihood to purchase. The analysis involved sta-

Behavioral Economics Guide 2021 151

Benny Cheung et al. The Value of Flexibility in the Subscription Economy

Abbie Letherby is a Senior Associate at Dectech. Samson, A. (2017). Mental money: The psy-
She has a BSc in Psychology and six years’ experi- chology of subscription payment options.
ence in conducting behavioural science research to BehavioralEconomics.com. https://www.behav-
solve clients’ commercial problems. Abbie designs ioraleconomics.com/mental-money-the-psy-
randomised controlled trials to optimise proposi- chology-of-subscription-payment-options/.
tions, customer communications and brand strat- Wilhelm, A. (2020, January 29). Microsoft shares
egy, in order to deliver actionable insights to senior rise after it beats revenue, profit expectations,
business stakeholders. Azure posts 62% growth. TechCrunch. https://
techcrunch.com/2020/01/29/microsoft-shares-
Alice Pearce is an Associate Research Analyst at rise-after-it-beats-revenue-profit-expecta-
Dectech. She holds an MSc in Behavioural & Eco- tions-azure-posts-62-growth/.
nomic Science at The University of Warwick, having
previously completed a BSc in Economics. Alice is
passionate about behavioural science insights and
has experience in market research, nudge design
and experiment implementation and analysis.

References:

Ampere Analysis Ltd. (2019). The UK VoD market:

Current status and future developments. Ofcom.
Annicelli, C. (2017, April 20). Amazon Prime has
room to grow in the UK. Business Insider. https://
www.businessinsider.com/amazon-prime-has-
room-to-grow-in-the-uk-2017-4.
Ariely, D., & Shampan’er, K. (2006). How small is
zero price? The true value of free products. SSRN.
https://papers.ssrn.com/sol3/papers.cfm?ab-
stract_id=951742.
Dodds, L. (2019, October 12). Rise of the subscrip-
tion economy – and what that means for your
pocket. The Telegraph. https://www.telegraph.
co.uk/technology/2019/10/12/rise-subscrip-
tion-economy-means-pocket/.
Mendelson, S. (2019, October 16). Netflix claims
64 million watched ‘Stranger Things’ Season
3 in the first month. Forbes. https://www.
forbes.com/sites/scottmendelson/2019/10/16/
stranger-things-season-3-netflix-ratings-64-
million-marvel-avengers-harry-potter-wino-
na-ryder-lost-x-files-mulder-scully-fringe-
wonder-woman/.
Mintel Press Team. (2019, March 7). The Amazon
effect - Nine in ten Brits shop on Amazon.
Mintel. https://www.mintel.com/press-centre/
retail-press-centre/the-amazon-effect-nine-
in-ten-brits-shop-on-amazon.

Behavioral Economics Guide 2021 152

 Career Coaching Builds Confidence in
Experienced Workers: What Governments Should
Be Thinking About for Reskilling Initiatives Post-
COVID – and Why
Gerard Creaner, Sinead Creaner,* Claire Wilson and Colm Creaner
GetReskilled

With the end of the COVID pandemic in sight for first-world countries, government focus will shift to re-
building economies. This will likely involve workforce reskilling initiatives for new jobs in thriving economic
areas. The authors have initial data to suggest that adding career coaching to technical training programmes
builds confidence in experienced workers, and results in an overall positive impact on their outlook regard-
ing the future. This is an unanticipated “bonus effect” in addition to the classic tangible metrics of reskilling
programmes. Since 2017, Confidence Metrics have been measured across 1,074 workers and, surprisingly,
been found to remain high during 2020 despite the economic fallout from COVID. The hypothesis relating to
the “bonus effects” of more confident and resilient workers needs further study, but if shown to be reliable,
including job-hunting skills within reskilling programmes, it could prove useful to both governments and
the private sector alike.

Introduction This paper offers insights into how to maxim-

ise the efficiency and effectiveness of such pro-
As vaccination roll out progresses (Public Health grammes.
England, 2021) and cases begin to fall (Office for
National Statistics, 2021), the COVID-19 pandemic Relevant Background Information
in first-world countries is changing. Government
focus will soon move from tackling a healthcare cri- The demographics of the 1,074 experienced work-
sis, to tackling the looming economic one (Beckett, ers (2017-2020) are as follows:
2021). A large part of this effort will involve retrain-
ing displaced workers for new jobs in thriving local • Based in Ireland
industries. • Taking same reskilling course
This paper examines the practical experience • 38% female; 62% male
and research background of one private training • 45% <40 years; 55% >40 years
provider with over 10 years’ experience reskilling • 28% employed; 72% unemployed
mid-career workers for employment in a growing
technical and highly regulated industry. They came from a range of other industries, in-
What’s more, that experience is in an online en- cluding food and beverage, finance, administration,
vironment and working alongside a long-running, healthcare and construction.
currently active government retraining scheme All received Irish Government funding through
(Higher Education Authority Ireland, 2021). the Springboard+ reskilling initiative (Higher Edu-
cation Authority Ireland, 2016), whereby the gov-
* Corresponding author: ernment paid either 90% or 100% of their course
[email protected] fees.

Behavioral Economics Guide 2021 153

Gerard Creaner et al. Career Coaching Builds Confidence in Experienced Workers

Our reskilling programmes have been delivered learning, which included a “Job-Hunting Skills”
online for over 10 years now. The team utilises an module for the experienced workers on our pro-
action research approach to updating course of- grammes.
ferings in response to new evidence or participant The module was designed to walk the experi-
feedback, whereby the authors work through a enced worker through finding their ideal job within
seven-step process in a continuous cycle – as per this new sector (aligning with relevant skills from
Ernest Stringer’s Action Research Interacting Spiral their work experience to date) and then through
(Stringer, 2007). the job-hunting process, step-by-step – including
Due to the online delivery method, programmes where to look, how to network, how to write a CV
were not impacted by the COVID-19 pandemic or and understanding key interview skills.
the associated social distancing requirements. This After spending five weeks studying guided ac-
is particularly relevant, since Ireland spent the tivities (approximately 50 hours of study time),
majority of time from March 2020 onward in level workers would submit an assignment that sim-
5 lockdown due to the ongoing pandemic (Depart- ulated a job application and recruitment process.
ment of the Taoiseach, 2020). They were given three real historical job adverts to
The programmes are all designed to transition choose from and complete tasks that included tai-
experienced workers from other industries into the loring their CV, writing a cover letter, considering
pharmaceutical and medical device manufacturing their relevant network contacts, writing follow-up
sector, which offers high-paying, high-tech jobs in emails and scripting sample interview answers as if
a stable, secure and growing sector in Ireland (Hal- they were applying for one of these jobs.
ligan, 2016).
During the pandemic, the pharmaceutical and Assessing the Effectiveness of Job-Hunting
medical device manufacturing industry in Ireland Skills
remained open, and hiring (Creaner, 2020). There
are over 62,000 people employed directly by this We assess the effectiveness of this Job-Hunting
sector, and a further 120,000 employed indirectly Skills module in four separate ways:
(GetReskilled, 2021). In total, this accounts for al-
most 10% of the Irish workforce (IDA Ireland, 2021). 1. Knowledge and Ability to Implement: Using a
multiple-choice assessment of 10 questions,
Adding Career Coaching to Technical participants choose which answer most close-
Modules ly matches their approach to various aspects of
job-hunting, following the action mapping ap-
Our business was initially delivering technical proach suggested by Cathy Moore (2007). Scores
modules to provide experienced workers from other do not count towards their grade in the module,
sectors with the industry-specific information they and participants are asked to answer honestly.
needed to secure an entry-level role. Scores show how well their answers align with
In 2016, we realised there was more we could do to current best practice job-hunting advice. This is
improve the success rate of our participants secur- the first task in the module and is revisited as one
ing employment in the industry. of the final tasks, to capture the change in their
We observed that many workers had poor knowledge and understanding.
job-hunting skills, particularly when moving into 2. Module Feedback: The experienced workers give
a new or an unfamiliar sector. While almost every- both quantitative (“star rating” out of 5, in re-
one thinks they know how to job hunt, it’s our ex- sponse to the question “How would you rate your
perience that very few can actually implement best overall experience of the Advanced Coaching pro-
practice in this regard. gramme so far?”) and qualitative feedback at the
To overcome this issue, we developed a career end of the Job-Hunting Skills module.
coaching strategy to complement the technical 3. Successful Outcomes: We measure “successful

Behavioral Economics Guide 2021 154

Gerard Creaner et al. Career Coaching Builds Confidence in Experienced Workers

outcomes” for all experienced workers on our Pre-Pandemic Group 2017-2019:

programmes. The successful outcomes are de- Establishing the Baseline
fined as a participant either finishing the course,
getting a job or both. This snapshot of success is Our pre-pandemic group consists of a total of 923
taken within six months of the course finishing. experienced workers: 329 in 2017, 315 in 2018 and
4. Confidence Metrics: We capture feedback from 279 in 2019.
participants at the end of their reskilling pro- A total of 597 of these completed the end-of-pro-
gramme – a key feature of this is a range of “Con- gramme survey: 217 in 2017, 197 in 2018 and 183 in
fidence Metrics” through which the participants’ 2019.
confidence is assessed across a range of future In 2017, the Job-Hunting Skills module was intro-
aspirations. duced as an optional module, with 20% of experi-
enced workers opting in.
The rest of this paper will first discuss the baseline In the latter half of 2018, and after successful
for these effectiveness measures, by considering proof of concept, the module was made mandatory.
results from a “pre-pandemic group” (experienced Since the 2018 group includes a mix of both opt-in
workers reskilling between 2017 and 2019), before and mandatory module participants, a total of 57%
assessing how that effectiveness has been affected of all 2018 participants completed the module.
by the pandemic, by outlining the results of our In 2019, the module was made mandatory for all
“pandemic cohort” (experienced workers reskilling experienced workers, so 100% of participants com-
during 2020). pleted it.

Figure 1: Results of knowledge and ability to implement assessment (2018-2019).

Behavioral Economics Guide 2021 155

Gerard Creaner et al. Career Coaching Builds Confidence in Experienced Workers

The effectiveness measures were as follows: 2. Module Feedback

1. Knowledge and Ability to Implement Table 1 shows the average results of the partici-
pant star ratings (out of 5) which were captured at
The multiple-choice assessment was introduced the end of module feedback from a total of 182 par-
within the module in 2018, so the data for Figure 1 ticipants during 2018 and 232 during 2019. Module
are gathered from 2018 and 2019 participants only. feedback was only implemented in 2018, so there
From Figure 1, we can see a change from the week are no data available for the 2017 participants.
one assessment scoring of experienced workers’ We can see from both qualitative and quantitative
knowledge and ability to implement best practice participant feedback that the experienced workers
job-hunting techniques (mean=3.85) and their enjoyed the module, with many saying they found
week five scores (mean=8.51), which is statistically more value than they expected.
significant (t=30.89, p<0.05). From analysing the qualitative answers and
This suggests that the module is successful- talking to participants, we established that there
ly equipping participants with knowledge of appears to be a common theme in terms of experi-
job-hunting best practice and the ability to imple- encing of the module:
ment it successfully.
Moreover, the initial assessment serves as an an- • Believing they already know best practice at the
chor by which the experienced worker can see their start of the module
own progress over the course of the module, there- • Being shocked at how low they score on the
by serving to increase self-belief in their new skills. knowledge and ability assessment – the first task

2017 2018 (n=182) 2019 (n=232)

Quantitative Feedback: N/A 4.29 4.18

Average Star Rating out of 5

Qualitative Feedback: “When I started this module, I thought I knew something about job-hunting, but after
completing this module I realised I really didn’t know anything.”

Participant Feedback “This module is amazing! I never realised how many mistakes I made before as regards
job-hunting. I feel so much more confident now that I have completed this course, and the
process will stay with me forever.”

Table 1: Pre-pandemic module feedback.

in week one We were struck by how similar this common ex-

• Approaching the module with a much more open perience mirrors that of a classic Delusion of Com-
mind to learning new techniques for job-hunting petence (Kruger & Dunning, 1999), and how effec-
• Taking on board these techniques, practising tively the module appears to help move participants
them and gaining confidence in how to imple- through each stage.
ment them This was particularly interesting given that the
• Reaching a level of confidence that now accurately optional opt-in was low (20%) – even when we told
reflects their knowledge and ability to implement people that previous participants found great value
best practice job-hunting techniques in the module. It seemed like most believed they

Behavioral Economics Guide 2021 156

Gerard Creaner et al. Career Coaching Builds Confidence in Experienced Workers

were the exception that truly already knew best programme.

practice.
We therefore took the decision to make the
Job-Hunting Skills module a mandatory part of the

2017 2018 2019

Finished full programme 142 (65%) 140 (71%) 118 (64%)

Got job 115 (52%) 90 (45%) 83 (45%)

… … … …
specifically in pharma/med
device 58 (50%) 56 (62%) 54 (65%)

Table 2: Pre-pandemic successful outcomes.

3. Successful Outcomes or decreased as a result of participating in the pro-

gramme.1
To compile the dataset in Table 2, we analysed the The nine aspirations assessed were:
survey results from the three pre-pandemic years.
From the above, it is evident that the move from • Engagement with lifelong learning
optional participation in the Job-Hunting Skills • Ambition for career advancement
module to mandatory participation had a statis- • Ambition to get a better-paid job
tically significant impact of increasing the per- • Enjoyment of further study
centage of experienced workers who successfully • Confidence in the future
found a job in the pharmaceutical or medical device • Determination to succeed
industry within six months of the programme fin- • Motivation for getting a rewarding career
ishing (moving from 50% to 65%; Z=-2.11, p<0.05). • Confidence in future job security
We can deduce, therefore, that the Job-Hunt- • Ambition to achieve mastery in their chosen ca-
ing Skills module equips students with effective reer
job-hunting skills for their target industry and gives
them the confidence to pursue those high-tech jobs, For ease of interpretation, this data has been
even without previous industry experience. summarised into the average number of work-
As a result of this evidence, we were happy to con- ers reporting an increase in confidence, averaged
tinue with our decision to keep the module manda- across all nine metrics.
tory. While this was not a deciding factor in making

4. Confidence Metrics 1 While there was no option for participants to indicate

that a metric “stayed the same” (which we acknowledge
Confidence metrics were assessed during the as a limitation), they could skip any question they did not
end-of-programme survey for all three years. wish to answer. Future iterations of this survey will in-
Participants were asked whether their confidence clude a “stayed the same” option, to make this point more
across a range of future aspirations had increased explicit.

Behavioral Economics Guide 2021 157

Gerard Creaner et al. Career Coaching Builds Confidence in Experienced Workers

2017 (n=217) 2018 (n=197) 2019 (n=183)

Job-hunting module partic- 83% 92% 93%

ipants

Non-job-hunting module 80% 86% N/A as Job-Hunting Skills

participant module made mandatory

Table 3: Pre-pandemic confidence metrics.

or keeping the module mandatory, we could see a as well as more stressed and anxious during the
statistically significant increase in confidence when lockdowns (Office for National Statistics, 2021).
comparing those who participated in the Job-Hunt- The observed effectiveness measures for this
ing Skills module in 2018, against those who did not group were as follows:
do so (Z=2.42, p<0.05).
We hypothesised that the 2018 iteration of the 1. Knowledge and Ability to Implement
Job-Hunting Skills module was increasing confi-
dence by reducing participants’ uncertainty. The
Figure 2 shows results for the knowledge and abil-
module takes a large and complex process and
ity to implement assessment for the 2020 group.
breaks it down into manageable, bitesize pieces. We
As expected, we once again see a shift in the mean
believe that, in doing so, participants are left feeling
scores of the experienced workers’ knowledge and
like they understand the job-hunting process better
their ability to implement best practice job-hunt-
and are motivated by their own ability to complete
ing techniques between week one (mean=3.98) and
each piece successfully.
week five (mean=8.59), which is statistically signif-
Not only does this seem to give participants hope
icant (t=-18.93, p<0.05).
of finding a job during their current job search, but
This confirms that the module is still success-
we believe it also provides them with reassurance
fully equipping participants with knowledge on
in terms of their ability to job hunt successfully in
job-hunting best practice and the ability to im-
the future, if they need to do so. This, in turn, ap-
plement it successfully, independent of the unique
pears to have a positive effect across the range of
destabilising effects of the pandemic.
future-looking Confidence Metrics.

Pandemic Cohort 2020 – Impact of the 2. Module Feedback

Pandemic on the Baseline
Table 4 shows the average star ratings from the
Our pandemic cohort consists of 151 experienced end-of-module feedback from 97 experienced
workers, each of whom took a reskilling programme workers in 2020.
in 2020. All participants were given the Job-Hunt- Similarly to previous years, we can see from the
ing Skills module, and 97 of them completed the participant feedback that the experienced workers
end-of-programme survey. enjoyed the module and took a good deal of value
Of the four effectiveness measures, we expected from it.
the first three to remain relatively unchanged by the Qualitative feedback again shows that partici-
circumstances of the pandemic and its economic pants are still moving through a cycle of not initial-
fallout. We did, however, hypothesise that we’d see ly being motivated by the prospect of the module,
a notable decrease in the Confidence Metrics, as the being shocked at their low current knowledge of
UK Office for National Statistics was reporting that best practice, learning and practicing best practice,
the population felt more worried about the future, before reaching an appropriate level of confidence.

Behavioral Economics Guide 2021 158

Gerard Creaner et al. Career Coaching Builds Confidence in Experienced Workers

Figure 2: Results for the knowledge and ability to implement assessment (2020).

2017 (n=217) 2018 (n=197) 2019 (n=183) 2020 (n=97)

Quantitative Feedback: N/A 4.29 4.18 4.31

Average Star Rating

out of 5

Qualitative Feedback: “At first, I thought this would not be something important to me… but now I see it is even more
important than technical skills.”

Participant Feedback “When I first began this module, I was skeptical, because I thought my CV and cover letter were good
enough and I felt it was just the lack of experience as the reason why I wasn’t securing employment
in the pharmaceutical industry. Now that I have completed this, I can see how wrong I was.”

Table 4: Pandemic module feedback.

3. Successful Outcomes There was, however, a statistically significant

increase in the percentage of experienced workers
In Table 5, we analysed successful outcomes re- finishing the programme, from 64% in 2019 to 80%
ported in the end-of-programme survey. in 2020 (Z=-2.57, p<0.05). Since the programme is
As expected, there was no significant difference delivered fully online and many workers were fur-
in the percentage of people getting a job (in phar- loughed at home during this time, this is perhaps an
ma/med device or not) when we compared 2020 to unsurprising outcome.
2019, as the industry kept hiring throughout the
lockdowns.

Behavioral Economics Guide 2021 159

Gerard Creaner et al. Career Coaching Builds Confidence in Experienced Workers

4. Confidence Metrics been summarised into the average number of work-

ers reporting an increase in confidence, averaged
We assessed the same nine aspirations as previ- across all nine metrics.
ously detailed. Unexpectedly, our hypothesis of a fall in Confi-
As before, for ease of interpretation the data have dence Metrics was not seen in the data. In fact, the

2017 2018 2019 2020

Finished full pro- 142 (65%) 140 (71%) 118 (64%) 78 (80%)
gramme

Got job 115 (52%) 90 (45%) 83 (45%) 39 (40%)

… … … … ….
specifically in phar-
ma/med device 58 (50%) 56 (62%) 54 (65%) 23 (58%)

Table 5: Pandemic successful outcomes.

2017 (n=217) 2018 (n=197) 2019 (n=183) 2020 (n=97)

Job-hunting module 83% 92% 93% 93%

participants

Non-job-hunting 80% 86% N/A as Job-Hunting N/A as Job-Hunting

module participant Skills module made Skills module made
mandatory mandatory

Table 6: Pandemic confidence metrics

“bonus effects” of the Job-Hunting Skills module this surprising occurrence.

stayed the same as 2019 (at a high of 93%) and ap-
peared to be independent of the economic impact of Interpreting our Surprising “Pandemic
the pandemic. Data”
Our belief in the module being in the best inter-
ests of our participants, and the decision to make its Based on the reports from the UK Office for Na-
inclusion mandatory, means that we don’t have a tional Statistics, our team had hypothesised that
control group (who didn’t take the module) to com- our Confidence Metrics would reduce within our
pare against. Nevertheless, we believe the ONS re- “pandemic cohort” in 2020 – but in fact they stayed
porting of the population feeling worried about the the same as the previous pre-pandemic group.
future, as well as stressed and anxious – see Figure We wanted to explore why this might be happen-
3 (Office for National Statistics, 2021) – provides an ing.
appropriate basis for comparison. There are some reasonable explanations that im-
We shall now consider possible explanations for mediately come to mind for why confidence may

Behavioral Economics Guide 2021 160

Gerard Creaner et al. Career Coaching Builds Confidence in Experienced Workers

have remained high for the pandemic cohort: delivered in an online format, and so studies could
continue independently of lockdowns
• These experienced workers were moving into the
pharmaceutical and medical device manufactur- However, as we had already observed increased
ing industry in Ireland, which continued to hire Confidence Metrics to the same level within the
throughout the lockdowns (NIBRT, 2020) pre-pandemic cohort – and shown that increasing
• The experienced workers were all beneficiaries participation in the Job-Hunting Skills module led
of a government-funded reskilling programme to increasing Confidence Metrics – the points above
(Higher Education Authority Ireland, 2021), so did not fully account for the results we were seeing
they faced little or no financial risk in 2020.
• Both the technical programme and the Job-Hunt- Why didn’t we see the same fall in Confidence
ing Skills module were purposely developed to be Metrics that was seen in other government and na-

Figure 3: ONS – Impact of COVID-19 on the adult population in Great Britain, April 2020. Source: Office for National
Statistics – Opinions and Lifestyle Survey

Behavioral Economics Guide 2021 161

Gerard Creaner et al. Career Coaching Builds Confidence in Experienced Workers

tional studies? tional success metrics across other circumstances.

The team moved to exploring the concept of psy- To do so, testing its implementation within a
chological resilience as a possible explanation. different industry niche, a different age and expe-
Studies show that one of the factors leading to a rience group and a different geographical location
person’s ability to demonstrate resilience is being would all be useful. For our own next steps, we plan
able to make realistic plans, being able to follow on replicating this project in another country.
those plans, having confidence in their strengths While job-hunting skills are typically low priority
and abilities and demonstrating solid communica- in large-scale reskilling initiatives, if these effects
tion and problem-solving skills (Padesky & Moon- can be shown to be reliable, we believe they could
ey, 2012). form an essential part of being able to transition
Building resilience for an individual means being successfully into a new industry, alongside gaining
prepared for challenges, crises and emergencies, the appropriate technical skills.
and feeling like they have reliable contingencies in It is important to note that while the inclusion of a
place to deal with unexpected future scenarios (de Job-Hunting Skills module will not result in a 100%
Terte & Stephens, 2014). job success rate, we found that the implementation
In other words, they are building self-efficacy of a mandated industry-specific component had the
beliefs, i.e. tying their strong beliefs to what they measurable impact of getting more people into jobs
understand to be their capability level (Bandura, in their intended industry.
1977). However, building a resilient workforce has ben-
The Job-Hunting Skills module makes an action- efits beyond “equipping them with knowledge” and
able process out of finding a job. It demystifies the “getting people into work.”
steps involved and breaks down the process into The traditional measurables of successfully
bitesize pieces that build on previous knowledge. completing a programme and/or getting a job,
It is taught as a cyclical and continuous process combined with the added “bonus effects” of the in-
that each person works through repeatedly, gain- creased Confidence Metrics, could leave future re-
ing confidence with each iteration. At the end of the skilling efforts with a realistic opportunity to really
programme, the experienced worker comes away go above and beyond.
with a plan that is applicable across all job hunts, In the private sector, this would be termed “de-
both now and in the future (i.e. it’s their reliable lighting the customer.” And while it’s not neces-
contingency). sarily a primary outcome measure for government
We hypothesise that the feeling of self-efficacy programmes, it is an opportunity for an addition-
gained from the module means that experienced al “successful outcome” as a result of the public
workers believe they now have the improved abil- spending involved in large-scale workforce reskill-
ities to deal with an ambiguous future. ing.
We see their resilience and confidence in the future As a private training provider, we acknowledge
as an unanticipated “bonus effect” of the current that we’re in a uniquely agile position to test and
Irish Government-funded reskilling programme, implement such initiatives that cannot necessarily
in which we bring together both job-hunting and be easily replicated by governments or traditional
technical skills. academia. Nonetheless, there are definitely oppor-
tunities for testing, and while it might not be easy,
Implications for Future Government our findings suggest that the benefits might be
Reskilling Programmes worth pursuing.
Imagine a society where a population displayed an
We note that our findings on “bonus effects” increase in the nine factors within our Confidence
are preliminary and that more research is needed Metrics despite extreme circumstances such as a
to confirm that implementing job-hunting skills pandemic. This unanticipated benefit of “delight-
continues to offer these, as well as improving tradi- ing the customer” suddenly becomes something to

Behavioral Economics Guide 2021 162

Gerard Creaner et al. Career Coaching Builds Confidence in Experienced Workers

strive for, in the public and private sectors alike. versity and spent seven years working in the phar-
And if that is indeed the case, we ask if govern- maceutical and medical device industries.
ments could make use of these observations to
develop and implement their own programmes, in Colm Creaner is a university student who is com-
order to assess and capture the “bonus effects” – as pleting a research internship with GetReskilled. He
well as traditional successes – when implementing has previously completed research work with the
reskilling initiatives in a post-COVID world. company in this area and participated in other pub-
lications with the research team.
The Authors
References
Gerard Creaner has over 30 years of experience
in the life sciences manufacturing industry across Bandura, A. (1977). Self-efficacy: Toward a unify-
a range of technical, managerial and business roles. ing theory of behavioral change. Psychological
He established a very successful engineering con- Review, 84(2), 191-215.
sultancy prior to founding GetReskilled, an online Beckett, D. (2021, May 12). Coronavirus and the im-
education and learning business, with offices in pact on output in the UK economy - Office for
Singapore, Ireland and Boston (USA), focused on National Statistics. Office for National Statistics.
the manufacture of safe and effective medicines https://www.ons.gov.uk/economy/grossdomes-
for the public. He is also a founding director of ticproductgdp/articles/coronavirusandtheim-
two Singapore-based philanthropic organisations, pactonoutputintheukeconomy/march2021.
namely the Farmleigh Fellowship and the Sin- Office for National Statistics. (2021, May 15).
gapore-Ireland Fund, both of which deepen the Coronavirus (COVID-19) latest insights - Office
well-established and historical Singapore-Ireland for National Statistics. https://www.ons.gov.uk/
relationship and deliver long-term benefits to both peoplepopulationandcommunity/healthandso-
countries. Gerard has an undergraduate degree cialcare/conditionsanddiseases/articles/coro-
in Chemical Engineering (UCD, 1980) and an MSc naviruscovid19/latestinsights#infections.
(Management) from Trinity College Dublin (2003), Creaner, G. (2020, October 26). Working in Ireland’s
and he is currently doing research for his Ph.D. on Pharma/MedTech sector? It’s time to negotiate.
the decision-making of experienced workers. GetReskilled. https://www.getreskilled.com/
time-to-negotiate/.
Sinead Creaner is research lead for the GetRe- de Terte, I., & Stephens, C. (2014). Psychological
skilled research team. She has co-authored a series resilience of workers in high-risk occupations.
of papers over the last couple of years looking at the Stress and Health, 30(5), 353-355.
decision-making processes of experienced workers Department of the Taoiseach. (2020, September).
in the field of behavioural science. She also works COVID-19 resilience & recovery 2021: The path
with GetReskilled’s Springboard+ and BSc students. ahead. Government of Ireland. https://www.gov.
She has extensive international experience, having ie/en/campaigns/resilience-recovery-2020-
lived, studied and worked in Ireland, Singapore, 2021-plan-for-living-with-covid-19/.
Australia, Boston MA and London UK over the last GetReskilled. (2021, May 26). List of 200
12 years. Pharmaceutical & Med Device factories by county in
Ireland. https://www.getreskilled.com/pharma-
Claire Wilson runs GetReskilled’s career coach- ceutical-jobs/ireland-factory-table/.
ing strategy – including the specially designed Halligan, U. (Ed.). (2016). Future skills needs of the
Job-Hunting Skills module. She is extremely enthu- biopharma industry in Ireland. Expert Group of
siastic about helping people reach their final goal of Future Skills Needs. http://www.skillsireland.ie/
employment in their new career path. Claire has a all-publications/2016/biopharma-skills-re-
BSc (Hons) in Medical Biology from Edinburgh Uni- port-final-web-version.pdf.

Behavioral Economics Guide 2021 163

Gerard Creaner et al. Career Coaching Builds Confidence in Experienced Workers

Higher Education Authority Ireland. (2021). HEA

- Springboard+. Springboard+. https://spring-
boardcourses.ie/about.
IDA Ireland. (2021). Bio-Pharmaceuticals &
Biotechnology Ireland. https://www.idaireland.
com/doing-business-here/industry-sectors/
bio-pharmaceuticals.
Kruger, J., & Dunning, D. (1999). Unskilled and un-
aware of it: How difficulties in recognizing one’s
own incompetence lead to inflated self-assess-
ments. Journal of Personality and Social Psychology,
77(6), 1121-1134.
Moore, C. (2007). Action mapping on one page.
Training Design - Cathy Moore. https://blog.
cathy-moore.com/online-learning-confer-
ence-anti-handout/.
NIBRT. (2020). NIBRT annual report 2020 (B.
O’Callaghan, Ed.). National Institute for
Bioprocessing Research and Training. https://
www.nibrt.ie/wp-content/uploads/2021/01/
NIBRT-Annual-Report-2020_final.pdf.
Padesky, C. A., & Mooney, K. A. (2012). Strengths-
based cognitive-behavioural therapy: A four-
step model to build resilience. Clinical Psychology
& Psychotherapy, 19(4), 283-290.
Public Health England. (2021, May). COVID-19
vaccine surveillance report Week 20 (No. 20). UK
Government. https://assets.publishing.service.
gov.uk/government/uploads/system/uploads/
attachment_data/file/988193/Vaccine_surveil-
lance_report_-_week_20.pdf.
Higher Education Authority Ireland. (2016).
Developing talent, changing lives: An evaluation
of Springboard+. Higher Education Authority
Ireland. https://springboardcourses.ie/pdfs/An-
Evaluation-of-Springboard+-2011-16.pdf.
Stringer, E. (2007). Action research in education (2nd
Edition) (2nd ed.). Pearson.

Behavioral Economics Guide 2021 164

III
RESOURCES
 Behavioral Science Concepts*

A
100 patients similar to Mr. Jones, 10 are estimated to
commit an act of violence”) more than an abstract
but equivalent probability frame (e.g. “Patients
similar to Mr. Jones are estimated to have a 10%
chance of committing an act of violence to others”)
Action bias (Slovic et al., 2000).
Affect-based judgments are more pronounced
Some core ideas in behavioral economics focus when people do not have the resources or time to
on people’s propensity to do nothing, as evident in reflect. For example, instead of considering risks
default bias and status quo bias. Inaction may be and benefits independently, individuals with a neg-
due to a number of factors, including inertia or an- ative attitude towards nuclear power may consider
ticipated regret. However, sometimes people have its benefits as low and risks as high under condi-
an impulse to act in order to gain a sense of control tions of time pressure. This leads to a more nega-
over a situation and eliminate a problem. This has tive risk-benefit correlation than would be evident
been termed the action bias (Patt & Zeckhauser, without time pressure (Finucane et al., 2000).
2000). For example, a person may opt for a medical The affect heuristic has been used as a possible
treatment rather than a no-treatment alternative, explanation for a range of consumer judgments, in-
even though clinical trials have not supported the cluding product innovations (King & Slovic, 2014),
treatment’s effectiveness. brand image (e.g. Ravaja et al., 2015), and product
Action bias is particularly likely to occur if we do pricing (e.g. the zero price effect; see Samson &
something for others or others expect us to act (see Voyer, 2012). It is considered another general pur-
social norm), as illustrated by the tendency for soc- pose heuristic similar to availability heuristic and
cer goal keepers to jump to left or right on penalty representativeness heuristic in the sense that affect
kicks, even though statistically they would be better serves as an orienting mechanism akin to similarity
off if they just stayed in the middle of the goal (Bar- and memorability (Kahneman & Frederick, 2002).
Eli et al., 2007). Action bias may also be more likely
among overconfident individuals or if a person has Altruism
experienced prior negative outcomes (Zeelenberg
et al., 2002), where subsequent inaction would be a According to neoclassical economics, rational
failure to do something to improve the situation. beings do whatever they need to in order to max-
imize their own wealth. However, when people
Affect heuristic make sacrifices to benefit others without expecting
a personal reward, they are thought to behave al-
The affect heuristic represents a reliance on good truistically (Rushton, 1984). Common applications
or bad feelings experienced in relation to a stimu- of this pro-social behavior include volunteering,
lus. Affect-based evaluations are quick, automatic, philanthropy, and helping others in emergencies
and rooted in experiential thought that is activated (Piliavin & Charng, 1990).
prior to reflective judgments (see dual-system the- Altruism is evident in a number of research find-
ory) (Slovic et al., 2002). For example, experiential ings, such as dictator games. In this game, one
judgments are evident when people are influenced participant proposes how to split a reward between
by risks framed in terms of counts (e.g. “of every himself and another random participant. While
some proposers (dictators) keep the entire reward
* Acknowledgements: The editor would like to thank for themselves, many will also voluntarily share
Andreas Haberl, Chelsea Hulse, and Roger Miles for their some portion of the reward (Fehr & Schmidt, 1999).
contributions to this encyclopedia. While altruism focuses on sacrifices made to ben-

Behavioral Economics Guide 2021 166

 Behavioral Science Concepts

efit others, similar concepts explore making sacri- the last three digits of their phone number multi-
fices to ensure fairness (see inequity aversion and plied by one thousand (e.g. 678 = 678,000). Results
social preferences). showed that people’s subsequent estimate of house
prices were significantly influenced by the arbitrary
Ambiguity (uncertainty) aversion anchor, even though they were given a 10 minute
presentation on facts and figures from the housing
Ambiguity aversion, or uncertainty aversion, is market at the beginning of the study. In practice,
the tendency to favor the known over the unknown, anchoring effects are often less arbitrary, as evident
including known risks over unknown risks. For the price of the first house shown to us by a real
example, when choosing between two bets, we are estate agent may serve as an anchor and influence
more likely to choose the bet for which we know perceptions of houses subsequently presented to us
the odds, even if the odds are poor, than the one for (as relatively cheap or expensive). Anchoring effects
which we don’t know the odds. have also been shown in the consumer packaged
This aversion has gained attention through the goods category, whereby not only explicit slogans
Ellsberg Paradox (Ellsberg, 1961). Suppose there are to buy more (e.g. “Buy 18 Snickers bars for your
two bags each with a mixture of 100 red and black freezer”), but also purchase quantity limits (e.g.
balls. A decision-maker is asked to draw a ball from “limit of 12 per person”) or ‘expansion anchors’
one of two bags with the chance to win $100 if red (e.g. “101 uses!”) can increase purchase quantities
is drawn. In one bag, the decision-maker knows (Wansink et al., 1998).
that exactly half of the pieces are red and half are
black. The color mixture of pieces in the second Asymmetrically dominated choice
bag is unknown. Due to ambiguity aversion, de-
cision-makers would favor drawing from the bag See Decoy effect
with the known mixture than the one with the un-
known mixture (Ellsberg, 1961). This occurs despite Availability heuristic
the fact that people would, on average, bet on red
or black equally if they were presented with just one Availability is a heuristic whereby people make
bag containing either the known 50-50 mixture or judgments about the likelihood of an event based
a bag with the unknown mixture. on how easily an example, instance, or case comes
Ambiguity aversion has also been documented to mind. For example, investors may judge the
in real-life situations. For example, it leads people quality of an investment based on information that
to avoid participating in the stock market, which was recently in the news, ignoring other relevant
has unknown risks (Easley & O’Hara, 2009), and to facts (Tversky & Kahneman, 1974). In the domain
avoid certain medical treatments when the risks are of health, it has been shown that drug advertising
less known (Berger, et al., 2013). recall affects the perceived prevalence of illnesses
(An, 2008), while physicians’ recent experience of
Anchoring (heuristic) a condition increases the likelihood of subsequently
diagnosing the condition (Poses & Anthony, 1991).
Anchoring is a particular form of priming effect In consumer research, availability can play a role in
whereby initial exposure to a number serves as a various estimates, such as store prices (Ofir et al.,
reference point and influences subsequent judg- 2008) or product failure (Folkes, 1988). The avail-
ments. The process usually occurs without our ability of information in memory also underlies the
awareness (Tversky & Kahneman, 1974) and has representativeness heuristic.
been researched in many contexts, including prob-
ability estimates, legal judgments, forecasting and
purchasing decisions (Furnham & Boo, 2011).
One experiment asked participants to write down

Behavioral Economics Guide 2021 167

 Behavioral Science Concepts

B
confidence, anchoring, and representativeness,
which lead investors to interpret increasing prices
as a trend that will continue, causing them to chase
the market (Fisher, 2014). Economic bubbles are
usually followed a sudden and sharp decrease in
Behavioral economics prices, also known as a crash.

The field of behavioral economics studies and

C
describes economic decision-making. According to
its theories, actual human behavior is less ration-
al, stable, and selfish than traditional normative
theory suggests (see also homo economicus), due to
bounded rationality, limited self-control, and so-
cial preferences. Certainty/possibility effects

Bias Changes in the probability of gains or losses do

not affect people’s subjective evaluations in linear
See Cognitive bias terms (see also prospect theory and “Zero price
effect”) (Tversky & Kahneman, 1981). For example,
Bounded rationality a move from a 50% to a 60% chance of winning a
prize has a smaller emotional impact than a move
Bounded rationality is a concept proposed by from a 95% chance to a 100% chance (certainty).
Herbert Simon that challenges the notion of human Conversely, the move from a 0% chance to a 5%
rationality as implied by the concept of homo eco- possibility of winning a prize is more attractive
nomicus. Rationality is bounded because there are than a change from 5% to 10%. People over-weight
limits to our thinking capacity, available informa- small probabilities, which explains the attractive-
tion, and time (Simon, 1982). Bounded rationality ness of gambling. Research suggests that problem
is a core assumption of the “natural assessments” gamblers’ probability perception of losing is not
view of heuristics and dual-system models of distorted and that their loss aversion is not signif-
thinking (Gilovich et al., 2002), and it is one of the icantly different from other people. However, they
psychological foundations of behavioral economics. are much more risk-taking and strongly overweight
(See also satisficing and fast and frugal.) small to medium probabilities of winning (Ring et
al., 2018).
(Economic) Bubble
Choice architecture
Economic (or asset) bubbles form when prices are
driven much higher than their intrinsic value (see This term coined by Thaler and Sunstein (2008)
also efficient market hypothesis). Well-known refers to the practice of influencing choice by “or-
examples of bubbles include the US Dot-com stock ganizing the context in which people make deci-
market bubble of the late 1990s and housing bub- sions” (Thaler et al., 2013, p. 428; see also nudge).
ble of the mid-2000s. According to Robert Shiller A frequently mentioned example is how food is
(2015), who warned of both of these events, specu- displayed in cafeterias, where offering healthy
lative bubbles are fueled by contagious investor en- food at the beginning of the line or at eye level can
thusiasm (see also herd behavior) and stories that contribute to healthier choices. Choice architecture
justify price increases. Doubts about the real value includes many other behavioral tools that affect de-
of investment are overpowered by strong emotions, cisions, such as defaults, framing, or decoy options.
such as envy and excitement.
Other biases that promote bubbles include over-

Behavioral Economics Guide 2021 168

 Behavioral Science Concepts

Choice overload chunking is, in fact, most effective when four to six
chunks are created (Mathy & Feldman, 2012). Al-
Also referred to as ‘overchoice’, the phenomenon though this seems to be a ‘magic number’, it is also
of choice overload occurs as a result of too many possible to learn to increase the size of those chunks
choices being available to consumers. Overchoice over time (Sullivan, 2009).
has been associated with unhappiness (Schwartz, In behavioral science, chunking has also been used
2004), decision fatigue, going with the default op- to refer to breaking up processes or tasks into more
tion, as well as choice deferral—avoiding making a manageable pieces (see for example Eşanu, 2019, on
decision altogether, such as not buying a product chunking in UX design or Wijland & Hansen, 2016,
(Iyengar & Lepper, 2000). Many different factors on mobile nudging in the banking sector).
may contribute to perceived choice overload, in-
cluding the number of options and attributes, time Cognitive bias
constraints, decision accountability, alignability
and complementarity of options, consumers’ pref- A cognitive bias (e.g. Ariely, 2008) is a systematic
erence uncertainty, among other factors (Chernev (non-random) error in thinking, in the sense that a
et al., 2015). judgment deviates from what would be considered
Choice overload can be counteracted by simplify- desirable from the perspective of accepted norms or
ing choice attributes or the number of available op- correct in terms of formal logic. The application of
tions (Johnson et al., 2012). However, some studies heuristics is often associated with cognitive biases.
on consumer products suggest that, paradoxically, Some biases, such as those arising from availability
greater choice should be offered in product domains or representativeness, are ‘cold’ in the sense that
in which people tend to feel ignorant (e.g. wine), they do not reflect a person’s motivation and are
whereas less choice should be provided in domains instead the result of errors in information process-
in which people tend to feel knowledgeable (e.g. soft ing. Other cognitive biases, especially those that
drinks) (Hadar & Sood, 2014). have a self-serving function (e.g. overconfidence),
are more motivated. Finally, there are also biases
Chunking that can be motivated or unmotivated, such as con-
firmation bias (Nickerson, 1998).
When the same information is presented in a As the study of heuristics and biases is a core el-
different form that is easier to process, our ability ement of behavioral economics, the psychologist
to receive and remember it is greater. People often Gerd Gigerenzer has cautioned against the trap of a
reorganize, regroup or compress information to aid “bias bias” – the tendency to see biases even when
in its understanding or recall. The resulting sub- there are none (Gigerenzer, 2018).
groups are ‘chunks’, which can be defined as a set
of information or items that are treated collectively Cognitive dissonance
as a single unit (Mathy & Feldman, 2012). Chunk-
ing may be done through strategic reorganization Cognitive dissonance, an important concept in
based on familiarity, prior knowledge, proximity or social psychology (Festinger, 1957), refers to the
other means to structure the information at hand. uncomfortable tension that can exist between two
For example, a phone number may be split up into simultaneous and conflicting ideas or feelings—of-
three subgroups of area code, prefix and number or ten as a person realizes that s/he has engaged in a
one might recognize a meaningful date in it, and so behavior inconsistent with the type of person s/he
can organize it more easily into different chunks. would like to be, or be seen publicly to be. According
In relation to the ideal amount of chunks, Miller to the theory, people are motivated to reduce this
(1956) found that humans best recall seven plus tension by changing their attitudes, beliefs, or ac-
or minus two units when processing information. tions. For example, smokers may rationalize their
More recently, various studies have shown that behavior by holding ‘self-exempting beliefs’, such

Behavioral Economics Guide 2021 169

 Behavioral Science Concepts

as “The medical evidence that smoking causes can- a particular brand and researches a new purchase
cer is not convincing” or “Many people who smoke may be motivated to seek out customer reviews on
all their lives live to a ripe old age, so smoking is not the internet that favor that brand. Confirmation
all that bad for you” (Chapman et al., 1993). bias has also been related to unmotivated processes,
Arousing dissonance can be used to achieve be- including primacy effects and anchoring, evident in
havioral change; one study (Dickerson et al., 1992), a reliance on information that is encountered early
for instance, made people mindful of their waste- in a process (Nickerson, 1998).
ful water consumption and then made them urge
others (publicly commit) to take shorter showers. Control premium
Subjects in this ‘hypocrisy condition’ subsequently
took significantly shorter showers than those who In behavioral economics, the control premium
were only reminded that they had wasted water or refers to people’s willingness to forego potential re-
merely made the public commitment. wards in order to control (avoid delegation) of their
own payoffs. In an experiment, participants were
Commitment asked to choose whether to bet on another person
or themselves answering a quiz question correctly.
Commitments (see also precommitment) are Although individuals’ maximizing their rewards
often used as a tool to counteract people’s lack of would bet on themselves in 56% of the decisions
willpower and to achieve behavior change, such as (based on their beliefs), they actually bet on them-
in the areas of dieting or saving. The greater the cost selves 65% of the time, suggesting an aggregate
of breaking a commitment, the more effective it is control premium of almost 10%. The average study
(Dolan et al., 2010). From the perspective of social participant was willing to sacrifice between 8 and
psychology, individuals are motivated to main- 15% of expected earnings to retain control (Owens
tain a consistent and positive self-image (Cialdini, et al., 2014). (See also overconfidence.)
2008), and they are likely to keep commitments to
avoid reputational damage (if done publicly) and/or Curse of knowledge
cognitive dissonance (Festinger, 1957). A field ex-
periment in a hotel, for example, found 25% greater Economists commonly assume that having more
towel reuse among guests who made a commitment information allows us to make better decisions.
to reuse towels at check-in and wore a “Friend of However, the information asymmetry that exists
the Earth” lapel pin to signal their commitment when one economic agent has more information
during their stay (Baca-Motes et al., 2012). The be- than another can also have negative effects for the
havior change technique of ‘goal setting’ is related better-informed agent. This is known as the curse
to making commitments (Strecher et al., 1995), of knowledge (Camerer et al., 1989), which occurs
while reciprocity involves an implicit commitment. because better-informed agents are unable to ig-
nore their own knowledge.
Confirmation bias The curse of knowledge can manifest itself in
many domains of economic life, such as setting
Confirmation bias (Wason, 1960) occurs when prices or estimating productivity. With respect to
people seek out or evaluate information in a way the latter, one study found that experts consistently
that fits with their existing thinking and preconcep- underestimate the amount of time required by nov-
tions. The domain of science, where theories should ices to perform a task (Hinds, 1999).
advance based on both falsifying and supporting A fun way to show the curse of knowledge in action
evidence, has not been immune to bias, which is is through a musical game in which participants are
often associated with people processing hypothe- either the “tapper” or a “listener.” In the game,
ses in ways that end up confirming them (Oswald the tapper selects a simple, well-known song, such
& Grosjean, 2004). Similarly, a consumer who likes a “Happy Birthday,” and taps out the rhythm on a

Behavioral Economics Guide 2021 170

 Behavioral Science Concepts

table. The listeners then try to guess the song. In preference for one option over another changes as
an early experiment, tappers expected the listeners a result of adding a third (similar but less attrac-
to correctly guess the song 50% of the time, yet, in tive) option. For example, people are more likely
reality, listeners were only correct 2.5% of the time to choose an elegant pen over $6 in cash if there
(Newton, 1990). is a third option in the form of a less elegant pen
(Bateman et al., 2008). While this effect has been
extensively studied in relation to consumer prod-

D
ucts, it has also been found in employee selection
(e.g. Slaughter et al., 2006), apartment choices
(Simonson, 1989), or as a nudge to increase cancer
screening (Stoffel et al., 2019).

Decision fatigue Default (option)

There are psychological costs to making deci- Default options are pre-set courses of action that
sions. Since choosing can be difficult and requires take effect if nothing is specified by the decision
effort, just like any other activity, long sessions of maker (Thaler & Sunstein, 2008), and setting de-
decision making can lead to poor choices. Similar faults is an effective nudge when there is inertia
to other activities that consume resources required or uncertainty in decision making (Samson, 2014).
for executive functions, decision fatigue is reflect- Since defaults do not require any effort by the de-
ed in self-regulation, such as a diminished ability cision maker, defaults can be a simple but powerful
to exercise self-control (Vohs et al., 2008). (See tool when there is inaction (Samson & Ramani,
also choice overload and ego depletion.) 2018). When choices are difficult, defaults may also
be perceived as a recommended course of action
Decision staging (McKenzie et al., 2006). Requiring people to opt
out if they do not wish to donate their organs, for
When people make complex or long decisions, example, has been associated with higher donation
such as buying a car, they tend to explore their rates (Johnson & Goldstein, 2003). Similarly, mak-
options successively. This involves deciding what ing contributions to retirement savings accounts
information to focus on, as well as choices be- has become automatic in some countries, such as
tween attributes and alternatives. For example, the United Kingdom and the United States.
when people narrow down their options, they often
tend to screen alternatives on the basis of a sub- Delusion of competence (Dunning-Kruger
set of attributes, and then they compare alterna- effect)
tives. Choice architects may not only break down
complex decisions into multiple stages, to make This is the case whereby, either socially or patho-
the process easier, but they can also work with an logically, a person lacks reflexive acknowledge-
understanding of sequential decision making by ment that they are not equipped to make a decision
facilitating certain comparisons at different stages or to act appropriately in relation to the demands
of the choice process (Johnson et al., 2012). of a situation. Kruger and Dunning (1999) ob-
served a divergence between perceived and actual
Decoy effect competence which explains a range of unsound
decision-making. The effect explains why, among
Choices often occur relative to what is on offer other real-world difficulties, management boards
rather than based on absolute preferences. The decide to promote products whose working they
decoy effect is technically known as an ‘asymmet- don’t understand, and why talent show contestants
rically dominated choice’ and occurs when people’s are unaware of their inability to sing, until ejected

Behavioral Economics Guide 2021 171

 Behavioral Science Concepts

by the judges. (The prevalence of this bias has made MP3 player, but on the actual trip I may mostly end
the producers of certain talent shows very wealthy.) up listening to my favorite rock music. When peo-
ple make simultaneous choices among things that
Dictator game can be classified as virtues (e.g. high-brow movies
or healthy deserts) or vices (e.g. low-brow movies
The dictator game is an experimental game (see or hedonic deserts), their diversification strategy
behavioral game theory) designed to elicit altruistic usually involves a greater selection of virtues (Read
aspects of behavior. In the ultimatum game, a pro- et al., 1999). (See also projection bias.)
posing player is endowed with a sum of money and
asked to split it with another (responding) player. Dual-self model
The responder may either accept the proposer’s of-
fer or reject it, in which case neither of the players In economics, dual-self models deal with the
will receive anything. Since expressed preferences inconsistency between the patient long-run self
in the ultimatum game may be due to factors other and myopic short-run self. With respect to savings
than altruism (e.g. fear of envy), the dictator game behavior, Thaler and Shefrin (1981) introduced the
is played without the responder being able to decide concepts of the farsighted planner and myopic doer.
whether to accept the offer or not (Camerer, 2003). At any point in time, there is a conflict between those
As a result, it only involves one actual player and selves with two sets of preferences. The approach
is not strictly a game. Whether or not these games helps economic theorists overcome the paradox
really better measure altruism, or something else, created by self-control in standard views of utility.
forms part of an interesting debate (e.g. Bardsley, The more recent dual-self model of impulse control
2008) (See also trust game.) (Fudenberg & Levine, 2006) explains findings from
the areas of time discounting, risk aversion, and
Discounting self-control (see also intertemporal choice). More
practically-oriented research on savings behavior
See Time discounting has attempted to make people feel more connected
to their future selves, making them appreciate that
Disposition effect they are the future recipients of current savings.
In an experiment, participants who were exposed
The disposition effect refers to investors’ reluc- to their future (as opposed to present) self in the
tance to sell assets that have lost value and greater form of an age-progressed avatar in virtual reality
likelihood of selling assets that have made gains environments allocated twice as much money to a
(Shefrin & Statman, 1985). This phenomenon can retirement account (Hershfield et al., 2011).
be explained by prospect theory (loss aversion),
regret avoidance and mental accounting. Dual-system theory

Diversification bias Dual-system models of the human mind contrast

automatic, fast, and non-conscious (System 1) with
People seek more variety when they choose controlled, slow, and conscious (System 2) thinking
multiple items for future consumption simultane- (see Strack & Deutsch, 2015, for an extensive re-
ously than when they make choices sequentially, view). Many heuristics and cognitive biases studied
i.e. on an ‘in the moment’ basis. Diversification is by behavioral economists are the result of intui-
non-optimal when people overestimate their need tions, impressions, or automatic thoughts generat-
for diversity (Read & Loewenstein, 1995). In other ed by System 1 (Kahneman, 2011). Factors that make
words, sequential choices lead to greater experi- System 1’s processes more dominant in decision
enced utility. For example, before going on vacation making include cognitive busyness, distraction,
I may upload classical, rock and pop music to my time pressure, and positive mood, while System

Behavioral Economics Guide 2021 172

 Behavioral Science Concepts

2’s processes tend to be enhanced when the deci- their disposal when they make choices. One of
sion involves an important object, has heightened these effort-reducing heuristics is referred to as
personal relevance, and when the decision maker is ‘elimination-by-aspects’. When it is applied, deci-
held accountable by others (Samson & Voyer, 2012; sion makers gradually reduce the number of alter-
Samson & Voyer, 2014). natives in a choice set, starting with the aspect that
they see as most significant. One cue is evaluated
at a time until fewer and fewer alternatives remain

E
in the set of available options (Tversky, 1972). For
example, a traveler may first compare a selection of
hotels at a target destination on the basis of clas-
sification, eliminating all hotels with fewer than
three stars. The person may then reduce the choice
Efficient market hypothesis set further by walking distance from the beach, fol-
lowed by guest reviews, etc., until only one option
According to the efficient market hypothesis, the remains.
price (market value) of a security reflects its true
worth (intrinsic value). In a market with perfectly (Hot-cold) Empathy gap
rational agents, “prices are right”. Findings in be-
havioral finance, by contrast, suggests that asset It is difficult for humans to predict how they will
prices also reflect the trading behavior of individ- behave in the future. A hot-cold empathy gap oc-
uals who are not fully rational (Barberis & Thaler, curs when people underestimate the influence of
2003), leading to anomalies such as asset bubbles. visceral states (e.g. being angry, in pain, or hungry)
on their behavior or preferences (Loewenstein,
Ego depletion 2005). In medical decision making, for example, a
hot-to-cold empathy gap may lead to undesirable
Ego depletion is a concept emanating from treatment choices when cancer patients are asked
self-regulation (or self-control) theory in psy- to choose between treatment options right after
chology. According to the theory, willpower oper- being told about their diagnosis.
ates like a muscle that can be exercised or exerted. In a study on the reverse, a cold-to-hot empa-
Studies have found that tasks requiring self-con- thy gap, smokers were assigned to different ex-
trol can weaken this muscle, leading to ego de- perimental conditions (Sayette et al., 2008). Some
pletion and a subsequently diminished ability to smokers in a hot (craving) state were asked to make
exercise self-control. In the lab, ego depletion has predictions about a high-craving state in a second
been induced in many different ways, such as hav- session. Others made the same prediction while
ing to suppress emotions or thoughts, or having to they were in a cold state. In contrast to those in the
make a range of difficult decisions. The resulting hot group, smokers in the cold group underpredict-
ego depletion leads people to make less restrained ed how much they would value smoking during the
decisions; consumers, for example, may be more second session. This empathy gap can explain poor
likely to choose candy over ‘healthy’ granola bars decisions among smokers attempting to quit that
(Baumeister et al., 2008). Some studies now sug- place them in high-risk situations (e.g. socializing
gest that the evidence for this resource depletion over a drink) and why people underestimate their
model of self-control has been overestimated (e.g. risk of becoming addicted in the first place.
Hagger & Chatzisarantis, 2016).
Endowment effect
Elimination-by-aspects
This bias occurs when we overvalue a good that
Decision makers have a variety of heuristics at we own, regardless of its objective market value

Behavioral Economics Guide 2021 173

 Behavioral Science Concepts

(Kahneman et al., 1991). It is evident when people unfairness, such as excessive CEO compensation,
become relatively reluctant to part with a good has been behaviorally associated with reduced work
they own for its cash equivalent, or if the amount morale among employees (Cornelissen et al., 2011).
that people are willing to pay for the good is lower
than what they are willing to accept when selling Fast and frugal
the good. Put more simply, people place a greater
value on things once they have established owner- Fast and frugal decision-making refers to the ap-
ship. This is especially true for goods that wouldn’t plication of ecologically rational heuristics, such as
normally be bought or sold on the market, usually the recognition heuristic, which are rooted in the
items with symbolic, experiential, or emotional psychological capacities that we have evolved as
significance. Endowment effect research has been human animals (e.g. memory and perceptual sys-
conducted with goods ranging from coffee mugs tems). They are ‘fast and frugal’ because they are
(Kahneman et al., 1990) to sports cards (List, 2011). effective under conditions of bounded rationality—
While researchers have proposed different reasons when knowledge, time, and computational power
for the effect, it may be best explained by psycho- are limited (Goldstein & Gigerenzer, 2002).
logical factors related to loss aversion (Ericson &
Fuster, 2014). Fear of missing out

Extrapolation bias Social media has enabled us to connect and inter-

act with others, but the number of options offered to
See Representativeness heuristic us through these channels is far greater than what
we can realistically take up, due to limited time and
practical constraints. The popular concept of FoMO,

F
or Fear of Missing Out, refers to “a pervasive ap-
prehension that others might be having rewarding
experiences from which one is absent” (Przybyl-
ski et al., 2013). People suffering from FoMO have
a strong desire to stay continually informed about
Fairness what others are doing (see also scarcity heuristic,
regret aversion, and loss aversion).
In behavioral science, fairness refers to our social
preference for equitable outcomes. This can pres- Framing effect
ent itself as inequity aversion, people’s tendency
to dislike unequal payoffs in their own or someone Choices can be presented in a way that high-
else’s favor. This tendency has been documented lights the positive or negative aspects of the same
through experimental games, such as the ultima- decision, leading to changes in their relative at-
tum, dictator, and trust games (Fehr & Schmidt, tractiveness. This technique was part of Tversky
1999). and Kahneman’s development of prospect theory,
A large part of fairness research in economics has which framed gambles in terms of losses or gains
focused on prices and wages. With respect to prices, (Kahneman & Tversky, 1979a). Different types of
for example, consumers are generally less accepting framing approaches have been identified, including
of price increases as result of a short term growth in risky choice framing (e.g. the risk of losing 10 out of
demand than rise in costs (Kahneman et al., 1986). 100 lives vs. the opportunity to save 90 out of 100
With respect to wages, employers often agree to lives), attribute framing (e.g. beef that is described
pay more than the minimum the employees would as 95% lean vs. 5% fat), and goal framing (e.g. mo-
accept in the hope that this fairness will be recip- tivating people by offering a $5 reward vs. imposing
rocated (e.g. Jolls, 2002). On the flip side, perceived a $5 penalty) (Levin et al., 1998).

Behavioral Economics Guide 2021 174

 Behavioral Science Concepts

H
The concept of framing also has a long history
in political communication, where it refers to the
informational emphasis a communicator chooses
to place in a particular message. In this domain,
research has considered how framing affects public
opinions of political candidates, policies, or broad- Habit
er issues (Busby et al., 2018).
Habit is an automatic and rigid pattern of behav-
ior in specific situations, which is usually acquired

G
through repetition and develops through associa-
tive learning (see also System 1 in dual-system the-
ory), when actions become paired repeatedly with
a context or an event (Dolan et al., 2010). ‘Habit
loops’ involve a cue that triggers an action, the ac-
Gambler’s fallacy tual behavior, and a reward. For example, habitual
drinkers may come home after work (the cue), drink
The term ‘gambler’s fallacy’ refers to the mis- a beer (the behavior), and feel relaxed (the reward)
taken belief held by some people that independent (Duhigg, 2012). Behaviors may initially serve to
events are interrelated; for example, a roulette or attain a particular goal, but once the action is au-
lottery player may choose not to bet on a number tomatic and habitual, the goal loses its importance.
that came up in the previous round. Even though For example, popcorn may habitually be eaten in
people are usually aware that successive draws of the cinema despite the fact that it is stale (Wood &
numbers are unrelated, their gut feeling may tell Neal, 2009). Habits can also be associated with sta-
them otherwise (Rogers, 1998). tus quo bias.

(Behavioral) Game theory Halo effect

Game theory is a mathematical approach to mod- This concept has been developed in social psy-
eling behavior by analyzing the strategic decisions chology and refers to the finding that a global
made by interacting players (Nash, 1950). In stand- evaluation of a person sometimes influences peo-
ard experimental economics, the theory assumes ple’s perception of that person’s other unrelated
homo economicus – a self-interested, rational max- attributes. For example, a friendly person may be
imizer. Behavioral game theory extends standard considered to have a nice physical appearance,
(analytical) game theory by taking into account whereas a cold person may be evaluated as less ap-
how players feel about the payoffs other players re- pealing (Nisbett & Wilson, 1977). Halo effects have
ceive, limits in strategic thinking, the influence of also been applied in other domains of psychology.
context, as well as the effects of learning (Camerer, For example, a study on the ‘health halo’ found that
2003). Games are usually about cooperation or fair- consumers tend to choose drinks, side dishes and
ness. Well-known examples include the ultimatum desserts with higher calorific content at fast‐food
game, dictator game and trust game. restaurants that claim to be healthy (e.g. Subway)
compared to others (e.g. McDonald’s) (Chandon &
Wansink, 2007).

Hedonic adaptation

People get used to changes in life experiences, a

process which is referred to as ‘hedonic adaptation’

Behavioral Economics Guide 2021 175

 Behavioral Science Concepts

or the ‘hedonic treadmill’. Just as the happiness as affect, “Availability heuristic”and representa-
that comes with the ownership of a new gadget or tiveness have a general purpose character; others
salary raise will wane over time, even the negative developed in social and consumer psychology are
effect of life events such as bereavement or disabili- more domain-specific, examples of which include
ty on subjective wellbeing tends to level off, to some brand name, price, and scarcity heuristics (Shah &
extent (Frederick & Loewenstein, 1999). When this Oppenheimer, 2008).
happens, people return to a relatively stable base-
line of happiness. It has been suggested that the Hindsight bias
repetition of smaller positive experiences (‘hedonic
boosts’), such as exercise or religious practices, has This bias, also referred to as the ‘knew-it-all-
a more lasting effect on our wellbeing than major along effect’, is a frequently encountered judgment
life events (Mochon et al., 2008). bias that is partly rooted in availability and repre-
sentativeness heuristics. It happens when being
Herd behavior given new information changes our recollection
from an original thought to something different
This effect is evident when people do what others (Mazzoni & Vannucci, 2007). This bias can lead to
are doing instead of using their own information or distorted judgments about the probability of an
making independent decisions. The idea of herding event’s occurrence, because the outcome of an
has a long history in philosophy and crowd psy- event is perceived as if it had been predictable. It
chology. It is particularly relevant in the domain of may also lead to distorted memory for judgments of
finance, where it has been discussed in relation to factual knowledge. Hindsight bias can be a problem
the collective irrationality of investors, including in legal decision-making. In medical malpractice
stock market bubbles (Banerjee, 1992). In other suits, for example, jurors’ hindsight bias tends to
areas of decision-making, such as politics, science, increase with the severity of the outcome (e.g. inju-
and popular culture, herd behavior is sometimes ry or death) (Harley, 2007).
referred to as ‘information cascades’ (Bikhchandi
et al., 1992). Herding behavior can be increased by Homo economicus
various factors, such as fear (e.g. Economou et al.,
2018), uncertainty (e.g. Lin, 2018), or a shared iden- The term homo economicus, or ‘economic man’,
tity of decision makers (e.g. Berger et al., 2018). denotes a view of humans in the social sciences,
particularly economics, as self-interested agents
Heuristic who seek optimal, utility-maximizing outcomes.
Behavioral economists and most psychologists,
Heuristics are commonly defined as cognitive sociologists, and anthropologists are critical of the
shortcuts or rules of thumb that simplify decisions, concept. People are not always self-interested (see
especially under conditions of uncertainty. They social preferences), nor are they mainly concerned
represent a process of substituting a difficult ques- about maximizing benefits and minimizing costs.
tion with an easier one (Kahneman, 2003). Heu- We often make decisions under uncertainty with in-
ristics can also lead to cognitive biases. There are sufficient knowledge, feedback, and processing ca-
disagreements regarding heuristics with respect pability (bounded rationality); we sometimes lack
to bias and rationality. In the fast and frugal view, self-control; and our preferences change, often in
the application of heuristics (e.g. the recognition response to changes in decision contexts.
heuristic) is an “ecologically rational” strategy that
makes best use of the limited information available Honesty
to individuals (Goldstein & Gigerenzer, 2002).
There are generally different classes of heuristics, Honesty is an important part of our everyday life.
depending on their scope. Some heuristics, such In both business and our private lives, relationships

Behavioral Economics Guide 2021 176

 Behavioral Science Concepts

I
are made and broken based on our trust in the other
party’s honesty and reciprocity.
A 2016 study investigated honesty, beliefs about
honesty and economic growth in 15 countries and
revealed large cross-national differences. Results
showed that average honesty was positively asso- Identity economics
ciated with GDP per capita, suggesting a relation-
ship between honesty and economic development. Identity economics describes the idea that we
However, expectations about countries’ levels of make economic choices based on monetary in-
honesty were not correlated with reality (the ac- centives and our identity. A person’s sense of self
tual honesty in reporting the results of a coin flip or identity affects economic outcomes. This was
experiment), but rather driven by cognitive biases outlined in Akerlof and Kranton’s (2000) seminal
(Hugh-Jones, 2016). paper which expanded the standard utility function
People typically value honesty, tend to have strong to include pecuniary payoffs and identity econom-
beliefs in their morality and want to maintain this ics in a simple game-theoretic model of behavior,
aspect of their self-concept (Mazar et al., 2008). further integrating psychology and sociology into
Self-interest may conflict with people’s honesty economic thinking.
as an internalized social norm, but the resulting When economic (or other extrinsic) incentives
cognitive dissonance can be overcome by engaging are ineffective in organizations, identity may be
in self-deception, creating moral “wiggle room” the answer: A worker’s self-image as jobholder and
that enables people to act in a self-serving man- her ideal as to how his job should be done, can be a
ner. When moral reminders are used, however, this major incentive in itself (Akerlof & Kranton, 2005).
self-deception can be reduced, as demonstrated in Organizational identification was found to be di-
laboratory experiments conducted by Mazar and rectly related to employee performance and even
colleagues (2008). It is not surprising, then, that a indirectly related with customer evaluations and
lack of social norms is a general driver of dishonest store performance in a study on 306 retail stores,
behavior, along with high benefits and low costs of for example (Lichtenstein et al., 2010). Also, when
external deception, a lack of self-awareness, as well employees were encouraged to create their own
as self-deception (Mazar & Ariely, 2006). job titles such that they better reflected the unique
Honesty must also be understood in the context value they bring to the job, identification increased,
of group membership. Employees of a large inter- and emotional exhaustion was reduced (Grant et
national bank, for example, behaved honestly on al., 2014). In some cases, identity can also have
average in an experiment’s control condition, but negative implications. Bankers whose professional
when their professional identity as bankers was identity was made salient, for example, displayed
rendered salient, a significant proportion of them more dishonest behavior (see honesty).
became dishonest. This suggests that the prevailing
business culture in the banking industry weakens IKEA effect
and undermines the honesty norm (Cohn et al.,
2014) (see also identity economics). While the endowment effect suggests that mere
ownership of a product increases its value to indi-
Hot and cold states viduals, the IKEA effect is evident when invested
labor leads to inflated product valuation (Norton et
See Empathy gap al., 2012). For example, experiments show that the
monetary value assigned to the amateur creations
Hyperbolic discounting of self-made goods is on a par with the value as-
signed to expert creations. Both experienced and
See Time discounting novice do-it-yourselfers are susceptible to the IKEA

Behavioral Economics Guide 2021 177

 Behavioral Science Concepts

effect. Research also demonstrates that the effect is deeds, as people may be perceived to have performed
not simply due to the amount of time spent on the the task for the incentives rather than for them-
creations, as dismantling a previously built product selves (Bénabou & Tirole, 2006). Similarly, perfor-
will make the effect disappear. mance incentives offered by an informed principal
The IKEA effect is particularly relevant today, (manager, teacher or parent) can adversely impact
given the shift from mass production to increas- an agent’s (worker, student or child) perception of
ing customization and co-production of value. The a task or of his own abilities, serving as only weak
effect has a range of possible explanations, such as reinforcers in the short run and negative reinforc-
positive feelings (including feelings of competence) ers in the long run (Bénabou & Tirole, 2003). (For an
that come with the successful completion of a task, interesting summary of when extrinsic incentives
a focus on the product’s positive attributes, and work and when they don’t in nonemployment con-
the relationship between effort and liking (Norton texts, see Gneezy et al., 2011).
et al., 2012), a link between our creations and our
self-concept (Marsh et al., 2018), as well as a psy- Inequity aversion
chological sense of ownership (Sarstedt et al., 2017.
The effort heuristic is another concept that pro- Human resistance to “unfair” outcomes is known
poses a link between perceived effort and valuation as ‘inequity aversion’, which occurs when peo-
(Kruger et al., 2004). ple prefer fairness and resist inequalities (Fehr &
Schmidt, 1999). In some instances, inequity aver-
Incentives sion is disadvantageous, as people are willing to
forego a gain in order to prevent another person
An incentive is something that motivates an indi- from receiving a superior reward. Inequity aversion
vidual to perform an action. It is therefore essential has been studied through experimental games,
to the study of any economic activity. Incentives, particularly dictator, ultimatum, and trust games.
whether they are intrinsic or extrinsic (traditional), The concept has been applied in various domains,
can be effective in encouraging behavior change, including business and marketing, such as research
such as ceasing to smoke, doing more exercise, on customer responses to exclusive price promo-
complying with tax laws or increasing public good tions (Barone & Tirthankar, 2010) and “pay what
contributions. Traditional incentives can effectively you want” pricing (e.g. Regner, 2015).
encourage behavior change, as they can help to both
create desirable and break undesirable habits. Pro- Inertia
viding upfront incentives can help the problem of
present bias – people’s focus on immediate gratifi- In behavioral economics, inertia is the endurance
cation. Finally, incentives can help people overcome of a stable state associated with inaction and the
barriers to behavior change (Gneezy et al., 2019). concept of status quo bias (Madrian & Shea 2001).
Traditionally, the importance of intrinsic incen- Behavioral nudges can either work with people’s
tives was underestimated, and the focus was put decision inertia (e.g. by setting defaults) or against
on monetary ones. Monetary incentives may back- it (e.g. by giving warnings) (Jung, 2019). In social
fire and reduce the performance of agents or their psychology the term is sometimes also used in re-
compliance with rules (see also over-justification lation to persistence in (or commitments to) atti-
effect), especially when motives such as the desire tudes and relationships.
to reciprocate or the desire to avoid social disap-
proval (see social norms) are neglected. These in- Information avoidance
trinsic motives often help to understand changes in
behavior (Fehr & Falk, 2002). Information avoidance in behavioral economics
In the context of prosocial behavior, extrinsic (Golman et al., 2017) refers to situations in which
incentives may spoil the reputational value of good people choose not to obtain knowledge that is freely

Behavioral Economics Guide 2021 178

 Behavioral Science Concepts

available. Active information avoidance includes options. Option A included 40 pieces, nine of which
physical avoidance, inattention, the biased inter- were broken. Option B included 24 pieces, all of
pretation of information (see also confirmation which were intact. Option A was superior, as it in-
bias) and even some forms of forgetting. In be- cluded 31 intact pieces, but when evaluated sepa-
havioral finance, for example, research has shown rately, individuals were willing to pay a higher price
that investors are less likely to check their portfolio for set B. In a joint evaluation of both options, on
online when the stock market is down than when the other hand, Option A resulted in higher willing-
it is up, which has been termed the ostrich effect ness to pay (Hsee, 1998).
(Karlsson et al., 2009). More serious cases of avoid-
ance happen when people fail to return to clinics to Licensing effect
get medical test results, for instance (Sullivan et al.,
2004). Also known as ‘self-licensing’ or ‘moral licens-
While information avoidance is sometimes stra- ing’, the licensing effect is evident when people al-
tegic, it usually has immediate hedonic benefits for low themselves to do something bad (e.g. immoral)
people if it prevents the negative (usually psycho- after doing something good (e.g. moral) first (Mer-
logical) consequences of knowing the information. ritt et al., 2010). The effect of licencing has been
It usually carries negative utility in the long term, studied for different behavioral outcomes, includ-
because it deprives people of potentially useful in- ing donations, cooperation, racial discrimination,
formation for decision making and feedback for fu- and cheating (Blanken et al., 2015). Well-publicized
ture behavior. Furthermore, information avoidance research in Canada asked participants to shop ei-
can contribute to a polarization of political opinions ther in a green or a conventional online store. In
and media bias. one experiment, people who shopped in a green
store shared less money in a dictator game. Another
Intertemporal choice experiment allowed participants to lie (about their
performance on a task) and cheat (take more money
Intertemporal choice is a field of research con- out of an envelope than they actually earned) and
cerned with the relative value people assign to pay- showed more dishonesty among green shoppers
offs at different points in time. It generally finds (Mazar & Zhong, 2010).
that people are biased towards the present (see
present bias) and tend to discount the future (see Loss aversion
time discounting and dual-self model).
Loss aversion is an important concept associated
with prospect theory and is encapsulated in the ex-

L
pression “losses loom larger than gains” (Kahne-
man & Tversky, 1979a). It is thought that the pain
of losing is psychologically about twice as powerful
as the pleasure of gaining. People are more willing
to take risks (or behave dishonestly, e.g. Schindler
Less-is-better effect & Pfattheicher, 2016) to avoid a loss than to make
a gain. Loss aversion has been used to explain the
When objects are evaluated separately rather than endowment effect and sunk cost fallacy, and it may
jointly, decision makers focus less on attributes also play a role in the status quo bias.
that are important and are influenced more by at- The basic principle of loss aversion can explain
tributes that are easy to evaluate. The less-is-better why penalty frames are sometimes more effective
effect suggests a preference reversal when objects than reward frames in motivating people (Gäch-
are considered together instead of separately. One ter et al., 2009) and has been applied in behavior
study presented participants with two dinner set change strategies. The website Stickk, for example,

Behavioral Economics Guide 2021 179

 Behavioral Science Concepts

allows people to publicly commit to a positive be- each mental account separately, losing out the big
havior change (e.g. give up junk food), which may picture of the portfolio. (See also partitioning and
be coupled with the fear of loss—a cash penalty in pain of paying for ideas related to mental account-
the case of non-compliance. (See also myopic loss ing.)
aversion and regret aversion.) Consumers’ tendency to work with mental ac-
People’s cultural background may influence the counts is reflected in various domains of applied
extent to which they are averse to losses (e.g. Wang behavioral science, especially in the financial ser-
et al., 2017) vices industry. Examples include banks offering
multiple accounts with savings goal labels, which
make mental accounting more explicit, as well as

M
third-party services that provide consumers with
aggregate financial information across different
financial institutions (Zhang & Sussman, 2018).

Mindless eating
Mental accounting
Various cues non-consciously affect the amount
Mental accounting is a concept associated with the and quality of people’s consumption of food. Cues
work of Richard Thaler (see Thaler, 2015, for a sum- often serve as benchmarks in the environment, and
mary). According to Thaler, people think of value in they may include serving containers, packaging,
relative rather than absolute terms. For example, people, labels, and atmospheric factors. They sug-
they derive pleasure not just from an object’s val- gest to the consumer what and how much is normal,
ue, but also the quality of the deal—its transaction appropriate, typical, or reasonable to consume.
utility (Thaler, 1985). In addition, humans often fail Perceptual biases contribute to a distorted sense of
to fully consider opportunity costs (tradeoffs) and consumption; for example, people underestimate
are susceptible to the sunk cost fallacy. calories in larger servings and tend to serve them-
Why are people willing to spend more when they selves more when using larger utensils, plates, or
pay with a credit card than cash (Prelec & Simester, bowls (Wansink et al., 2009).
2001)? Why would more individuals spend $10 on a Brian Wansink, the most prominent academic in
theater ticket if they had just lost a $10 bill than if behavioral food science, has faced allegations of
they had to replace a lost ticket worth $10 (Kahne- scientific misconduct and several article retractions
man & Tversky, 1984)? Why are people more likely (Ducharme, 2018).
to spend a small inheritance and invest a large one
(Thaler, 1985)? Money illusion
According to the theory of mental accounting,
people treat money differently, depending on fac- The term ‘money illusion’ has been coined by Ir-
tors such as the money’s origin and intended use, ving Fisher (1928) and refers to people’s tendency
rather than thinking of it in terms of the “bottom to think of monetary values in nominal rather than
line” as in formal accounting (Thaler, 1999). An real terms. This usually occurs when we neglect to
important term underlying the theory is fungibili- consider money’s decrease in purchasing power as
ty, the fact that all money is interchangable and has a result of inflation. Investors, for example, may
no labels. In mental accounting, people treat assets focus on more salient nominal returns rather than
as less fungible than they really are. Even seasoned real returns that also account for inflation (Shafir
investors are susceptible to this bias when they view et al., 1997).
recent gains as disposable “house money” (Thal-
er & Johnson, 1990) that can be used in high-risk
investments. In doing so, they make decisions on

Behavioral Economics Guide 2021 180

 Behavioral Science Concepts

Myopic loss aversion Nudge

Myopic loss aversion occurs when investors take According to Thaler and Sunstein (2008, p. 6), a
a view of their investments that is strongly focused nudge is
on the short term, leading them to react too neg- any aspect of the choice architecture that
atively to recent losses, which may be at the ex- alters people’s behavior in a predictable way
pense of long-term benefits (Thaler et al., 1997). without forbidding any options or signifi-
This phenomenon is influenced by narrow framing, cantly changing their economic incentives.
which is the result of investors considering specif- To count as a mere nudge, the intervention
ic investments (e.g. an individual stock or a trade) must be easy and cheap to avoid. Nudges are
without taking into account the bigger picture (e.g. not mandates. Putting the fruit at eye level
a portfolio as a whole or a sequence of trades over counts as a nudge. Banning junk food does
time) (Kahneman & Lovallo, 1993). A large-scale not.
field experiment has shown that individuals who Perhaps the most frequently mentioned nudge is
receive information about investment performance the setting of defaults, which are pre-set courses of
too frequently tend to underinvest in riskier assets, action that take effect if nothing is specified by the
losing out on the potential for better long-term decision-maker. This type of nudge, which works
gains (Larson et al., 2016). with a human tendency for inaction, appears to be
particularly successful, as people may stick with a
choice for many years (Gill, 2018).

N
On a cost-adjusted basis, the effectiveness of
nudges is often greater than that of traditional ap-
proaches (Benartzi et al., 2017).
Questions about the theoretical and practical val-
ue of nudging have been explored (Kosters & Van
Naive allocation der Heijden, 2015) with respect to their ability to
produce lasting behavior change (Frey & Rogers,
Decision researchers have found that people pre- 2014), as well as their assumptions of irrationality
fer to spread limited resources evenly across a set and lack of agency (Gigerenzer, 2015). There may
of possibilities (see also 1/N heuristic). This can be also be limits to nudging due to non-cognitive
referred to as ‘naive allocation’. For example, con- constraints and population differences, such as a
sumers may invest equal amounts of money across lack of financial resources if nudges are designed
different investment options regardless of their to increase savings (Loibl et al., 2016). Limits in the
quality. Similarly, the diversification bias shows application of nudges speak to the value of experi-
that consumers like to spread out consumption mentation in order to test behavioral interventions
choices across a variety of goods. Research suggests prior to their implementation.
that choice architects can work with these tenden- As a complementary approach that addresses
cies due to decision makers’ partition dependence. the shortcomings of nudges, Hertwig and Grüne-
For instance, by separating healthy food menu op- Yanoff (2017) propose the concept of boosts, a deci-
tions into different menu categories (e.g. ‘fruits’, sion-making aid that fosters people’s competence
‘vegetables’) and combining unhealthy options into to make informed choices. (See also choice archi-
one single menu category (e.g. ‘candies and cook- tecture.)
ies’), one can steer consumers toward choosing
more healthy options and fewer unhealthy options 1/N (heuristic)
(Johnson et al., 2012).
1/N is a trade-off heuristic, one that assigns equal
weights to all cues or alternatives (Gigerenzer &

Behavioral Economics Guide 2021 181

 Behavioral Science Concepts

Gaissmaier, 2011). Under the 1/N rule, resources chances of success (Moore & Healy, 2008). Among
are allocated equally to each of N alternatives. For investors, overconfidence has been associated
example, in the (one-shot) ultimatum game, par- with excessive risk-taking (e.g. Hirshleifer & Luo,
ticipants most frequently split their money equally. 2001), concentrated portfolios (e.g. Odean, 1998)
Similarly, people often hedge their money in in- and overtrading (e.g. Grinblatt & Keloharju, 2009).
vestments by allocating equal amounts to different The planning fallacy is another example of over-
options. 1/N is a form of naive allocation of resourc- confidence, where people underestimate the length
es. of time it will take them to complete a task, often
ignoring past experience (Buehler et al., 1994). (See
also optimism bias.)

O Over-justification effect

This effect occurs when a person’s intrinsic in-

terest in a previously unrewarded activity decreas-
Optimism bias es after they engage in that activity as a means to
achieving an extrinsic goal (e.g. financial reward)
People tend to overestimate the probability of (Deci et al., 1999). As a result, the number of hours
positive events and underestimate the probability worked by volunteers, for instance, may be nega-
of negative events happening to them in the future tively affected by small financial rewards (Frey &
(Sharot, 2011). For example, we may underesti- Goette, 1999) (see also incentives).
mate our risk of getting cancer and overestimate
our future success on the job market. A number of

P
factors can explain unrealistic optimism, including
perceived control and being in a good mood (Hel-
weg-Larsen & Shepperd, 2001). (See also overcon-
fidence.)

Ostrich effect Pain of paying

See Information avoidance People don’t like to spend money. We experience

pain of paying (Zellermayer, 1996), because we are
Overconfidence (effect) loss averse. The pain of paying plays an important
role in consumer self-regulation to keep spending
The overconfidence effect is observed when peo- in check (Prelec & Loewenstein, 1998). This pain
ple’s subjective confidence in their own ability is is thought to be reduced in credit card purchases,
greater than their objective (actual) performance. because plastic is less tangible than cash, the deple-
It is frequently measured by having experimental tion of resources (money) is less visible, and pay-
participants answer general knowledge test ques- ment is deferred. Different personality types expe-
tions. They are then asked to rate how confident rience different levels of pain of paying, which can
they are in their answers on a scale. Overconfidence affect spending decisions. Tightwads, for instance,
is measured by calculating the score for a person’s experience more of this pain than spendthrifts. As a
average confidence rating relative to the actual pro- result, tightwads are particularly sensitive to mar-
portion of questions answered correctly. keting contexts that make spending less painful
A big range of issues have been attributed to over- (Rick, 2018). (See also mental accounting.)
confidence more generally, including the high rates
of entrepreneurs who enter a market despite the low

Behavioral Economics Guide 2021 182

 Behavioral Science Concepts

Partition dependence Planning fallacy

See Naive allocation Originally proposed by Kahneman and Tversky

(1979b), the planning fallacy is the tendency for
Partitioning individuals or teams to underestimate the time and
resources it will take to complete a project. This
The rate of consumption can be decreased by error occurs when forecasters overestimate their
physically partitioning resources into smaller units, ability and underestimate the possible risk associ-
for example cookies wrapped individually or mon- ated with a project. Without proper training teams
ey divided into several envelopes. When a resource of individuals can exacerbate this phenomena caus-
is divided into smaller units (e.g. several packs of ing projects to be based on the team’s confidence
chips), consumers encounter additional decision rather than statistical projections.
points—a psychological hurdle encouraging them One way to combat the planning fallacy is to use
to stop and think. In addition to the cost incurred a method termed Reference Class Forecasting (Fly-
when resources are used, opening a partitioned vbjerg et al., 2005; Kahneman & Tversky, 1979b).
pool of resources incurs a psychological trans- This method begins by creating a benchmark using
gression cost, such as feelings of guilt (Cheema & data on similar projects. Then estimates are built
Soman, 2008). Related research has found that sep- based on variances from the benchmark, depending
arate mental payment accounts (i.e. envelopes with on variables related to the project at hand. For ex-
money) can disrupt a shopping momentum effect ample, a construction company might estimate that
that may occur after an initial purchase (Dhar et al., building a house will take five weeks instead of the
2007). (For related ideas, see also mental account- average reference class time of six weeks, because
ing). the team at hand is larger and more skilled than
previous project teams. (See also optimism bias,
Peak-end rule overconfidence.)

According to the peak-end rule, our memory of Possibility effect

past experience (pleasant or unpleasant) does not
correspond to an average level of positive or nega- See Certainty/possibility effects
tive feelings, but to the most extreme point and the
end of the episode (Kahneman, 2000b). The rule de- Precommitment
veloped from the finding that evaluations of a past
episode seem to be determined by a weighted av- Humans need a continuous and consistent
erage of ‘snapshots’ of an experience, such as mo- self-image (Cialdini, 2008). In an effort to align fu-
ments in a film, thus neglecting its actual duration ture behavior, being consistent is best achieved by
(Fredrickson & Kahneman, 1993), as well research making a commitment. Thus, precommitting to a
showing that people would prefer to repeat a painful goal is one of the most frequently applied behavio-
experience if it is followed by a slightly less painful ral devices to achieve positive change. Committing
one (Kahneman et al., 1993). In terms of memories, to a specific future action (e.g. staying healthy by
remembered utility is more important than total going to the gym) at a particular time (e.g. at 7am on
utility (Kahneman, 2000a). People’s memories of Mondays, Wednesdays and Fridays) tends to better
prototypical moments are related to the judgments motivate action while also reducing procrastina-
made when people apply a representativeness heu- tion (Sunstein, 2014).
ristic (Kahneman, 2000b). The ‘Save More Tomorrow’ program, aimed at
helping employees save more money (Thaler & Be-
nartzi, 2004), illustrates precommitment alongside
other ideas from behavioral economics. The pro-

Behavioral Economics Guide 2021 183

 Behavioral Science Concepts

gram also avoids the perception of loss that would the present time when considering trade-offs be-
be felt with a reduction in disposable income, be- tween two future moments (O’Donoghue & Rabin,
cause consumers commit to saving future increases 1999). For example, a present-biased person might
in income. People’s inertia makes it more likely that prefer to receive ten dollars today over receiving
they will stick with the program, because they have fifteen dollars tomorrow, but wouldn’t mind wait-
to opt out to leave. ing an extra day if the choice were for the same
amounts one year from today versus one year and
Preference one day from today (see time discounting). The
concept of present bias is often used more generally
In economics, preferences are evident in theoret- to describe impatience or immediate gratification
ically optimal choices or real (behavioral) choices in decision-making.
when people decide between alternatives. Prefer-
ences also imply an ordering of different options Primacy effect
in terms of expected levels of happiness, gratifi-
cation, utility, etc. (Arrow, 1958). Measurement of See Serial-position effect
preferences may rely on willingness to pay (WTP)
and willingness to accept (WTA). Preferences are (Conceptual) Priming
sometimes elicited in survey research, which may
be associated with a range of problems, such as Conceptual priming is a technique and process
the hypothetical bias, when stated preferences are applied in psychology that engages people in a
different from those expressed in actual choices, or task or exposes them to stimuli. The prime con-
response effects, when subjects return the answer sists of meanings (e.g. words) that activate asso-
that they perceive the researcher ‘expects’. Armin ciated memories (schema, stereotypes, attitudes,
Falk and colleagues have developed cross-cultur- etc.). This process may then influence people’s
ally valid survey questions that are good predictors performance on a subsequent task (Tulving et al.,
of preferences in behavioral experiments. These 1982). For example, one study primed consumers
include questions about risk taking (see prospect with words representing either ‘prestige’ US retail
theory), social preferences (e.g. about reciprocity) brands (Tiffany, Neiman Marcus, and Nordstrom)
and time discounting (Falk et al., 2012). or ‘thrift’ brands (Wal-Mart, Kmart, and Dollar
Store). In an ostensibly unrelated task, partici-
Preference reversal pants primed with prestige names then gave higher
preference ratings to prestige as opposed to thrift
Preference reversal (Lichtenstein & Slovic, 1973) product options (Chartrand et al., 2008). Conceptu-
refers to a change in the relative frequency by which al priming is different from processes that do not
one option is favored over another in behavioral ex- rely on activating meanings, such as perceptual
periments, as may be evident in the less-is-better priming (priming similar forms), the mere expo-
effect or ratio bias, for example, or framing effects sure effect (repeated exposure increases liking),
more generally. The preferred ordering of a pair of affective priming (subliminal exposure to stimuli
choices is often found to depend on how the choice evokes positive or negative emotions) (Murphy &
is presented; this effect contradicts the predictions Zajonc, 1993), or the perception-behavior link (e.g.
of rational choice theory. (See also transitive/in- mimicry) (Chartrand & Bargh, 1999).
transitive preferences.) The technique of conceptual priming has become
a promising approach in the field of economics,
Present bias particularly in the study of the economic effects of
social identity (see identity economics) and social
The present bias refers to the tendency of people norms (Cohn & Maréchal, 2016).
to give stronger weight to payoffs that are closer to

Behavioral Economics Guide 2021 184

 Behavioral Science Concepts

GAINS LOSSES

HIGH PROBABILITY 95% chance to win $10,000 95% chance to lose $10,000

(Certainty Effect) Fear of disappointment Hope to avoid loss

RISK-AVERSE RISK-SEEKING

LOW PROBABILITY 5% chance to win $10,000 5% chance to lose $10,000

(Possibility Effect) Hope of large gain Fear of large loss

RISK-SEEKING RISK-AVERSE

Figure 1. Prospect Theory Quadrant

(Myopic) Procrastination Projection bias also affects choices in other set-

tings, such as medical decisions (Loewenstein,
People often put off decisions, which may be due 2005), gym attendance (Acland & Levy, 2015), cat-
to self-control problems (leading to present bias), alog orders (Conlin et al., 2007), as well as car and
inertia, or the complexity of decision-making (see housing markets (Busse et al., 2012).
choice overload). Various nudge tools, such as pre-
commitment, can be used to help individuals over- Prospect theory
come procrastination. Choice architects can also
help by providing a limited time window for action Prospect theory is a behavioral model that shows
(see scarcity heuristic) or a focus on satisficing how people decide between alternatives that involve
(Johnson et al., 2012). risk and uncertainty (e.g. % likelihood of gains or
losses). It demonstrates that people think in terms
Projection bias of expected utility relative to a reference point (e.g.
current wealth) rather than absolute outcomes.
In behavioral economics, projection bias refers Prospect theory was developed by framing risky
to people’s assumption that their own tastes or choices and indicates that people are loss-averse;
preferences will remain the same over time (Loe- since individuals dislike losses more than equiv-
wenstein et al., 2003). Both transient preferences in alent gains, they are more willing to take risks to
the short-term (e.g. due to hunger or weather con- avoid a loss. Due to the biased weighting of prob-
ditions) and long-term changes in tastes can lead abilities (see certainty/possibility effects) and loss
to this bias. For example, people may overestimate aversion, the theory leads to the following pattern
the positive impact of a career promotion due to an in relation to risk (Kahneman & Tversky, 1979a;
under-appreciation of (hedonic) adaptation, put Kahneman, 2011).
above-optimal variety in their planning for future Prospect theory has been applied in diverse eco-
consumption (see diversification bias), or underes- nomic settings, such as consumption choice, labor
timate the future selling price of an item by not tak- supply, and insurance (Barberis, 2013).
ing into account the endowment effect. Consumers’
under-appreciation of habit formation (associated
with higher consumption levels over time) may lead
to projection bias in planning for the future, such as
retirement savings.

Behavioral Economics Guide 2021 185

 Behavioral Science Concepts

R
by agreeing to the conceded request (Cialdini et al.,
1975).

Recency effect

Ratio bias See Serial-position effect

We find it harder to deal with proportions or ratios Recognition heuristic

than with absolute numbers. For example, when
asked to evaluate two movie rental plans with a While a core heuristic in the heuristics and biases
contracted scale (e.g. 7 and 9 new movies per week tradition of Tversky and Kahneman is availability,
for Plans A and B, respectively) as opposed to an a conceptually similar heuristic proposed in Gig-
equivalent offering with an expanded scale (364 and erenzer’s fast and frugal tradition is recognition.
468 movies per year, respectively), consumers favor In the fast and frugal view, the application of heu-
the better plan (Plan B) more in the scale expansion ristics is an “ecologically rational” strategy that
than contraction condition (Burson et al., 2009). makes best use of the limited information available
This is because our experiential system—unlike the to individuals (Goldstein & Gigerenzer, 2002). Rec-
rational system—encodes information as concrete ognition is an easily accessible cue that simplifies
representations, and absolute numbers are more decision-making and indicates that sometimes less
concrete than ratios or percentages (Kirkpatrick knowledge can lead to more accurate inferences. In
& Epstein, 1992). (See also framing, dual-system one experiment, participants had to judge which
theory, affect heuristic.) one of two cities had the greater population size. Re-
sults showed that the vast majority of choices were
Reciprocity based on recognition of the city name. What’s more,
the study indicated a less-is-more effect, whereby
Reciprocity is a social norm that involves in-kind people’s guesses are more accurate in a domain of
exchanges between people—responding to an- which they have little knowledge than one about
other’s action with another equivalent action. It is which they know a lot. American participants did
usually positive (e.g. returning a favor), but it can better on German cities, while German participants
also be negative (e.g. punishing a negative action) had higher scores on American cities (Goldstein &
(Fehr & Gächter, 2000). Reciprocity is of interest to Gigerenzer, 2002). (See also satisficing.)
behavioral economists because it does not involve
an economic exchange, and it has been studied by Reference dependence
means of experimental games (see behavioral game
theory). Organizations often apply reciprocity Reference dependence is one of the fundamental
norms in practice. Charities take advantage of reci- principles of prospect theory and behavioral eco-
procity if they include small gifts in solicitation let- nomics more generally. In prospect theory (Kahne-
ters (e.g. Falk, 2007), while hospitals may ask for- man & Tversky, 1979a), people evaluate outcomes
mer patients for donations (e.g. Chuan et al., 2018). relative to a reference point, and then classify gains
Reciprocity is also used as a social influence tool and losses (see also loss aversion, endowment ef-
in the form of ‘reciprocal concessions’, an approach fect). Reference dependence can apply to any deci-
also known as the ‘door-in-the-face’ technique. It sion involving risk and uncertainty. Online privacy
occurs when a person makes an initial large request research, for example, has shown that identical
(e.g. to buy an expensive product), followed up by privacy notices do not always result in the same
a smaller request (e.g. a less expensive option), if levels of disclosure (Adjerid et al., 2013). Consumers
the initial request is denied by the responder. The evaluate privacy notices relative to the status quo—
responder then feels obligated to ‘return the favor’ their current level of protection. When privacy no-

Behavioral Economics Guide 2021 186

 Behavioral Science Concepts

tices are preceded by notices that are less protec- er (2004) found that ‘gain frames’ in advertising
tive, people disclose more compared to those who (“Get energized”) lead to more favorable attitudes
have experienced no change in privacy protection. when the body of the advertising message is written
The converse is the case if preceding privacy notices in promotional terms (e.g. emphasizing the en-
are more protective. ergy benefits of drinking grape juice), whilst ‘loss
frames’ (“Don’t miss out on getting energized!”)
Regret aversion have a more favorable effect when the main body
of the ad focuses on prevention (e.g. stressing the
When people fear that their decision will turn cancer reduction benefits of drinking grape juice).
out to be wrong in hindsight, they exhibit regret
aversion. Regret-averse people may fear the conse- Representativeness heuristic
quences of both errors of omission (e.g. not buying
the right investment property) and commission Representativeness is one of the major general
(e.g. buying the wrong investment property) (Seiler purpose heuristics, along with availability”Availa-
et al., 2008). The effect of anticipated regret is par- bility heuristic”and affect. It is used when we judge
ticularly well-studied in the domain of health, such the probability that an object or event A belongs to
as people’s decisions about medical treatments. A class B by looking at the degree to which A resem-
meta-analysis in this area suggests that anticipated bles B. When we do this, we neglect information
regret is a better predictor of intentions and behav- about the general probability of B occurring (its
ior than other kinds of anticipated negative emo- base rate) (Kahneman & Tversky, 1972). Consider
tions and evaluations of risk (Brewer et al., 2016). the following problem:
(See also loss aversion, status quo bias, sunk cost
fallacy, fear of missing out, information avoid- Bob is an opera fan who enjoys touring art museums
ance, and action bias.) when on holiday. Growing up, he enjoyed playing chess
with family members and friends. Which situation is
Regulatory focus theory more likely?

The psychological theory of regulatory focus (Flo- A. Bob plays trumpet for a major symphony orchestra
rack et al., 2013; Higgins, 1998) holds that human B. Bob is a farmer
motivation is rooted in the approach of pleasure and
the avoidance of pain and differentiates a promo-
tion focus from a prevention focus. The former in- A large proportion of people will choose A in the
volves the pursuit of goals that are achievement- or above problem, because Bob’s description matches
advancement-related, characterized by eagerness, the stereotype we may hold about classical musi-
whereas the latter focuses on security and pro- cians rather than farmers. In reality, the likelihood
tection, characterized by vigilance. For example, of B being true is far greater, because farmers make
a person can become healthy by either engaging up a much larger proportion of the population.
in physical activity and eating organic food, or re- Representativeness-based evaluations are a com-
fraining from bad habits such as smoking or eating mon cognitive shortcut across contexts. For exam-
junk food. Prevention and promotion orientations ple, a consumer may infer a relatively high product
are a matter of both enduring dispositions and sit- quality from a store (generic) brand if its packaging
uational factors. is designed to resemble a national brand (Kardes
According to regulatory fit theory, messages and et al., 2004). Representativeness is also at work if
frames that are presented as gains are more in- people think that a very cold winter is indicative of
fluential under a promotion focus, whereas those the absence of global warming (Schubert & Stadel-
presented as losses carry more weight in a preven- mann, 2015) or when gamblers prefer lottery tickets
tion focus. For example, research by Lee and Aak- with random-looking number sequences (e.g. 7, 16,

Behavioral Economics Guide 2021 187

 Behavioral Science Concepts

23, …) over those with patterned sequences (e.g. Scarcity (heuristic)

10, 20, 30, ….) (Krawczyk & Rachubik, 2019). In fi-
nance, investors may prefer to buy a stock that had When an object or resource is less readily avail-
abnormally high recent returns (the extrapolation able (e.g. due to limited quantity or time), we tend
bias) or misattribute a company’s positive charac- to perceive it as more valuable (Cialdini, 2008).
teristics (e .g. high quality goods) as an indicator of Scarcity appeals are often used in marketing to in-
a good investment (Chen et al., 2007). duce purchases. Marketing messages with limited
quantity appeals are thought to be more effective
Risk-as-feelings than limited time appeals, because they create a
sense of competition among consumers (Aggarw-
‘Consequentialist’ perspectives of decision-mak- al et al., 2011). An experiment (Lee & Seidle, 2012)
ing under risk or uncertainty (risky-choice theo- that used wristwatch advertisements as stimuli ex-
ries, see e.g. prospect theory) tend to either focus posed participants to one of two different product
on cognitive factors alone or consider emotions as descriptions “Exclusive limited edition. Hurry, lim-
an anticipated outcome of a decision. ited stocks” or “New edition. Many items in stock”.
The risk-as-feelings hypothesis (Loewenstein et They then had to indicate how much they would be
al., 2001), on the other hand, also includes emotions willing to pay for the product. The average consum-
as an anticipatory factor, namely feelings at the er was willing to pay an additional 50% if the watch
moment of decision-making. was advertised as scarce.
In contrast to theories such as the affect heuristic, Scarcity can be used as an effective strategy by
where feelings play an informational role helping choice architects to get people who put off deci-
people to decide between alternatives, risk-as- sions (myopic procrastinators) to act (Johnson et
feelings can account for cases where choices (e.g. al., 2012).
refusal to fly due to a severe anxiety about air trav-
el) diverge from what individuals would objectively Scarcity (psychology of)
consider the best course of action.
People have a “mental bandwidth,” or brainpow-
er, made up of attention, cognition, and self-con-

S
trol (Mullainathan & Sharif, 2013), which consists
of finite resources that may become reduced or
depleted. The scarcity mindset entails a feeling of
not having enough of something. According to Mul-
lainathan and Sharif, anyone can experience cogni-
Satisficing tive scarcity, but it is particularly pronounced for
people living in poverty. On the positive side, this
According to Herbert Simon, people tend to make may induce limited focus that can be used produc-
decisions by satisficing (a combination of suffic- tively. The downside is ‘tunneling’, which inhibits
ing and satisfying) rather than optimizing (Simon, the cognitive power needed to solve problems, rea-
1956); decisions are often simply ‘good enough’ son, or retain information. Reduced bandwidth also
in light of the costs and constraints involved. As a impairs executive control, compromising people’s
heuristic, satisficing individuals will choose op- ability to plan and increasing impulsiveness where-
tions that meet their most basic decision criteria. A by the focus becomes immediate—put food on the
focus on satisficing can be used by choice architects table, find shelter, or pay the utility bill (See also
when decision makers are prone to procrastination present bias).
(Johnson et al., 2012). The financial and life worries associated with
poverty, and the difficult tradeoffs low-income in-
dividuals must make on a regular basis, all reduce

Behavioral Economics Guide 2021 188

 Behavioral Science Concepts

their cognitive capacity. Limits on self-control or can only be done with a phone call, and complicated
planning may lead some individuals to sacrifice fu- or long government student aid application forms.
ture rewards in favor of short-term needs. Procras- Even when a sludge is associated with a benefi-
tination over important tasks is also more likely, as cial behavior (as in student aid, voter registrations
is avoidance of expressing negative emotions. or driver’s licenses, for example), costs can be ex-
cessive. These costs may be a difficulty in acquiring
Self-control information, unnecessary amounts of time spent,
or psychological detriments, such as frustration
Self-control, in psychology, is a cognitive process (Sunstein, 2020).
that serves to restrain certain behaviors and emo-
tions vis-a-vis temptations and impulses. This as- Social norm
pect of self-regulation allows individuals to achieve
goals (Diamond, 2013). (See also intertemporal Social norms signal appropriate behavior and are
choice, present bias, dual-self model, dual-system classed as behavioral expectations or rules within
theory, ego depletion, and decision fatigue.) a group of people (Dolan et al., 2010). Social norms
of exchange, such as reciprocity, are different from
Serial-position effect market exchange norms (Ariely, 2008). Normative
feedback (e.g. how one’s energy consumption level
The serial-position effect refers to the finding compares to the regional average) is often used in
that items (e.g. word, picture or action) that are behavior change programs (Allcott, 2011) and has
located either at the beginning (primacy effect) been particularly effective to prompt pro-environ-
or end (recency effect) of a list are more easily re- mental behavior (Farrow et al., 2017). This feedback
membered (Ebbinghaus, 1913). These effects have can either be descriptive, representing what most
also been extensively studied in social psychology. people do for the purpose of comparison (e.g. “The
Research on persuasion, for example, has found majority of guests in this room reuse their towels”;
primacy effects to be stronger when the issue in a Goldstein et al., 2008), or injunctive, communicat-
message is relevant or familiar to individuals, and ing approved or disapproved behavior (e.g. “Please
recency effect more likely to occur when the issue is don’t….”, Cialdini et al., 2006). The latter is often
less relevant or familiar to them (Haugtvedt & We- more effective when an undesirable behavior is
gener, 1994; Lana, 1961). more prevalent than desirable behavior (Cialdini,
The serial-position effect should not be confused 2008).
with more general order effects, which refers to
context effects produced by the order of items, such Social preferences
as questions in a research instrument. (See also an-
choring and peak-end rule.) Social preferences (e.g. Fehr & Fischbacher, 2002)
are one type of preference investigated in behav-
Sludge ioral economics and relate to the concepts of reci-
procity, altruism, inequity aversion, and fairness.
The two defining characteristics of a sludge (Thal-
er, 2018) are “friction and bad intentions” (Gold- Social proof
hill, 2019). While Richard Thaler strongly advocates
nudging for good by making desirable behavior The influence exerted by others on our behav-
easier, a sludge does the opposite: It makes a pro- ior can be expressed as being either normative or
cess more difficult in order to arrive at an outcome informational. Normative influence implies con-
that is not in the best interest of the sludged. Exam- formity in order to be accepted or liked (Aronson et
ples of sludges include product rebates that require al., 2005), while informational influence occurs in
difficult procedures, subscription cancellations that ambiguous situations where we are uncertain about

Behavioral Economics Guide 2021 189

 Behavioral Science Concepts

how to behave and look to others for information Sunk cost fallacy
or cues. Social proof is an informational influence
(or descriptive norm) and can lead to herd behav- Individuals commit the sunk cost fallacy when
ior. It is also sometimes referred to as a heuristic. they continue a behavior or endeavor as a result
Research suggests that receiving information about of previously invested resources (time, money or
how others behave (social proof) leads to greater effort) (Arkes & Blumer, 1985). This fallacy, which
compliance among people from collectivist cul- is related to loss aversion and status quo bias, can
tures, whereas information on the individual’s past also be viewed as bias resulting from an ongoing
behavior (consistency/commitment) is associated commitment.
with greater compliance for people from individu- For example, individuals sometimes order too
alist cultures (Cialdini et al., 1999). much food and then over-eat just to "get their
money’s worth". Similarly, a person may have a $20
Status quo bias ticket to a concert and then drive for hours through
a blizzard, just because s/he feels that s/he has to
Status quo bias is evident when people prefer attend due to having made the initial investment.
things to stay the same by doing nothing (see also If the costs outweigh the benefits, the extra costs
inertia) or by sticking with a decision made previ- incurred (inconvenience, time or even money) are
ously (Samuelson & Zeckhauser, 1988). This may held in a different mental account than the one as-
happen even when only small transition costs are sociated with the ticket transaction (Thaler, 1999).
involved and the importance of the decision is great. Research suggests that rats, mice and humans are
Field data from university health plan enrol- all sensitive to sunk costs after they have made the
ments, for example, show a large disparity in health decision to pursue a reward (Sweis et al., 2018).
plan choices between new and existing enrollees.
One particular plan with significantly more favora- System 1/2
ble premiums and deductibles had a growing mar-
ket share among new employees, but a significantly See Dual-system theory
lower share among older enrollees. This suggests
that a lack of switching could not be explained by

T
unchanging preferences.
Samuelson and Zeckhauser note that status quo
bias is consistent with loss aversion, and that it
could be psychologically explained by previously
made commitments, sunk cost thinking, cogni-
tive dissonance, a need to feel in control and regret Take-the-best (heuristic)
avoidance. The latter is based on Kahneman and
Tversky’s observation that people feel greater re- Take-the-best is a simple decision-making
gret for bad outcomes that result from new actions shortcut that people may apply when choosing be-
taken than for bad consequences that are the con- tween alternatives. It is a one-reason decision rule,
sequence of inaction (Kahneman & Tversky, 1982). a type of heuristic where judgments are based on
While status quo bias is frequently considered a single “good” reason only, ignoring other cues
to be irrational, sticking to choices that worked in (Gigerenzer & Gaissmaier, 2011). Using the take-
the past is often a safe and less difficult decision the-best heuristic, a decision maker will base the
due to informational and cognitive limitations (see choice on one attribute that is perceived to discrim-
bounded rationality). For example, status quo bias inate most effectively between the options (Giger-
is more likely when there is choice overload (Dean enzer & Goldstein, 1996). Airport customs officers,
et al., 2017) or high uncertainty and deliberation for example, may determine whether a passenger is
costs (Nebel, 2015). selected for a search by choosing the best of vari-

Behavioral Economics Guide 2021 190

 Behavioral Science Concepts

ous cues, such as airport of origin, nationality, or nor occurs at a constant rate. It is usually studied by
amount of luggage (Pachur & Marinello, 2013). One asking people questions such as “Would you rather
study investigated voters’ perceptions of how US receive £100 today or £120 a month from today?” or
presidential candidates would handle the single is- “Would you rather receive £100 a year from today
sue that voters regarded as most important, such as or £120 a year and one month from today?” Results
the state of the economy or foreign policy. A model show that people are happier to wait an extra month
based on this issue (as a take-the-best attribute for a larger reward when it is in the distant future.
used by potential voters) correctly chose the winner In hyperbolic discounting, values placed on rewards
of the popular vote in 97% of all predictions (Graefe decrease very rapidly for small delay periods and
& Armstrong, 2012). then fall more slowly for longer delays (Laibson,
1997). (See also present bias.)
Take-the-first (heuristic) Research has shown different ways to reduce dis-
counting, such as primed future focus (Sheffer et
Take-the-first is a fluency heuristic. Fluen- al., 2016), mental simulation of future experiences
cy-based decision-making strategies occur when (e.g. Stein et al., 2016), and interactions with visual
different alternatives are recognized, but the one representations of one’s future self (Hershfield et
that is recognized faster is given higher value with al., 2011).
respect to a criterion (Gigerenzer & Gaissmaier,
2011). In the case of take-the-first, decision-mak- Transitive/intransitive preferences
ers simply choose the first alternative that comes
to mind (Johnson & Raab, 2003). Similar to other Preference transitivity is a hallmark of rational
fast and frugal approaches, this strategy is most choice theory. It holds that if, out of a set of options,
suitable in situations that present limitations to A is preferred to B and B to C, then A must also be
people’s ability to analyze information carefully. preferred to C (e.g. von Neumann & Morgenstern,
When experienced handball players were asked to 1947),. Intransitive preferences (i.e. C is preferred to
decide between taking a shot or passing the ball in A) violate the transitivity assumption and are some-
video sequences, the first option that came to mind times used to indicate System 1 vs 2 decision-mak-
tended to be superior to later options or a condition ing (Gallo et al., 2016). (See also preference reversal
under which when they had more time to analyze and decoy effect.)
the situation.
Trust
Time (temporal) discounting
Trust pervades human societies. It is indispensa-
Time discounting research investigates differ- ble in friendships, love, family, organizations and
ences in the relative valuation placed on rewards politics. Interpersonal trust is a mental construct
(usually money or goods) at different points in with implications for social functioning and eco-
time by comparing its valuation at an earlier date nomic behavior as studied by trust games, for ex-
with one for a later date (Frederick et al., 2002). ample.
Evidence shows that present rewards are weight- Although neoclassical economic theory suggests
ed more heavily than future ones. Once rewards that trust in strangers is irrational, trust and trust-
are very distant in time, they cease to be valuable. worthiness can be widely observed across socie-
Delay discounting can be explained by impulsivity ties. In fact, reciprocity exists as a basic element of
and a tendency for immediate gratification (see human relationships and behavior, and this is ac-
self-control), and it is particularly evident for ad- counted for in the trust extended to an anonymous
dictions such as nicotine (Bickel et al., 1999). counterpart (Berg et al., 1995). The nature of trust-
Hyperbolic discounting theory suggests that dis- ing behavior is a multi-faceted part of psychology,
counting is not time-consistent; it is neither linear investigated in terms of underlying dispositions,

Behavioral Economics Guide 2021 191

 Behavioral Science Concepts

U
intergroup processes, and cognitive expectations
(Evans & Krueger, 2009). Behavioral and biological
evidence indicates that trusting is not simply a spe-
cial case of risk-taking, but based rather on impor-
tant forms of social preferences, such as betrayal
aversion (Fehr, 2010). Ultimatum game
Both trust and trustworthiness increase when in-
dividuals are closer socially, but the latter declines The ultimatum game is an early example of re-
when partners come from different social groups, search that uncovered violations of standard as-
such as nationality or race. Furthermore, high sta- sumptions of rationality (see homo economicus). In
tus individuals are found to be able to elicit more the experiment, one player (the proposer/allocator)
trustworthiness in others (Glaeser et al., 2000). For is endowed with a sum of money and asked to split
example, CEOs are considerably more trusting and it between him/herself and an anonymous player
exhibit more trustworthiness than students. Trust (the responder/recipient). The recipient may either
seems to reinforce trustworthy behavior. In a be- accept the allocator’s proposal or reject it, in which
havioral experiment, trustworthiness was highest case neither of the players will receive anything.
when the threat to punish was available but not From a traditional game-theoretic perspective, the
used, and lowest when the threat to punish was allocator should only offer a token amount and the
actually used. Paradoxically, however, most CEOs recipient should accept it. However, results showed
and students used the punishment threat; although that most allocators offered more than just a token
CEOs made use of it significantly less (Fehr & List, payment, and many went as far as offering an equal
2004). split. Some offers were declined by recipients, sug-
gesting that they were willing to make a sacrifice
Trust game when they felt that the offer was unfair (see also
inequity aversion and fairness) (Guth et al., 1982).
Similar to the dictator game, this game asks (See also dictator game and trust game.)
participants to split money between themselves
and someone else. However, the trust game first Utility
asks Player A to determine an initial endowment of
zero or a higher value (e.g. $5). The money is then In economics, utility (e.g. Stigler, 1950) refers to
multiplied (e.g. tripled to $15) by the experimenter the benefits (satisfaction or happiness) consumers
and given to Player B, who is then asked to return derive from a good, and it can be measured based
an amount of zero or a higher value back to Player on individuals’ choices between alternatives or
A. The game is about reciprocity and trust, because preferences evident in their willingness to pay or
Player A must decide how much of the endowment accept. Behavioral economists have questioned
to give to Player B in the hope of receiving at least past assumptions that utility is always maximized,
the same amount in return. In the original exper- and they have worked with both traditional and new
iment (Berg et al., 1995), 30 out of 32 first players utility measures.
sent money, and 11 of these 30 decisions resulted in • Expected utility (Bernoulli, 1954 [1738]) has been
a payback that was greater than the initial amount used in economics as well as game and decision
sent. This finding confounds the prediction offered theory, including prospect theory, and is based on
by standard economic assumptions (see homo eco- choices with uncertain outcomes.
nomicus) that there would be no trust. However, • Discounted utility is a form of utility used in the
as with other games, critics have raised questions intertemporal choice domain of behavioral eco-
about what the trust game actually measures (Brül- nomics (Berns et al., 2007).
hart & Usunier, 2012). (See also ultimatum game.) • Experience(d) utility (Kahneman et al., 1997) re-
lates to actual (hedonic) experiences associated

Behavioral Economics Guide 2021 192

 Behavioral Science Concepts

with an outcome (in contrast to choice-based and WTA should be very close to WTP for a given
decision utility), which is associated with theories good.
on forecasting errors like the diversification bias. Behavioral economics, however, has shown that
• Remembered utility (Kahneman et al., 1997) sug- WTP and WTA may be context-dependent. For ex-
gests that people’s choices are also based on their ample, Thaler (1985) found evidence that people
memories of past events or experiences and is in- presented with a hypothetical scenario of lying on
voked in the peak-end rule. a beach and craving a beer would be willing to pay
• Instant utility and forecasted utility have been used significantly more for a beer purchased at a resort
in the area of intertemporal choice, such as re- hotel as opposed to a rundown grocery store (see
search on the empathy gap, showing that fore- also transaction utility and mental accounting). In
casted utility is biased in the direction of instant addition, sometimes the average WTA for a good
utility (Camerer & Loewenstein, 2004). exceeds its WTP, which may be indicative of an en-
• Procedural utility is relevant if people value not dowment effect, i.e. people value something more
only outcomes, but also the processes that lead to if they already own it. Research has also shown that
these outcomes (Frey, Benz, & Stutzer, 2004). the farther a good is from being an ordinary pri-
• Social utility has been proposed in relation to game vate (market) good, the more likely it is that WTA
theory, where players not only always act self-in- exceeds WTP. The WTA-to-WTP ratio is particu-
terestedly, but also show concerns about the per- larly high for health/safety and public/non-market
ceived intentions of other players and fairness goods (Horowitz & McConnel, 2002).
(Camerer, 1997).
• Transaction utility accounts for perceived merit Winner’s curse
or quality of a deal, rather than just the value of
a good or service relative to its price captured by The winner’s curse describes the phenomenon
acquisition utility (Thaler, 1985). that the winning bid of an auction tends to exceed
the true (and uncertain to the bidders) value of the
commodity, resulting, in effect, in the winner over-

W
paying. Emotion, cognitive biases and incomplete
information seem to account for this behavior,
which can, in extremis, lead to bubbles in the stock
or real estate markets.
In his seminal paper, “Anomalies: The Winner’s
Willingness to pay (WTP) / willingness to accept Curse”, Richard Thaler (1988) stated that if he were
(WTA) to auction of a jar of coins amongst his students, (1)
the average bid would be significantly less than the
In economics, willingness to accept (WTA) and actual value of the coins (bidders are risk averse)
willingness to pay (WTP) are measures of pref- and (2) the winning bid would exceed the value of
erence that do not rely on actual choices between the jar (even if it might be overpriced). This is not
alternative options. Instead, they ask individuals consistent with the idea of all bidders being ration-
to specify monetary amounts. WTA is a measure of al. In theory, if perfect information were available
the minimum financial compensation that a person to everyone and all participants were completely
would need in order to part with a good or to put up rational in their decision-making and skilled at
with something undesirable (such as pollution or valuation, no overpayments should occur. However,
crime). Willingness to pay (WTP) is the opposite— the winner’s curse, a robust and persistent devia-
the maximum amount of money someone is willing tion from theoretical predictions established in ex-
to pay for a good or to avoid something undesirable. perimental economics, reflects bounded rationality
According to standard economic intuition, WTP quite well, since people have difficulty in perform-
should be relatively stable across decision contexts ing contingent reasoning on future events (Char-

Behavioral Economics Guide 2021 193

 Behavioral Science Concepts

References
ness & Levin, 2009) (see intertemporal choice). Not
surprisingly, in an experimental demonstration of
the winner’s curse, the degree of uncertainty con- Acland, D., & Levy, M. R. (2015). Naiveté, projection
cerning the value of the commodity and the number bias, and habit formation in gym attendance.
of competing bidders were identified as the two fac- Management Science, 61(1), 146-160.
tors that affect the incidence and magnitude of this Adjerid, I., Acquisti, A., Brandimarte, L., &
curse (Bazerman & Samuelson, 1983). Loewenstein, G. (2013). Sleights of privacy:
In an attempt to overcome the winner’s curse, an Framing, disclosures, and the limits of transpar-
experiment has identified two factors that account ency. Proceedings of the Ninth Symposium on
for its persistence: a variability in the environment, Usable Privacy and Security (SOUPS). Retrieved
which leads to ambiguous feedback (i.e. choices and from http://cups.cs.cmu.edu/soups/2013/pro-
outcomes being only partially correlated), and the ceedings/a9_Adjerid.pdf
tendency of decision makers to learn adaptively. Aggarwal, P., Jun, S. Y., & Huh, J. H. (2011). Scarcity
Therefore, reducing the variance in the feedback messages. Journal of Advertising, 40(3), 19-30.
(such that choices and outcomes are correlated), Akerlof, G., & Kranton, R. (2000). Economics and
performance can be significantly improved (Bere- identity. The Quarterly Journal of Economics,
by-Meyer & Grosskopf, 2008). 115(3), 715-753.
Akerlof, G., & Kranton, R. (2005). Identity and the
economics of organizations. Journal of Economic

Z
Perspectives, 19(1), 9-32.
Allcott, H. (2011). Social norms and energy conser-
vation. Journal of Public Economics, 95(5), 1982-
2095.
An, S. (2008). Antidepressant direct-to-consumer
Zero price effect advertising and social perception of the prev-
alence of depression: Application of the avail-
The zero price effect suggests that traditional ability heuristic. Health Communication, 23(6),
cost-benefits models cannot account for the psy- 499-505.
chological effect of getting something for free. A Ariely, D. (2008). Predictably Irrational. New York:
linear model assumes that changes in cost are the Harper Collins.
same at all price levels and benefits stay the same. Arkes, H. R., & Blumer, C. (1985). The psychology of
As a result, a decrease in price will make a good sunk costs. Organizational Behavior and Human
equally more or less attractive at all price points. Decision Processes, 35, 124-140.
The zero price model, on the other hand, suggests Aronson, E., Wilson, T., & Akert, A. (2005). Social
that there will be an increase in a good’s intrinsic Psychology (5th ed.). Upper Saddle River, NJ:
value when the price is reduced to zero (Shampanier Prentice Hall.
et al., 2007). Free goods have extra pulling power, as Arrow, K. (1958). Utilities, attitudes, choices: A re-
a reduction in price from $1 to zero is more powerful view note. Econometrica, 26(1): 1-23.
than a reduction from $2 to $1. This is particular- Baca-Motes, K., Brown, A., Gneezy, A., Keenan, E.
ly true for hedonic products—things that give us A., & Nelson, L. D. (2012). Commitment and be-
pleasure or enjoyment (e.g. Hossain & Saini, 2015). havior change: Evidence from the field. Journal of
A core psychological explanation for the zero price Consumer Research, 39(5), 1070-1084.
effect has been the affect heuristic, whereby op- Banerjee, A. (1992). A simple model of herd behav-
tions that have no downside (no cost) trigger a more ior. Quarterly Journal of Economics, 107, 797-817.
positive affective response. Barberis, N. C. (2013). Thirty years of prospect
theory in economics: A review and assessment.
Journal of Economic Perspectives, 27(1), 173-96.

Behavioral Economics Guide 2021 194

 Behavioral Science Concepts

Barberis, N. C., & Thaler, R. H. (2003). A survey of Berger, S., Feldhaus, C., & Ockenfels, A. (2018). A
behavioral finance. In G. M. Constantinides, M. shared identity promotes herding in an infor-
Harris, & R. M. Stulz (Eds.), Handbook of the mation cascade game. Journal of the Economic
economics of finance (pp. 1053-1128). Elsevier. Science Association, 4(1), 63-72.
Bardsley, N. (2008). Dictator game giving: Altruism Bernoulli, D. (1954 [1738]). Exposition of a new the-
or artefact? Experimental Economics, 11(2), 122- ory on the measurement of risk. Econometrica,
133. 22(1): 23–36.
Bar-Eli, M., Azar, O. H., Ritov, I., Keidar-Levin, Y., & Berns, G. S., Laibson, D., & Loewenstein, G. (2007).
Schein, G. (2007). Action bias among elite soccer Intertemporal choice: Toward an integrative
goalkeepers: The case of penalty kicks. Journal of framework. Trends in Cognitive Sciences, 11(11),
Economic Psychology, 28(5), 606-621. 482-488.
Barone, M. J., & Tirthankar, R. (2010). Does exclu- Bickel, W., Odum, A., & Madden, G. (1999).
sivity always pay off? Exclusive price promotions Impulsivity and cigarette smoking: Delay dis-
and consumer response. Journal of Marketing, counting in current, never, and ex-smokers.
74(2), 121-132. Psychopharmacology, 146(4),447-454.
Bateman, I. J., Munro, A., & Poe, G. L. (2008). Decoy Bikhchandi, S., Hirschleifer, D., & Welch, I. (1992).
effects in choice experiments and contin- A theory of fads, fashion, custom and cultur-
gent valuation: Asymmetric dominance. Land al change as informational cascades. Journal of
Economics, 84(1), 115-127. Political Economy, 100, 992-1026.
Baumeister, R. F., Sparks, E. A., Stillman, T. F., & Blanken, I., van de Ven, N., & Zeelenberg, M. (2015).
Vohs, K. D. (2008). Free will in consumer be- A meta-analytic review of moral licensing.
havior: Self-control, ego depletion, and choice. Personality and Social Psychology Bulletin, 41(4),
Journal of Consumer Psychology, 18(1), 4-13. 540-558.
Bazerman, M. & Samuelson, W. (1983). I won the Brewer, N. T., DeFrank, J. T., & Gilkey, M. B. (2016).
auction but don’t want the prize. Journal of Anticipated regret and health behavior: A me-
Conflict Resolution, 27(4), 618-634. ta-analysis. Health Psychology, 35(11), 1264-1275.
Bénabou, R. & Tirole, J. (2003). Intrinsic and extrin- Brülhart, M., & Usunier, J.-C. (2012). Does the trust
sic motivation. Review of Economic Studies, 30, game measure trust? Economic Letters, 115, 20-
489-520. 23.
Bénabou, R. & Tirole, J. (2006). Incentives and Buehler, R., Griffin, D., & Ross, M. (1994). Exploring
prosocial behavior. American Economic Review, the “planning fallacy”: Why people underes-
96(5), 1652-1678. timate their task completion times. Journal of
Benartzi, S., Beshears, J., Milkman, K. L., Sunstein, Personality and Social Psychology, 67(3), 366-381.
C. R., Thaler, R. H., Shankar, M., Tucker-Ray, W., Burson, K. A., Larrick, R. P., & Lynch, J. G., Jr. (2009).
Congdon, W. J., & Galing, S. (2017). Should gov- Six of one, half dozen of the other: expanding
ernments invest more in nudging? Psychological and contracting numerical dimensions produces
Science, 28(8), 1041-1055. preference reversals. Psychological Science, 20(9),
Bereby-Meyer, Y. & Grosskopf, B. (2008). 1074-1078.
Overcoming the winner’s curse: An adaptive Busby, E., Flynn, D. J., & Druckman, J. N. (2018).
learning perspective. Behavioral Decision Making, Studying framing effects on political preferenc-
21(1), 15-27. es: Existing research and lingering questions. In
Berg, J., Dickhaut, J. & McCabe, K. (1995). Trust, rec- P. D'Angelo (Ed.), Doing News Framing Analysis II
iprocity, and social history. Games and Economic (pp. 67-90). New York: Routledge.
Behavior, 10(1), 122-142. Busse, M. R., Pope, D. G., Pope, J. C., & Silva-Risson,
Berger, L., Bleichrodt, H., & Eeckhoudt, L. (2013). J. (2012). Projection bias in the housing and car
Treatment decisions under ambiguity. Journal of markets. NBER Working Paper. Retrieved from
Health Economics, 32, 559-569. https://www.nber.org/papers/w18212.

Behavioral Economics Guide 2021 195

 Behavioral Science Concepts

Camerer, C. (2003). Behavioral game theory. reciprocity decays over time. Proceedings of the
Princeton, NJ: Princeton University Press. National Academy of Sciences, 115(8), 1766-1771.
Camerer, C. F. (1997). Progress in behavioral game Cialdini, R. B. (2008). Influence: Science and Practice,
theory. Journal of Economic Perspectives, 11, 167- 5th ed. Boston: Pearson.
188. Cialdini, R. B., Demaine, L. J., Sagarin, B. J., Barrett,
Camerer, C., Loewenstein, G., & Weber, M. (1989). D. W., Rhoads, K., & Winter, P. L. (2006).
The curse of knowledge in economic settings: Managing social norms for persuasive impact.
An experimental analysis. Journal of Political Social Influence, 1(1), 3-15.
Economy, 97(5), 1232-1254. Cialdini, R. B., Vincent, J. E., Lewis, S. K., Catalan,
Camerer, C. F., & Loewenstein, G. (2004). Behavioral J., Wheeler, D., & Darby, B. L. (1975). Reciprocal
economics: past, present and future. In C. F. concessions procedure for inducing compli-
Camerer, G. Loewenstein and M. Rabin (Eds.), ance: The door-in-the-face technique. Journal
Advances in Behavioral Economics (pp. 3-51). of Personality and Social Psychology, 31, 206-215.
Princeton, NJ: Princeton University Press. Cialdini, R. B., Wosinska, W., Barrett, D. W., Butner,
Chandon, P., & Wansink, B. (2007). The biasing J., Gornik-Durose, M. (1999). Compliance with a
health halos of fast-food restaurant health request in two cultures: The differential influ-
claims: Lower calorie estimates and higher ence of social proof and commitment/consisten-
side-dish consumption intentions. Journal of cy on collectivists and individualists. Personality
Consumer Research, 34(3), 301-314. and Social Psychology Bulletin, 25, 1242-1253.
Chapman, S., Wong, W. L., & Smith, W. (1993). Self- Cohn, A., Fehr, E. & Maréchal, M. (2014). Business
exempting beliefs about smoking and health: culture and dishonesty in the banking industry.
Differences between smokers and ex-smokers. Nature, 516, 86-89.
American Journal of Public Health, 83(2), 215-219. Cohn, A., & Maréchal, M. A. (2016). Priming in eco-
Charness, G., & Levin, D. (2009). The origin of the nomics. Current Opinion in Psychology, 12, 17-21.
Winner’s Curse: A laboratory study. American Conlin, M., O'Donoghue, T., & Vogelsang, T.
Economic Journal: Microeconomics, 1(1), 207-36. J. (2007). Projection bias in catalog orders.
Chartrand, T. L., & Bargh, J. A. (1999). The chamele- American Economic Review, 97(4), 1217-1249.
on effect: The perception-behavior link and so- Cornelissen, T., Himmler, O., & Koenig, T. (2011).
cial interaction. Journal of Personality and Social Perceived unfairness in CEO compensation and
Psychology, 76(6), 893-910. work morale. Economics Letters, 110, 45-48.
Chartrand, T. L., Huber, J., Shiv, B., & Tanner, R. Dean, M., Kibris, O., & Masatlioglu, Y. (2017).
(2008). Nonconscious goals and consumer Limited attention and status quo bias. Journal of
choice. Journal of Consumer Research, 35, 189-201. Economic Theory, 169, 93-127.
Cheema, A., & Soman, D. (2008). The effect of par- Deci, E. L., Koestner, R., & Ryan, R. M. (1999). A
titions on controlling consumption. Journal of meta-analytic review of experiments examining
Marketing Research, 45(6), 665-675. the effects of extrinsic rewards on intrinsic mo-
Chen, G., Kim, K. A., Nofsinger, J. R., & Rui, O. M. tivation. Psychological Bulletin, 125, 627-668.
(2007). Trading performance, disposition effect, Dhar, R., Huber, J., & Khan, U. (2007). The shopping
overconfidence, representativeness bias, and momentum effect. Journal of Marketing Research,
experience of emerging market investors. Journal 44, 370-378.
of Behavioral Decision Making, 20, 425-451. Diamond, A. (2013). Executive functions. Annual
Chernev, A., Böckenholt, U., & Goodman, J. (2015). Review of Psychology, 64, 135-168.
Choice overload: A conceptual review and me- Dickerson, C. A., Thibodeau, R., Aronson, E., &
ta-analysis. Journal of Consumer Psychology, Miller, D. (1992). Using cognitive dissonance to
25(2), 333-358. encourage water conservation. Journal of Applied
Chuan, A., Kessler, J. B., & Milkman, K. L. (2018). Social Psychology, 22(11), 841-854.
Field study of charitable giving reveals that Dolan, P., Hallsworth, M., Halpern, D., King, D., &

Behavioral Economics Guide 2021 196

 Behavioral Science Concepts

Vlaev, I. (2010). MINDSPACE: Influencing behav- Fehr, E. & Falk, A. (2002). Psychological founda-
iour through public policy. London, UK: Cabinet tions of incentives. European Economic Review,
Office. 46(4-5), 687-724.
Ducharme, J. (2018, September 21). A Prominent Fehr, E., & Fischbacher, U. (2002). Why social
researcher on eating habits has had more stud- preferences matter: The impact of non-selfish
ies retracted. Time. Retrieved from http://time. motives on competition, cooperation and incen-
com/5402927/brian-wansink-cornell-re- tives. The Economic Journal, 112(478), C1-C33.
signed/. Fehr, E., & Gächter, S. (2000). Fairness and retal-
Duhigg, C. (2012). The power of habit: Why we do what iation: The economics of reciprocity. Journal of‌‌‌‌‌‌
we do in life and business. New York: Random Economic Perspectives, 14, 159-181.
House. Fehr, E. & List, J. (2004). The hidden costs and re-
Easley, D., & O’Hara, M. (2009). Ambiguity and non- turns of incentives: Trust and trustworthiness
participation: the role of regulation. The Review among CEOs. Journal of the European Economics
of Financial Studies, 22(5), 1817-1843. Association, 2(5), 743-771.
Ebbinghaus, H. (1913). On memory: A contribution Fehr, E., & Schmidt, K. M. (1999). A theory of fair-
to experimental psychology. New York, NY: ness, competition, and cooperation. The Quarterly
Teachers College. Journal of Economics, 114, 817-868.
Economou, F., Hassapis, C., & Philippas, N. (2018). Festinger, L. (1957). A theory of cognitive disso-
Investors’ fear and herding in the stock market. nance. Stanford: Stanford University Press.
Applied Economics, 50(34-35), 3654-3663. Finucane, M. L., Alhakami, A., Slovic, P., & Johnson,
Ellsberg, D. (1961). Risk, ambiguity, and the savage S. M. (2000). The affect heuristic in judgments of
axioms. The Quarterly Journal of Economics, 75(4), risks and benefits. Journal of Behavioral Decision
643-669. Making, 13, 1-17.
Ericson, K. M. M., & Fuster, A. (2014). The endow- Fisher, G. S. (2014). Advising the behavioral inves-
ment effect. Annual Review of Economics, 6(1), tor: Lessons from the real world. In H. K. Barker
555-579. & V. Ricciardi (Eds.), Investor behavior: The psy-
Eşanu, E. (2019, November 1). 10 powerful user chology of financial planning and investing (pp.
nudges illustrated. UX Planet. Retrieved from 265-283). New York: John Wiley & Sons.
https://uxplanet.org/10-powerful-user-nudg- Fisher, I. (1928). The money illusion. New York:
es-illustrated-540ce4063f9a. Adelphi Company.
Evans, A. & Krueger, J. (2009). The psychology Fiske, S. T., & Taylor, S. E. (1991). Social Cognition
(and economics) of trust. Social and Personality (2nd ed.). New York: McGraw-Hill.
Psychology Compass, 3(6), 1003-1017. Florack, A., Keller, J., & Palcu, J. (2013). Regulatory
Falk, A. (2007). Gift exchange in the field. focus in economic contexts. Journal of Economic
Econometrica, 75, 1501-1511. Psychology, 38, 127–137.
Falk, A., Becker, A., Dohmen, T., Huffman, D. & Flyvbjerg, B., Skamris Holm, M. K., & Buhl, S. L.
Sunde, U. (2012). An experimentally validat- (2005). How (in)accurate are demand forecasts
ed preference module. Retrieved from http:// in public works projects? The case of transporta-
cups.cs.cmu.edu/soups/2013/proceedings/a9_ tion. Journal of the American Planning Association,
Adjerid.pdf 71(2), 131-146.
Farrow, K., Grolleau, G., & Ibanez, L. (2017). Social Folkes, V. S. (1988). The availability heuristic and
norms and pro-environmental behavior: A re- perceived risk. Journal of Consumer research,
view of the evidence. Ecological Economics, 140, 15(1), 13-23.
1-13. Frederick, S., & Loewenstein, G. (1999). Hedonic
Fehr, E. (2010). On the economics and biolo- adaptation. In D. Kahneman, E. Diener, & N.
gy of trust. Journal of the European Economics Schwarz (Eds.), Well-being: The foundations of
Association, 7(2-3), 235-266. hedonic psychology (pp. 302-329). New York:

Behavioral Economics Guide 2021 197

 Behavioral Science Concepts

Russell Sage Foundation. the fast and frugal way: Models of bounded ra-
Frederick, S., Loewenstein, G., & O’Donoghue, T. tionality. Psychological Review, 103, 650-669.
(2002). Time discounting and time preference: A Gill, D. (2018, February 20). When ‘nudging’ is for-
critical review. Journal of Economic Literature, 40, ever: The case of Sweden. Chicago Booth Review.
351-401. Retrieved from http://review.chicagobooth.edu/
Fredrickson, B. L., & Kahneman, D. (1993). Duration behavioral-science/2018/article/when-nudg-
neglect in retrospective evaluations of affec- ing-forever-case-sweden.
tive episodes. Journal of Personality and Social Gilovich, T., Griffin, D., & Kahneman, D. (Eds.).
Psychology, 65(1), 45-55. (2002). Heuristics and biases: The psychology of
Frey, B. S., & Goette, L. (1999). Does pay motivate intuitive judgment. Cambridge, UK: Cambridge
volunteers? Working Paper Series No. 7. Institute University Press.
for Empirical Research in Economics. University Glaeser, E., Laibson, D., Scheinkman, J. & Soutter, C.
of Zurich. (2000). Measuring trust. The Quarterly Journal of
Frey, B. S., Benz, M., & Stutzer, A. (2004). Introducing Economics, 115(3), 811-846.
procedural utility: Not only what, but also how Gneezy, U., Kajackaite, A., & Meier, S. (2019).
matters. Journal of Institutional and Theoretical Incentive-based interventions. Forthcoming in
Economics, 160, 377-401. the Handbook of Behavior Change.
Frey, E., & Rogers, T. (2014). Persistence: How Gneezy, U., Meier, S. & Rey-Biel, P. (2011). When and
treatment effects persist after interventions why incentives (don’t) work to modify behavior.
stop. Policy Insights from the Behavioral and Brain Journal of Economic Perspectives, 25(4), 191-210.
Sciences, 1(1), 172-179. Goldhill, O. (2019, July 31). Politicians love nudge
Fudenberg, D., & Levine, D. K. (2006). A dual-self theory. But beware its doppelgänger “sludge”.
model of impulse control. American Economic Quartz. Retrieved from: https://qz.com/1679102/
Review, 96(5), 1449-1476. sludge-takes-nudge-theory-to-new-manipu-
Furnham, A., & Boo, H. C. (2011). A literature review lative-levels/.
of the anchoring effect. The Journal of Socio- Goldstein, D. G., & Gigerenzer, G. (2002). Models of
Economics, 40(1), 35-42. ecological rationality: the recognition heuristic.
Gächter, S., Orzen, H., Renner, E., & Starmer, C. Psychological Review, 109(1), 75-90.
(2009). Are experimental economists prone to Goldstein, N. J., Cialdini, R. B., & Griskevicius, V.
framing effects? A natural field experiment. (2008). A room with a viewpoint: Using social
Journal of Economic Behavior & Organization, 70, norms to motivate environmental conservation
443-446. in hotels. Journal of Consumer Research, 35(3),
Gallo, I., Sood, S., Mann, T. C. , & Gilovich, T. (2016). 472-482.
The heart and the head: On choosing experi- Golman, R., Hagmann, D., & Loewenstein, G. (2017).
ences intuitively and possessions deliberative- Information avoidance. Journal of Economic
ly. Journal of Behavioral Decision Making. DOI: Literature, 55(1), 96-135.
10.1002/bdm.1997. Graefe, A., & Armstrong, J. S. (2012). Predicting
Gigerenzer, G. (2018), The bias bias in behavioral elections from the most important issue: A
economics. Review of Behavioral Economics, 5(3- test of the take‐the‐best heuristic. Journal of
4), 303-336. Behavioral Decision Making, 25(1), 41-48.
Gigerenzer, G. (2015). On the supposed evidence for Grant, A., Berg, J. & Cable, D. (2014). Job titles
libertarian paternalism. Review of Philosophy and as identity badges: How self-reflective titles
Psychology, 6, 361-383. can reduce emotional exhaustion. Academy of
Gigerenzer, G., & Gaissmaier, W. (2011). Heuristic Management Journal, 57(4), 1201-1225.
decision making. Annual Review of Psychology, Grinblatt, M., & Keloharju, M. (2009). Sensation
62(1), 451-482 seeking, overconfidence, and trading activity.
Gigerenzer, G., & Goldstein, D. G. (1996). Reasoning Journal of Finance, 64(2), 549-578.

Behavioral Economics Guide 2021 198

 Behavioral Science Concepts

Guth, W., Schmittberger, R., & Schwarz, B. (1982). An of WTA/WTP studies. Journal of Environmental
experimental analysis of ultimatum bargaining. Economics and Management, 44, 426-447.
Journal of Economic Behavior and Organization, 3, Hossain, M. T., & Saini, R. (2015). Free indulgences:
367-388. Enhanced zero-price effect for hedonic options.
Hadar, L., & Sood, S. (2014). When knowledge is International Journal of Research in Marketing,
demotivating: Subjective knowledge and choice 32(4), 457-460.
overload. Psychological Science, 25(9), 1739-1747. Hsee, C. K. (1998). Less is better: When low-value
Hagger, M. S., & Chatzisarantis, N. L. D. (2016). A options are valued more highly than high-value
multilab preregistered replication of the ego-de- options. Journal of Behavioral Decision Making, 11,
pletion effect. Perspectives on Psychological 107-121.
Science, 11, 546-573. Hugh-Jones, D. (2016). Honesty, beliefs about
Harley, E.M. (2007). Hindsight bias in legal decision honesty, and economic growth in 15 countries.
making. Social Cognition, 25(1), 48-63. Journal of Economic Behavior & Organization, 127,
Haugtvedt, C. P., & Wegener, D. T. (1994). Message 99-114.
order effects in persuasion: An attitude strength Iyengar, S., & Lepper, M. (2000). When choice is de-
perspective. Journal of Consumer Research, motivating: Can one desire too much of a good
21(1), 205-218. thing? Journal of Personality and Social Psychology,
Heath, C., & Heath, D. (2007). Made to stick: why 79, 995-1006.
some ideas survive and others die. New York: Johnson, E. J., & Goldstein, D. G. (2003). Do defaults
Random House. save lives? Science, 302, 1338-1339.
Helweg-Larsen, M., & Shepperd, J. A. (2001). Do Johnson, E. J., Shu, S. B., Dellaert, B. G.C., Fox, C. R.,
moderators of the optimistic bias affect personal Goldstein, D. G.,Häubl, G., Larrick, R. P., Payne, J.
or target risk estimates? A review of the litera- W., Peters, E., Schkade, D., Wansink, B., & Weber,
ture. Personality and Social Psychology Review, E. U. (2012). Beyond nudges: Tools of a choice ar-
5(1), 74-95. chitecture, Marketing Letters, 23, 487-504.
Hershfield, H. E., Goldstein, D. G., Sharpe, W. F., Fox, Johnson, J. G., & Raab, M. (2003). Take the first:
J., Yeykelvis, L., Carstensen, L. L., & Bailenson, Option generation and resulting choices.
J. (2011). Increasing saving behavior through Organizational Behavior and Human Decision
age-progressed renderings of the future self. Processes, 91, 215-229.
Journal of Marketing Research, 48, 23-37. Jolls, C. (2002). Fairness, minimum wage law, and
Hertwig, R., & Grüne-Yanoff, T. (2017). Nudging employee benefits. New York University Law
and boosting: Steering or empowering good Review, 77, 47-70.
decisions. Perspectives on Psychological Science, Jung. D. (2019, March 19). Nudge action: Overcoming
12(6), 973-986. decision inertia in financial planning tools.
Higgins, E. T. (1998). Promotion and prevention: Behavioraleconomics.com. Retrieved from https://
Regulatory focus as a motivational principle. www.behavioraleconomics.com/nudge-ac-
In M. P. Zanna (Ed.). Advances in Experimental tion-overcoming-decision-inertia-in-finan-
Psychology (Vol. 30, pp. 1–46). San Diego, CA: cial-planning-tools/.
Academic Press. Kahneman, D. (2000a). Experienced utility and ob-
Hinds, P. J. (1999). The curse of expertise: The ef- jective happiness: A moment-based approach. In
fects of expertise and debiasing methods on D. Kahneman & A. Tversky (Eds.), Choices, values,
prediction of novice performance. Journal of and frames (pp. 673–692). New York: Cambridge
Experimental Psychology: Applied, 5(2), 205-221. University Press.
Hirshleifer, D., & Luo, G. Y. (2001). On the survival of Kahneman, D. (2000b). Evaluation by moments:
overconfident traders in a competitive securities Past and future. In D. Kahneman & A. Tversky
market. Journal of Financial Markets, 4(1), 73-84. (Eds.), Choices, values, and frames (pp. 693–708).
Horowitz, J. K., & McConnell, K. E. (2002). A review New York: Cambridge University Press.

Behavioral Economics Guide 2021 199

 Behavioral Science Concepts

Kahneman, D. (2003). Maps of bounded rational- Tversky (Eds.), Choices, values and frames (pp.
ity: Psychology for behavioral economics. The 2-23). New York: Cambridge University Press.
American Economic Review, 93, 1449-1475. Kahneman, D., Wakker, P., & Sarin, R. (1997). Back
Kahneman, D. (2011). Thinking, fast and slow. to Bentham: Explorations of experienced utility.
London: Allen Lane. Quarterly Journal of Economics, 112, 375-405.
Kahneman, D., & Frederick, S. (2002). Kardes, F. R., Posavac, S. S., & Cronley, M. L.
Representativeness revisited: Attribute sub- (2004). Consumer inference: a review of pro-
stitution in intuitive judgment. In T. Gilovich, cesses, bases, and judgment contexts. Journal of
D. Griffin, & D. Kahneman (Eds.), Heuristics of Consumer Psychology, 14(3), 230-256.
intuitive judgment: Extensions and applications Karlsson, N., Loewenstein, G., & Seppi, D. (2009).
(pp. 49–81). New York: Cambridge University The ostrich effect: Selective attention to in-
Press. formation. Journal of Risk and Uncertainty, 38,
Kahneman, D., Fredrickson, B. L., Schreiber, C. A., & 95–115.
Redelmeier, D. A. (1993). When more pain is pre- King, J., & Slovic, P. (2014). The affect heuristic in
ferred to less: Adding a better end. Psychological early judgments of product innovations. Journal
Science, 4(6), 401-405. of Consumer Behaviour, 13(6), 411-428.
Kahneman, D., Knetsch, J. L., & Thaler, R. H. Kirkpatrick, L. A., & Epstein, S. (1992). Cognitive-
(1986). Fairness as a constraint on profit seek- experiential self theory and subjective probabil-
ing: Entitlements in the market. The American ity: further evidence for two conceptual systems.
Economic Review, 76(4), 728-741. Journal of Personality and Social Psychology, 63(4),
Kahneman, D., Knetsch, J. L., & Thaler, R. H. (1990). 534–544.
Experimental tests of the endowment effect and Kosters, M., & Van der Heijden, J. (2015). From
the Coase theorem. Journal of Political Economy, mechanism to virtue: Evaluating nudge theory.
98(6), 1325-1348. Evaluation, 21(3), 276-291.
Kahneman, D., Knetsch, J. L., & Thaler, R. H. (1991). Krämer, W., & Gigerenzer, G. (2005). How to con-
Anomalies: The endowment effect, loss aver- fuse with statistics or: the use and misuse of
sion, and status quo bias. Journal of Economic conditional probabilities. Statistical Science,
Perspectives, 5(1), 193-206. 20(3), 223-230.
Kahneman, D., & Lovallo, D. (1993). Timid choices Krawczyk, M. W., & Rachubik, J. (2019). The repre-
and bold forecasts: A cognitive perspective on sentativeness heuristic and the choice of lottery
risk taking. Management Science, 39, 17-31. tickets: A field experiment. Judgment and Decision
Kahneman, D., & Tversky, A. (1972). Subjective Making, 14(1), 51-57.
probability: A judgment of representativeness. Kruger, J., & Dunning, D. (1999). Unskilled and un-
Cognitive Psychology, 3, 430-454. aware of it: How difficulties in recognizing one’s
Kahneman, D., & Tversky, A. (1979a). Prospect own incompetence lead to inflated self-as-
theory: An analysis of decision under risk. sessments. Journal of Personality and Social
Econometrica, 47, 263-291. Psychology, 77(6), 1121-1134.
Kahneman, D., & Tversky, A. (1979b). Intuitive pre- Kruger, J., Wirtz, D., Van Boven, L., & Altermatt,
diction: Biases and corrective procedures. TIMS T. W. (2004). The effort heuristic. Journal of
Studies in Management Science, 12, 313-327. Experimental Social Psychology, 40(1), 91-98.
Kahneman, D., & Tversky, A. (1982). The psychology Laibson, D. (1997). Golden eggs and hyperbolic
of preference. Scientific American, 246, 160-173. discounting. Quarterly Journal of Economics, 112,
Kahneman, D., & Tversky, A. (1984). Choices, val- 443-477.
ues, and frames. American Psychologist, 39(4), Lana, R. E. (1961). Familiarity and the order of pres-
341-350. entation of persuasive communications. Journal
Kahneman, D., & Tversky, A. (1999). Evaluation by of Abnormal and Social Psychology, 62(3), 573-
moments: Past and future. In D. Kahneman & A. 577.

Behavioral Economics Guide 2021 200

 Behavioral Science Concepts

Larson, F., List, J. A., & Metcalfe, R. D. (2016). Can 96-105.

myopic loss aversion explain the equity premi- Loewenstein, G., O’Donoghue, T., & Rabin, M.
um puzzle? Evidence from a natural field exper- (2003). Projection bias in predicting future util-
iment with professional traders. NBER Working ity. Quarterly Journal of Economics, 118(4), 1209-
Paper. Retrieved from https://www.nber.org/ 1248.
papers/w22605. Loewenstein, G., Weber, E. U., Hsee, C. K., & Welch,
Latané, B., & Darley, J. (1970). The unresponsive N. (2001). Risk as feelings. Psychological Bulletin,
bystander: why doesn’t he help? New York: 127(2), 267-286.
Appleton-Century-Crofts. Loibl, C., Jones, L. E., Haisley, E., & Loewenstein,
Lee, A. Y., & Aaker, J. L. (2004). Bringing the frame G. (2016). Testing strategies to increase saving
into focus: The influence of regulatory fit on and retention in individual development account
processing fluency and persuasion. Journal of programs. https://papers.ssrn.com/sol3/papers.
Personality and Social Psychology, 86, 205-218. cfm?abstract_id=2735625
Lee, S. Y., & Seidle, R. (2012). Narcissists as consum- Madrian, B., & Shea, D. (2001). The power of sug-
ers: The effects of perceived scarcity on process- gestion: Inertia in 401(k) participation and sav-
ing of product information. Social Behavior and ings behavior. Quarterly Journal of Economics, 116,
Personality, 40(9), 1485-1500. 1149-1187.
Levin, I. P., Schneider, S. L., & Gaeth, G. J. (1998). Mathy, F., & Feldman, J. (2012). What’s magic about
All frames are not created equal: A typology and magic numbers? Chunking and data compres-
critical analysis of framing effects. Organizational sion in short-term memory. Cognition, 122(3),
Behavior and Human Decision Processes, 76, 149- 346–362.
188. March, J. G. (1978). Bounded rationality, ambiguity,
Levine, D. (1997). Modeling altruism and spite- and the engineering of choice. The Bell Journal of
fulness in experiments. Review of Economic Economics, 9(2), 587-608.
Dynamics, 1(3), 593-622. Marsh, L. E., Kanngiesser, P., & Hood, B. (2018).
Lichtenstein, D., Maxham, J. & Netemeyer, R. When and how does labour lead to love? The
(2010). The relationships among manager-, em- ontogeny and mechanisms of the IKEA effect.
ployee-, and customer-company identification: Cognition, 170, 245-253.
Implications for retail store financial perfor- Mazar, N., & Ariely, D. (2006). Dishonesty in every-
mance. Journal of Retailing, 86(1), 85-93. day life and its policy implications. Journal of
Lichtenstein, S., & Slovic, P. (1973). Reversals of Public Policy & Marketing, 25(1), 117-126.
preference between bids and choices in gambling Mazar, N., & Zhong, C. (2010). Do green products
decisions. Journal of Experimental Psychology, make up better people? Psychological Science, 21,
89(1), 46-55. 494-498.
Lin, M. C. (2018). The impact of aggregate uncer- Mazar, N., Amir, O., & Ariely, D. (2008). The dis-
tainty on herding in analysts' stock recommen- honesty of honest people: A theory of self-con-
dations. International Review of Financial Analysis, cept maintenance. Journal of Marketing Research,
57, 90-105. 45(6), 633-644.
List, J. A. (2011). Does market experience elimi- Mazzoni, G., & Vannucci, M. (2007). Hindsight bias,
nate market anomalies? The case of exogenous the misinformation effect, and false autobio-
market experience. American Economic Review, graphical memories. Social Cognition, 25(1), 203-
101(3), 313-17. 220.
Loewenstein, G. (2005). Hot-cold empathy gaps and McKenzie, C. R., Liersch, M. J., & Finkelstein, S. R.
medical decision-making. Health Psychology, (2006). Recommendations implicit in policy de-
24(Suppl. 4), S49-S56. faults. Psychological Science, 17(5), 414-420.
Loewenstein, G. (2005). Projection bias in medical Merritt, A., Effron, D. A., Monin, B. (2010). Moral
decision making. Medical Decision Making, 25(1), self-licensing: When being good frees us to be

Behavioral Economics Guide 2021 201

 Behavioral Science Concepts

bad. Social and Personality Psychology Compass, Ofir, C., Raghubir, P., Brosh, G., Monroe, K. B., &
4/5, 344-357. Heiman, A. (2008). Memory-based store price
Miller, G. A. (1956). The magical number seven, plus judgments: the role of knowledge and shopping
or minus two: Some limits on our capacity for experience. Journal of Retailing, 84(4), 414-423.
processing information. Psychological Review, Oosterbeek, H., Sloof, R., & van de Kuilen, G. (2004).
63(2), 81–97. Cultural differences in ultimatum game ex-
Mochon, D., Norton, M. I., & Ariely, D. (2008). periments: evidence from a meta-analysis.
Getting off the hedonic treadmill, one step at a Experimental Economics, 7, 171-188.
time: The impact of regular religious practice Oswald, M. E., & Grosjean, S. (2004). Confirmation
and exercise on wellbeing. Journal of Economic bias. In R. F. Pohl (Ed.), Cognitive illusions: A hand-
Psychology, 29, 632-642. book on fallacies and biases in thinking, judgement
Moore, D. A., & Healy, P. J. (2008). The trouble with and memory (pp. 79–96). New York: Psychology
overconfidence. Psychological Review, 115(2), Press.
502-517. Owens, D., Grossman, Z., & Fackler, R. (2014). The
Mullainathan, S., & Sharif, E. (2013). Scarcity: Why control premium: A preference for payoff auton-
having too little means so much. London: Allen omy. American Economic Journal: Microeconomics,
Lane. 6(4), 138-161.
Murphy, S. T., & Zajonc, R. B. (1993). Affect, cog- Pachur, T., & Marinello, G. (2013). Expert intuitions:
nition, and awareness: Affective priming with How to model the decision strategies of airport
optimal and suboptimal stimulus exposures. customs officers? Acta Psychologica, 144(1), 97-
Journal of Personality and Social Psychology, 64, 103.
723-729. Patt, A., & Zeckhauser, R. (2000). Action bias and
Nash, J. F. (1950). Equilibrium points in n-person environmental decisions. Journal of Risk and
games. Proceedings of the National Academy of Uncertainty, 21, 45-72.
Sciences, 36(1), 48-49. Piliavin, J. A., & Charng, H. W. (1990). Altruism: A
Nebel, J. M. (2015). Status quo bias, rationality, and review of recent theory and research. Annual
conservatism about value. Ethics, 125(2), 449- Review of Sociology, 16(1), 27-65.
476. Poses, R. M., & Anthony, M. (1991). Availability,
Newton, L. (1990). Overconfidence in the commu- wishful thinking, and physicians' diagnostic
nication of intent: heard and unheard melodies. judgments for patients with suspected bacter-
Unpublished doctoral dissertation, Stanford emia. Medical Decision Making, 11(3), 159-168.
University. Prelec, D., & Loewenstein, G. (1998). The red and the
Nickerson, R. S. (1998). Confirmation bias: A ubiq- black: Mental accounting of savings and debt.
uitous phenomenon in many guises. Review of Marketing Science, 17(1), 4-28.
General Psychology, 2, 175-220. Prelec, D., & Simester, D. (2001). Always leave home
Nisbett, R., & Wilson, T. D. (1977). The Halo Effect: without it: A further investigation of the cred-
Evidence for unconscious alteration of judg- it-card effect on willingness to pay. Marketing
ments. Journal of Personality and Social Psychology, Letters. 12(1), 5–12.
35, 250-256. Przybylski, A. K., Murayama, K., DeHaan, C. R., &
Norton, M. I., Mochon, D., & Ariely, D. (2012). The Gladwell, V. (2013). Motivational, emotional,
IKEA effect: When labor leads to love. Journal of and behavioral correlates of fear of missing out.
Consumer Psychology, 22, 453-460. Computers in Human Behavior, 29(4), 1841-1848.
Odean, T. (1998). Volume, volatility, price, and Ravaja, N., Aula, P., Falco, A., Laaksonen, S.,
profit when all traders are above average. Journal Salminen, M., & Ainamo, A. (2015). Online
of Finance, 53(6), 1887-1934. news and corporate reputation. Journal of Media
O’Donoghue, T., & Rabin, M. (1999). Doing it now or Psychology, 27(3), 118-133.
later. American Economic Review, 89(1), 103-124. Read, D., Loewenstein, G., Kalyanaraman, S. (1999).

Behavioral Economics Guide 2021 202

 Behavioral Science Concepts

Mixing virtue and vice: Combining the immedia- Samuelson, W., & Zeckhauser, R. J. (1988). Status
cy effect and the diversification heuristic. Journal quo bias in decision making. Journal of Risk and
of Behavioral Decision Making, 12, 257-273. Uncertainty, 1, 7-59.
Read, D., & Loewenstein, G. (1995). Diversification Sarstedt, M., Neubert, D., & Barth, K. (2017). The
bias: Explaining the discrepancy in variety seek- IKEA Effect. A conceptual replication. Journal of
ing between combined and separated choices. Marketing Behavior, 2(4), 307-312.
Journal of Experimental Psychology: Applied, 1, 34- Sayette, M. A., Loewenstein, G., Griffin, K. M., &
49. Black, J. J. (2008). Exploring the cold-to-hot
Regner, T. (2015). Why consumers pay voluntarily: empathy gap in smokers. Psychological Science,
Evidence from online music. Journal of Behavioral 19(9), 926-932.
and Experimental Economics, 57, 205-214. Schindler, S., & Pfattheicher, S. (2017). The frame of
Rick, S. I. (2018). Tightwads and spendthrifts: An the game: Loss-framing increases dishonest be-
interdisciplinary review. Financial Planning havior. Journal of Experimental Social Psychology,
Review, 1(1-2), e1010. Retrieved from https://doi. 69, 172-177.
org/10.1002/cfp2.1010. Schubert, R., & Stadelmann, M. (2015). Energy-
Rick, S. I., Cryder, C. E., & Loewenstein, G. (2008). using durables: Why consumers refrain from
Tightwads and spendthrifts. Journal of Consumer economically optimal choices. Frontiers in Energy
Research, 34, 767-782. Research, 3, 7.
Ring, P., Probst, C. C., Neyse, L., Wolff, S., Kaernbach, Schwartz, B. (2004). The paradox of choice: Why
C., van Eimeren, T., Camerer, C. F., & Schmidt, more is less. New York: Ecco.
U. (2018). It’s all about gains: Risk preferences Scott, P. J., & Lizieri, C. 92012). Consumer house
in problem gambling. Journal of Experimental price judgments: New evidence of anchoring and
Psychology: General, 147(8), 1241-1255. arbitrary coherence. Journal Of Property Research,
Rogers, P. (1998). The cognitive psychology of lot- 29, 49-68.
tery gambling: A theoretical review. Journal of Seiler, M., Seiler, V., Traub, S., & Harrison, D. (2008).
Gambling Studies, 14, 111-134. Regret aversion and false reference points in
Rushton, J. P. (1984). The altruistic personality. residential real estate. Journal of Real Estate
Development and Maintenance of Prosocial Research, 30(4), 461-474.
Behavior, 271-290. Shafir, E., Diamond, P., & Tversky, A. (1997). Money
Samson, A. (2014, February 25). A simple change illusion. The Quarterly Journal of Economics,
that could help everyone drink less. Psychology 112(2), 341-374.
Today. Retrieved from https://www.psychology- Shah, A. K., & Oppenheimer, D. M. (2008). Heuristics
today.com/intl/blog/consumed/201402/simple- made easy: An effort-reduction framework.
change-could-help-everyone-drink-less. Psychological Bulletin, 134(2), 207-222.
Samson, A., & Ramani, P. (2018, August 27). Finding Shampanier, K., Mazar, N., & Ariely D. (2007). Zero
the right nudge for your clients. Investment News. as a special price: The true value of free products.
Retrieved from https://www.investmentnews. Marketing Science, 26, 742-757.
com/article/20180827/BLOG09/180829939/ Sharot, T. (2011). The optimism bias. Current Biology,
finding-the-right-nudge-for-your-clients. 21(23), R941-R945.
Samson, A., & Voyer, B. (2012). Two minds, three Sheffer, C. E., Mackillop, J., Fernandez, A.,
ways: Dual system and process models in con- Christensen, D., Bickel, W. K., Johnson, M. W.,
sumer psychology. Academy of Marketing Science ... & Mathew, M. (2016). Initial examination of
Review, 2, 48–71. priming tasks to decrease delay discounting.
Samson, A., & Voyer, B. (2014). Emergency purchas- Behavioural Processes, 128, 144-152.
ing situations: Implications for consumer deci- Shefrin, H., & Statman, M. (1985). The disposition
sion-making. Journal of Economic Psychology, 44, to sell winners too early and ride losers too long:
21-33. Theory and evidence. The Journal of Finance, 40,

Behavioral Economics Guide 2021 203

 Behavioral Science Concepts

777-790. P., Glasgow, R., DeVellis, B., Meertens, R. M., &

Shepperd, J. A., Carroll, P., Grace, J., & Terry, M. Bulger, D. W. (1995). Goal setting as a strategy
(2002). Exploring the causes of comparative op- for health behavior change. Health Education
timism. Psychologica Belgica, 42, 65-98. Quarterly, 22, 190-200.
Shiller, R. J. (2015). Irrational exuberance. NJ: Sullivan, L. E. (2009). The SAGE glossary of the social
Princeton University Press. and behavioral sciences. Los Angeles, CA: SAGE.
Simon, H. A. (1956). Rational choice and the struc- Sullivan, P. S., Lansky, A., & Drake, A. (2004). Failure
ture of the environment. Psychological Review to return for HIV test results among persons at
63(2), 129-138. high risk for HIV infection: Results from a mul-
Simon, H. A. (1982). Models of bounded rationality. tistate interview project. JAIDS Journal of Acquired
Cambridge, MA: MIT Press. Immune Deficiency Syndromes, 35(5), 511–518.
Simonson, I. (1989). Choice based on reasons: Sunstein, C. R. (2014). Nudging: A very short guide.
The case of attraction and compromise effects. Journal of Consumer Policy, 37(4), 583-588.
Journal of Consumer Research, 16(2), 158-174. Sunstein, C. (2020). Sludge audits. Behavioural
Slaughter, J. E., Bagger, J., & Li, A. (2006). Context Public Policy. https://doi.org/10.1017/
effects on group-based employee selection bpp.2019.32.
decisions. Organizational Behavior and Human Sweis, B. M., Abram, S. V., Schmidt, B. J., Seeland, K.
Decision Processes, 100(1), 47-59. D., MacDonald, A. W., Thomas, M. J., & Redish,
Slovic, P., Finucane, M. L., Peters, E., & MacGregor, A. D. (2018). Sensitivity to “sunk costs” in mice,
D. G. (2002). The affect heuristic. In T. Gilovich, rats, and humans. Science, 361(6398), 178-181.
D. Griffin, & D. Kahneman (Eds.), Heuristics and Thaler, R. H. (1985). Mental accounting and con-
biases: The psychology of intuitive judgment sumer choice. Marketing Science, 4(3), 199-214.
(pp. 397-420). New York: Cambridge University Thaler, R. (1988). Anomalies: The Winner’s Curse.
Press. Journal of Economic Perspectives, 2(1), 191-202.
Slovic, P., Monahan, J., & MacGregor, D. M. (2000). Thaler, R. H. (1990). Anomalies: Saving, fungibili-
Violence risk assessment and risk communica- ty, and mental accounts. The Journal of Economic
tion: The effects of using actual cases, providing Perspectives, 4, 193-205.
instructions, and employing probability vs. fre- Thaler, R. H. (1999). Mental accounting matters.
quency formats. Law and Human Behavior, 24(3), Journal of Behavioral Decision Making, 12, 183-
271-296. 206.
Stein, J. S., Wilson, A. G., Koffarnus, M. N., Daniel, Thaler, R. H. (2015). Misbehaving: The making of
T. O., Epstein, L. H., & Bickel, W. K. (2016). behavioral economics. New York, NY: W. W.
Unstuck in time: Episodic future thinking re- Norton & Company.
duces delay discounting and cigarette smoking. Thaler, R. H. (2018). Nudge, not sludge. Science,
Psychopharmacology, 233(21-22), 3771-3778. 361(6401), 431.
Stigler, G. J. (1950). The development of utility the- Thaler, R. H., & Benartzi, S. (2004). Save More
ory. Journal of Political Economy, 58(4), 307-327. Tomorrow: Using behavioral economics to
Stoffel, S. T., Yang, J., Vlaev, I., & von Wagner, C. increase employee saving. Journal of Political
(2019). Testing the decoy effect to increase in- Economy, 112, S164-S187.
terest in colorectal cancer screening. PLOS ONE, Thaler, R. H., & Johnson, E. J. (1990). Gambling
14(3), e0213668. https://doi.org/10.1371/journal. with the house money and trying to break even:
pone.0219811Strack, F., & Deutsch, R. (2015). The effects of prior outcomes on risky choice.
The duality of everyday life: Dual-process and Management Science, 36(6), 643-660.
dual system models in social psychology. APA Thaler, R. H., & Shefrin, H. M. (1981). An econom-
Handbook of Personality and Social Psychology, 1, ic theory of self-control. Journal of Political
891-927. Economy, 89(2), 392-406.
Strecher, V. J., Seijts, G. H., Kok, G. J., Latham, G. Thaler, R. H., Sunstein, C. R., & Balz, J. P. (2013).

Behavioral Economics Guide 2021 204

 Behavioral Science Concepts

Choice architecture. In E. Shafir (Ed.), The behav- Mindless eating and healthy heuristics for the
ioral foundations of public policy (pp. 428-439). irrational. American Economic Review, 99, 165-
Princeton, NJ: Princeton University Press. 169.
Thaler, R. H., & Sunstein, C. (2008). Nudge: Wansink, B., Kent, R. J., & Hoch, S. J. (1998). An
Improving decisions about health, wealth, and anchoring and adjustment model of purchase
happiness. New Haven, CT: Yale University Press. quantity decisions. Journal Of Marketing Research,
Thaler, R. H., Tversky, A., Kahneman, D., & Schwartz, 35(1), 71–81.
A. (1997). The effect of myopia and loss aver- Wason, P. C. (1960). On the failure to eliminate hy-
sion on risk taking: An experimental test. The potheses in a conceptual task. Quarterly Journal of
Quarterly Journal of Economics, 112(2), 647-661. Experimental Psychology, 12(3), 129-140.
Thomas, M., Desai, K. K, & Seenivasan, S. (2011). Wijland, R. & Hansen, P. (2016). Mobile nudging:
How credit card payments increase unhealthy Youth engagement with banking apps. Journal of
food purchases: Visceral regulation of vices. Financial Services Marketing, 21, 51–63.
Journal of Consumer Research, 38, 505-524. Wood, W., & Neal, D. T. (2009). The habitual con-
Trautmann, S., & van de Kuilen, G. (2015). Ambiguity sumer. Journal of Consumer Psychology, 19, 579-
attitudes. The Wiley Blackwell Handbook of 592.
Judgement and Decision Making, 89-116. Zeelenberg, M., Van den Bos, K., Van Dijk, E., &
Tulving, E., Schacter, D. L., & Stark, H. A. (1982). Pieters, R. (2002). The inaction effect in the psy-
Priming effects in word fragment completion chology of regret. Journal of Personality and Social
are independent of recognition memory. Journal Psychology, 82(3), 314-327.
of Experimental Psychology: Learning, Memory and Zellermayer, O. (1996). The pain of paying. (Doctoral
Cognition, 8(4), 336-342. dissertation). Department of Social and
Tversky, A. (1972). Elimination by aspects: A theory Decision Sciences, Carnegie Mellon University,
of choice. Psychological Review, 79, 281-299. Pittsburgh, PA.
Tversky, A., & Kahneman, D. (1974). Judgment Zhang, C. Y., & Sussman, A. B. (2018). Perspectives
under uncertainty: Heuristics and biases. Science on mental accounting: An exploration of budg-
(New Series), 185, 1124-1131. eting and investing. Financial Planning Review,
Tversky, A., & Kahneman, D. (1981). The Framing of 1(1-2), e1011.
Decisions and the Psychology of Choice. Science,
211(4481), 453-458.
Tversky, A., & Kahneman, D. (1983). Extensional
versus intuitive reasoning: the conjunction fal-
lacy in probability judgement. Psychology Review,
90(4), 293-315.
Vohs, K. D., Baumeister, R. F., Schmeichel, B. J.,
Twenge, J. M., Nelson, N. M., & Tice, D. M. (2008).
Making choices impairs subsequent self-control:
A limited-resource account of decision making,
self-regulation, and active initiative. Journal of
Personality and Social Psychology, 94, 883-898.
Von Neumann, J., & Morgenstern, O. (1947), Theory
of games and economic behavior, Princeton, NJ:
Princeton University Press.
Wang, M., Rieger, M. O., & Hens, T. (2017). The
impact of culture on loss aversion. Journal of
Behavioral Decision Making, 30(2), 270-281.
Wansink B., Just, D. R., & Payne, C. R. (2009).

Behavioral Economics Guide 2021 205

Explore the mind of the consumer Program features
through The Chicago School’s
Behavioral Economics programs. Dedicated, engaged faculty
who are highly experienced
With foundations in advanced psychology, the Behavioral professionals and leaders in
Economics programs at The Chicago School provide students with their respective fields.
two pathways to building skills in understanding and influencing
consumer behavior: the Certificate in Behavioral Economics,
a customizable and abbreviated credential situated within the A student-faculty partnership
Behavioral Economics program, and the M.A. in Behavioral model that encourages
Economics, a traditional full master’s degree with elective options. collaborative work between
students and instructors,
Our M.A. in Behavioral Economics and Certificate in Behavioral enhancing professional,
Economics programs blend elements of consumer, social, and
academic, and community
cognitive psychology to provide a psychological perspective to
consumer behavior. engagement.

Those who earn their degree or certificate are prepared

to deliver professional services, perform research, excel Integrated learning that
as leaders and policy advisers, and serve diverse populations balances classroom instruction
in business, marketing, and politics with sensitivity and “real work” research
and inclusion. and application.

ABOUT THE CHICAGO SCHOOL

The Chicago School of Professional Psychology is a nonprofit, accredited A curriculum that values
institution with more than 5,700 students at campuses across the country exposure to a variety of
(Chicago, Dallas, Southern California, Washington, D.C., and online). The strategies for understanding
Chicago School has been an innovator in the field of psychology and related
and researching diverse human
behavioral sciences since 1979. The Chicago School offers more than 30
degree programs and several opportunities for international experiences. experience and behaviors.

CONTACT US TODAY TO LEARN MORE: 800-721-8072 | www.thechicagoschool.edu

M.A. IN BEHAVIORAL ECONOMICS
The online M.A. Behavioral Economics non-licensure program is designed for working
adults interested in psychological perspectives of human decision-making, risk
assessment, and consumer behavior. This program provides students an alternative to
the traditional MBA by offering a curriculum with a foundation in advanced psychology
that addresses broader business applications to decision-making, negotiation,
marketing, and consumer behavior.

The M.A. in Behavioral Economics utilizes a competency-based model grounded in

consumer, social, cognitive and consulting psychology, as well as political science and
infuses multicultural perspectives from diverse market audiences.The curriculum is
interdisciplinary in approach and integrates theories of consumer decision-making,
consulting, and financial literacy, including coursework in choice architecture,
neuromarketing, and persuasive messaging to generate a richer understanding of
human behavior.

Graduates are prepared to deliver professional services, perform research, excel as

leaders and policy advisers, and to sensitively and inclusively serve diverse populations
in business, marketing, and politics.

WHAT DISTINGUISHES THIS PROGRAM?

• The online Behavioral Economics M.A. program provides students with an alternative
to the traditional MBA by combining social psychological theory with a practical
application toward decision-making and consumer behavior within the context of a
psychology degree.

• The program is distinct from those of competing institutions both in its flexible
online delivery model and its curriculum, which blends elements of consumer,
social, and cognitive psychology while providing a psychological perspective to
behavioral economics.

• Upon successful completion of the online M.A. in Behavioral Economics program,

students who meet admissions requirements will be prepared to enter The Chicago
School’s Business Psychology Ph.D. program, allowing them to pursue additional
postgraduate and career opportunities.

CAREER POSSIBILITIES
Graduates can consider careers in the following fields:
• Consulting • Public relations • Human resources
• Public service • Health care • Nonprofit
• Marketing • Higher education • Government

CONTACT US TODAY TO LEARN MORE: 800-721-8072 | www.thechicagoschool.edu

M.A. PROGRAM M.A. STUDENT EXPERIENCE
SPECIFICATIONS The M.A. in Behavioral Economics program is designed to
support interaction and learning among students and faculty by
The M.A. in Behavioral Economics incorporating cohort membership, small groupings, a blended
is a non-licensure 40 credit hour delivery system, active learning, and pedagogical “best practices”
program. The program includes: within the design.

Cohort model: Students in the Behavioral Economics M.A. program

• 18 credit hours of core move through a sequence of courses collectively. The common
course work goal of starting and completing the program together encourages
students to work collectively, which promotes the development of
• 16 credit hours of research personal relationships and the building of a professional network.
course work Cohort membership enables students to support and learn from
other students.
• 6 credit hours of elective
course work Small groupings: The program strategically allows for arrangement
of students in small groups for online learning that is advantageous
The program culminates in for active learning. As approximations, online courses have fewer
an Applied Research Project than 20 students.
in which students will apply
behavioral economics concepts Diverse delivery system: This program utilizes both synchronous
to an approved topic. Students and asynchronous instructional modalities to provide students an
will complete classwork over the accommodative learning environment that encourages interaction
course of their studies that will among students and faculty, supports active learning, and respects
guide them through the process diverse talents and ways of learning. Asynchronous learning
of writing the Applied Research includes the use of online forums, as well as audio and video
Project. A faculty member will recordings. Synchronous learning includes the use of live chat
approve and supervise the project sessions and virtual meetings.
through these courses. Student services: Online students have access to a range of
students support services including: access to Library Services,
professional skill development through Career Services,
opportunities to study abroad, the chance to present original
research at the Graduate Research Forum, and engagement
opportunities through student groups and societies.

Certificate in Behavioral Economics

Also available is our Certificate in Behavioral Economics. This
program requires fewer credit hours than the M.A. yet also blends
behavioral economics and business psychology to provide a
unique alternative to a traditional MBA. Curriculum begins with an
introduction to the fundamentals of behavioral economics. Students
then choose two electives that suit their professional goals.
Total program credits: 9-10 credit hours
Length of program: 3 terms
Delivery format: online

CONTACT US TODAY TO LEARN MORE: 800-721-8072 | www.thechicagoschool.edu

EXECUTIVE MSc
BEHAVIOURAL SCIENCE
UNCOVER THE SCIENCE
BEHIND BEHAVIOUR
An increasing number of organisations now engage
with the idea of applying behavioural insights to
their organisational challenges.

The Executive MSc Behavioural Science, based in

LSE’s Department of Psychological and Behavioural
Science, is taught by experts at the forefront of
international research in behavioural science.

Our programme provides rigorous training for

professionals who are seeking to expand their
knowledge in this emerging and exciting field.
Many of our alumni are now prominent behavioural
science leaders and experts in a range of
organisations around the world.

CONTACT US
[email protected]

For further information, please visit

lse.ac.uk/EMScBehaviouralScience
EXECUTIVE MSc
BEHAVIOURAL SCIENCE
UNCOVER THE SCIENCE
BEHIND BEHAVIOUR
LSE’s Executive MSc Behavioural Science is taught by specialists at the
forefront of international research in behavioural science. Our programme
provides the opportunity for full-time professionals working in any sector to
obtain a graduate qualification in behavioural science, allowing you to pursue
new and expanded opportunities within this emerging and exciting field.

The programme starts in September each year with teaching being delivered during three
two-week intensive teaching blocks at the LSE campus in London. You are not required
to be in attendance on campus outside of these weeks and can therefore continue to
live and work outside of London and the UK. Between teaching sessions you work
independently on various assignments. After the final teaching session you complete
a dissertation on a topic of your choice with support from your supervisor.

The programme includes unique and innovative modules such as:

• Behavioural Science and Policy

• Behavioural Decision Science
• Research Methods for Behavioural Science
• Frontiers in Behavioural Science Methods
• Policy Appraisal and Ethics
• Behavioural Science in an Age of New Technology
• Corporate Behaviour and Decision Making
• Organisational Culture

OUR STUDENTS
Our students come from a wide range of academic and professional
backgrounds from all over the world, but one thing binds them
together: a passion for behavioural science and a desire to
CONTACT US
better understand how principles from behavioural science [email protected]
can be applied in their professional (and personal) lives.
For further information, please visit
lse.ac.uk/EMScBehaviouralScience
WHAT OUR ALUMNI HAVE TO SAY ABOUT THE PROGRAMME

I use my knowledge of behavioural science This degree provided networking

in my work all the time. Understanding how opportunities and access to academics
people respond and react allows me to design at a level that isn’t available anywhere
the workplace with this top of mind. else in the world.
Ellen, 2018 graduate Joshua, 2019 graduate

The EMSc in Behavioural Science was This is a life changing experience and
a deeply rewarding experience both you will be back to your professional
personally and professionally. life with a set of new tools that will
Claire, 2020 graduate change your career path.
Guilherme, 2021 graduate

CONTACT US
[email protected]

For further information, please visit

lse.ac.uk/EMScBehaviouralScience
 Learn the theory, apply the tools,
and make a difference
Penn’s Master of Behavioral and Decision Sciences program equips students with theoretical and
practical tools to understand how individuals and groups make decisions, how to affect those decisions,
and how social norms play a role in motivating and changing social behaviors. Led by world-renowned
faculty, researchers, and practitioners, MBDS creates unique opportunities for students to engage
with an exceptional advisory board, apply tools and knowledge in design challenges, and pursue
independent, cross-disciplinary research throughout Penn.

From our alumni:

“I facilitated a four-month behavioral consulting project for graduate students who

delivered a top-notch final brief and research paper for a real-world problem—
as good as anything I’ve seen from professional consulting firms. The students
superbly applied their knowledge of behavioral science to everyday situations
and answered tough questions from national and regional leaders during the
final presentation.”

Alex Willard, MBDS ‘19

Marketing Strategist, US Army Enterprise Marketing Office

“One of the most rewarding aspects of the MBDS program is the design challenge.
Not only does it provide the opportunity to apply theoretical concepts to solve
real-world industry challenges, but it also encourages you to proactively identify
existing the behavioral biases driving these very challenges. In my current role
as a data scientist, I find this to be a valuable skill that has enabled me to better
interpret the end result of certain decisions.”

Anu Raghuram, MBDS ‘19

Data Scientist, BlackRock

”Design challenges were a great way to apply what we learned in the classroom
towards solving real-world problems. These experiences taught me how to work
with stakeholders to define clear project objectives, design, and run effective field
research and use data to inform strategic decisions. These skills have proven to be
essential for my job in marketing research and guest insights.”

Paula Pereira, MBDS ‘20

Category Manager and Insights at Tim Hortons, Restaurant Brands International

Learn more about our engaged and well-connected alumni at

www.upenn.edu/mbds
Meet the Master of Behavioral and
Decision Sciences program’s
founding director
Cristina Bicchieri
Founding Director, Master of Behavioral and Decision Sciences
S.J. Patterson Harvie Professor of Social Thought and Comparative Ethics,
Departments of Philosophy and Psychology
Director, Center for Social Norms and Behavioral Dynamics

Cristina Bicchieri is a world authority on social norms and has

consulted with UNICEF, the World Bank, the Gates Foundation,
the United Kingdom’s Department for International Development,
and many other organizations. She is the founder of the Master of
Behavioral and Decision Sciences program, the Penn Social Norms
Group (PENN SoNG), and the Behavioral Ethics Lab. She is also the
“Wherever there is a human
Director of the Center for Social Norms and Behavioral Dynamics, a
group there are social norms.”
-Cristina Bicchieri
newly formed research center at Penn that aims to support positive
behaviors on a global scale.

The Center for Social Norms and

Behavioral Dynamics
The Center for Social Norms and Behavioral Dynamics at Penn, led
by Director Cristina Bicchieri, aims to support positive behaviors
on a global scale, across both informal and organizational settings.
The Center has undertaken a range of projects with partner
organizations around the world by leveraging their expertise in
measuring behavior, analyzing behavioral data, and identifying
systematic behavioral drivers. Master of Behavioral and Decision
Sciences students can explore research and learning opportunities
at the Center around social norms frameworks and theory. Through
the Center, both students and professionals have access to the
NoBeC (Norms and Behavioral Change) Talks, which showcase
interdisciplinary early career and senior researchers working on
norms and behavioral change around the world.

To learn more about the MBDS program’s world-renowned faculty and researchers, visit:

www.upenn.edu/mbds
Unparalleled connections,
exceptional opportunities
A defining feature of the University of Pennsylvania’s Master of Behavioral and Decision Sciences
program (MBDS) is its network of outstanding industry and research partners who help bring students
exceptional networking experiences, such as internships, design challenges, and employer-driven
projects.

From our advisory board:

“As more bright candidates enter the field, we recognize a growing need for higher
education to prepare and qualify the next generation of behavioral practitioners.
With its academic rigor and access to leading corporate field settings, Penn MBDS is
uniquely positioned to provide a best-in-class experience.”

Charlotte Blank
Chief Behavioral Officer, Maritz

“I have unique experiences in combining my academic background and industry

experience, and I love to be able to share these experiences to help people who are
interested in making a similar bridge in behavior science between academics and
industry.”

Namika Sagara
President, North America, Behavioral Science Center, Ipsos

“MBDS students are being taught to look at problems in more integrated and
human-centered ways. This is at the core of how we work at Deloitte Consulting,
and I’m delighted to support a program that is so well aligned with what the market
needs from consultants and other problem-solvers.”

Greg Szwartz
Practice Lead – Healthcare Predictive Modeling, Deloitte Consulting

About our Design Challenges

The MBDS program hosts Design Challenges every spring to drive collaboration between student
groups and industry partners who, together, integrate cutting edge knowledge into real-life problems.
This spring, fifty-four students teamed up to tackle twelve challenges in health, wellness, sustainability,
technology integration in marketing, and finance over six weeks.

Learn more about how MBDS connects students and industry at

www.upenn.edu/mbds
 MSc in Behavioural
and Economic Science
26 November 2015, 6.30 pm
26 November 2015, 6.30 pm
Modern Records Centre
Modern Records Centre

The Departments of Psychology and Economics at the University of

Warwick offer innovative new courses in the growing area of decision
science and behavioural economics. The MSc draws on the excellent,
ground-breaking research being undertaken in the departments of LAUNCH
Psychology, Economics and the Warwick Business School. EVENT

The MSc will suit those with a quantitive background ALL WELCOME
(e.g. maths, sciences, economics, psychology).

Further Details:
Email: [email protected] Tel: +44 (0)24 7657 5527

www.warwick.ac.uk/bes
 MSc in Behavioural and Economic Science

Why should I take

this course?
This inter-disciplinary course emphasises both theoretical foundations
and real-world applications of Behavioural Science, and is aimed at those
intending to work in business, public policy implementation or research.
26 November 2015, 6.30 pm
26Modern
November
Modules 2015,
Centre 6.30 pm
will include
Records
Modern Recordsgrounding
 A thorough Centre covering both the theory
and real-world application, in behavioural
economics and the cognitive science of
judgement and decision making.
 Modules on the design, conduction and analysis
of behavioural experiments and the analysis of
LAUNCH
large-scale datasets. EVENT
 An empirical research project.
ALL WELCOME
Our previous students have gone on to take positions at The Busara Center for Behavioral
Economics, The UK Behavioural Insights Team, Google, Frontier Economics, Facebook,
Ogilvy Change and more.

Further Details:
Email: [email protected] Tel: +44 (0)24 7657 5527

www.warwick.ac.uk/bes
 Why Warwick?
You will be taught by leading researchers from the Departments of
Psychology and Economics and Warwick Business School.
Three leading departments in this area of research.
26 Warwick
Novemberhas been2015, 6.30
ranked top pm
of the specialist subject table for Economics in
26Modern
November
The Times 2015,
and the
Records Sunday
Centre 6.30
TimespmUniversity League Tables for 2020. Behavioural
Science
Modern was identified
Records Centre as an area of significant academic achievement in the
Research Excellence Framework.
Warwick is a global community. Our students come from all over the world,
including South America, Asia, Europe, USA and the Middle East and from many
backgrounds including undergraduate study, industry and the public sector.
LAUNCH
Find out more about Postgraduate Study at Warwick EVENT
www.warwick.ac.uk/study/postgraduate ALL WELCOME

Further Details:
Email: [email protected] Tel: +44 (0)24 7657 5527

www.warwick.ac.uk/bes
 Postgraduate Programs
(Taught in English)

University School/Department Program

United States

Brown University School of Public Health Master of Public Health (Health Behavior con-
centration)

Department of Economics PhD in Economics

California Institute of Technology Division of the Humanities and PhD in Social and Decision Neuroscience
(Caltech) Social Science

Carnegie Mellon University Department of Social and PhD in Social and Decision Sciences
Decision Sciences

Tepper School of Business PhD in Behavioral Economics

(see also Dynamic Decision Making Laboratory)

(see also Center for Behavioral and Decision

Research)

Chapman University Economic Science Institute MS in Behavioral and Computational Economics

The Chicago School of Professional Masters in Behavioral Economics

Psychology
See pp. 206-208

Claremont Graduate University School of Social Science, Policy, PhD in Economics

and Evaluation
(see also Center for Neuroeconomics Studies)

Columbia University Columbia Business School MBA, MS, and PhD in Business

(see also Center for Decision Sciences)

Department of Economics MA and PhD in Economics

(see also Cognitive and Behavioral Economics

Initiative)

(see also Cognition & Decision Lab)

Cornell University Charles H. Dyson School of PhD in Applied Economics and Management
Applied Economics and
Management Master of Professional Studies (MPS) in
Applied Behavioral Economics and
Individual Choice

(see also Lab for Experimental Economics &

Decision Research)

(see also Cornell Center for Behavioral

Economics in Child Nutrition Programs)

Duke University The Fuqua School of Business MBA and PhD in Business Administration (Mar-
keting or Decision Sciences track)

Behavioral Economics Guide 2021 218

 Postgraduate Programs

Franklin University College of Arts, Sciences & Master’s in Business Psychology

Technology

Georgia State University Andrew Young School of Policy PhD in Economics

Studies
MA in Economics

(see also Experimental Economics Center)

Harvard University Department of Economics PhD in Economics

School of Public Health
MS and PhD in Social and Behavioral Sciences

Johns Hopkins University Johns Hopkins Bloomberg PhD in Social and Behavioral Sciences
School of Public Health

Massachusetts Institute of Technology Department of Brain and PhD in Brain and Cognitive Sciences
Cognitive Sciences
Masters in Management, Analytics,
MIT Sloan School of Applied Economics
Management
(see also MIT Sloan Neuroeconomics
Laboratory)

New York University Graduate School of Arts & MAs and PhDs in Economics, Politics and Psy-
Science chology

(see also Center for Experimental Social Science)

(see also Institute for the Study of Decision

Making)

Ohio State University Department of Psychology PhD in Psychology (Decision Psychology)

(see also Decision Sciences Collaborative)

Stanford University Department of Economics PhD in Economics (Behavioral & Experimental

specialization)

(see also Institute for Economic Policy Research)

Texas A&M University Department of Economics PhD in Economics

(see also Economic Research Laboratory)

University of Arizona Eller College of Management PhD in Economics

(see also Institute for Behavioral

Economics)

University of California, Berkeley Haas School of Business PhDs in Marketing, Psychology and
Economics
Department of Psychology
(see also Initiative for Behavioral Economics &
Department of Economics Finance)

(see also Berkeley Decision Science

Research Group)

University of California, Los Angeles Anderson School of Manage- PhD Behavioral Decision Making
ment

Behavioral Economics Guide 2021 219

 Postgraduate Programs

University of California, San Diego Rady School of Management MBA and PhD in Management

(see also Atkinson Behavioral Research Lab)

University of California, Santa Barbara College of Letters & Science PhD in Economics

(see also Experimental and Behavioral

Economics Laboratory)

University of Chicago Booth School of Business MBA

PhD in Behavioral Science

(see also Center for Decision Research)

University of Kansas College of Liberal Arts and MA in Applied Behavioral Science

Sciences
PhD in Behavioral Psychology

(see also KU Applied Behavioral

Economics Laboratory)

University of Maryland College of Behavioral & Social PhD in Social, Decision, and
Sciences Organizational Sciences

University of Oregon College of Arts and Science MA and PhD in Psychology

PhD in Economics

Lundquist College of Business PhD in Marketing

(see also Institute of Cognitive and

Decision Sciences)

University of Pennsylvania School of Arts & Sciences Master of Behavioral and Decision Sciences

See pp. 212-214

(see also Behavioral Ethics Lab)

(see also Social Norms Group)

University of Pittsburgh Katz Graduate School of PhD in Marketing

Business

Dietrich School of Arts & PhD in Economics

Sciences

University of Southern California Dana and David Dornsife PhD in Economics

College of Letters, Arts, and
Sciences (see also Los Angeles Behavioral
Economics Laboratory)

University of Wisconsin School of Human Ecology MS and PhD in Human Ecology:

Consumer Behavior and Family
Economics

(see also Behavioral Research Insights Through

Experiments Lab)

Washington University in St. Louis School of Arts and Sciences PhD in Behavior, Brain and Cognition

(see also Behavioral Economics Laboratory)

Behavioral Economics Guide 2021 220

 Postgraduate Programs

Yale University Yale School of Management Doctoral Programs in Financial

Economics, Marketing, and
Organizations and Management

(See also Yale-Ipsos Consumer Marketing &

Behavioral Economics Think Tank)

United Kingdom

Abertay University Division of Psychology MSc, MPhil, PhD in Behavioural Science

City University London Interdisciplinary MSc in Behavioural Economics

School of Arts and Social PhDs in Economics and Psychology

Sciences
(see also Decision Making and Behavioural
Economics Research Group)

Durham University Department of Psychology MSc in Behavioural Science

Durham Business School MSc in Experimental Economics

Kingston University Faculty of Arts and Social MSc in Behavioural Decision Science
Sciences

Lancaster University Management School PhD Behavioural and Experimental

Economics

London School of Economics and Department of Psychological MSc in Behavioural Science

Political Science and Behavioural Science
Executive MSc in Behavioural Science

See pp. 209-211

Departments of Management, PhDs in Management (Marketing), Social Pol-

Social Policy, Economics and icy, Economics and Psychological and Behav-
Psychological and Behavioural ioural Science
Science
(see also LSE Behavioural Lab for Teaching and
Research)

Middlesex University Business School MSc in Behavioural Economics in Action

Queen Mary University of London School of Economics and MSc in Behavioural Finance
Finance
University College London Division of Psychology And Executive Programme in Behavioural
Language Sciences Science

Division of Psychology And MSc in Cognitive and Decision Sciences

Language Sciences
MSc in Behaviour Change

School of Management and the PhD in Experimental Psychology

Behavioural Insights Team
PhDs in Management with Behavioural Science
and Policy

University of Bath MSc Applied Psychology and

Economic Behaviour

Behavioral Economics Guide 2021 221

 Postgraduate Programs

University of Cambridge Judge Business School MBA, Executive MBA and PhDs in
Business Economics, Marketing, etc.

Faculty of Economics PhD in Economics

(see also Cambridge Experimental and Behav-

ioural Economics Group)

University of East Anglia Department of Economics MSc in Behavioural and Experimental

Economics

MSc in Behavioural Economics and Data Sci-

ence

(see also Behavioural and Experimental Eco-

nomics Group)

University of Essex Department of Economics MSc in Behavioural Economics

University of Exeter School of Business MSc in Behavioural Economics and Finance

University of Huddersfield MSc in Behavioural Economics and Decision

Science

University of Leeds Leeds University Business School MSc in Business Analytics and Decision Scienc-
es

(see also Centre for Decision Research)

University of Nottingham School of Economics MSc in Behavioural Economics

PhD in Economics

(see also Centre for Decision Research and

Experimental Economics)

University of Oxford Department of Economics DPhil in Economics

(see also Behavioural Economics Research

Group)

(see also Nuffield Centre for Experimental

Social Sciences)

University of Reading Henley Business School MSc Behavioural Finance

Graduate Institute of Interna- MSc in Consumer Behaviour

tional Development, Agriculture
and Economics

University of Stirling Stirling Management School and MSc in Behavioural Science for
Behavioural Science Centre Management

(see also Behavioural Science Centre)

Behavioral Economics Guide 2021 222

 Postgraduate Programs

University of Warwick Interdisciplinary MSc in Behavioural and Economic Science

See pp. 215-217

Department of Psychology PhD in Psychology

(see also Behavioural Science Group)

Department of Psychology & De- MSc Behavioural and Data Science

partment of Computer Science

The Netherlands
Erasmus University Rotterdam Erasmus School of Economics Master in Economics and Business
(Behavioural Economics specialization)

PhD in Applied Economics (Behavioural Eco-

nomics group)

Leiden University Institute of Psychology Master in Psychology (Economic and

Consumer Psychology specialization)

Maastricht University School of Business and Master in Human Decision Science

Economics

Radboud University Nijmegen Department of Social Science Master in Behavioural Science

Master in Economics (Economics,

Behaviour and Policy specialization)

Tilburg University Department of Social Master in Social Psychology (Economic

Psychology Psychology track)

School of Social and Behavioral Research Master in Social and

Sciences Behavioral Sciences

Tilburg University Graduate Research Master and PhDs in

Schools Economics, Business (Marketing track) and
Social & Behavioural Sciences

(see also Tilburg Institute for

Behavioural Economics Research)

University of Amsterdam (Amsterdam School of Economics MSc in Economics (Behavioural Economics and
Business School / School of Economics) Game Theory track)

PhD in Economics (Behavioural Economics

research priority area)

University of Groningen Faculty of Behavioural and Research Master in Behavioural and Social
Social Sciences Sciences

Utrecht University Graduate School of Social and PhD in Social and Behavioural Sciences
Behavioural Sciences
(see also Behaviour in Social Context)

Wageningen University & Research MSc in Statistical Science for the Life and Be-
havioural Sciences

Behavioral Economics Guide 2021 223

 Postgraduate Programs

Germany

Friedrich-Schiller University Jena Jena Graduate School PhD in Human Behaviour in Social and Econom-
ic Change

Applied University at Hamm-Lippstadt Intercultural Business Psychology Masters

(Economic Psychology concentration)

Ludwig-Maximilians University Mu- Munich Graduate School of PhD in Economics

nich Economics
(see also Munich Experimental Laboratory for
Economic and Social Sciences)

TH Köln MA in Behavioral Ethics, Economics and

Psychology

University of Bonn Bonn Graduate School of PhD in Economics

Economics
(see also Center for Economics and Neurosci-
ence)

(see also Bonn Laboratory for Experi-

mental Economics)

University of Kassel MSc in Economic Behaviour and

Governance

University of Konstanz Graduate School of Decision PhDs at the Graduate School of Decision Scienc-
Sciences es (interdisciplinary)

Other Countries

Australia

Monash University Faculty of Business and Master of Applied Economics and Econometrics
Economics

School of Business, Monash PhDs in Business and Economics

University Malaysia.
(see also Monash Laboratory for
Experimental Economics)

(see also Monash Business

Behavioural Laboratory)

RMIT University Master of Business (Behavioural Economics

specialization)

PhD in Economics, Finance & Marketing (Be-

havioural Economics specialization)

(see also Behavioural Business Lab)

University of Melbourne School of Psychological Sciences Master of Applied Psychology

University of Queensland School of Economics Master and PhD in Economics

(see also Risk and Sustainable Management

Group)

Behavioral Economics Guide 2021 224

 Postgraduate Programs

University of Technology Sydney (UTS) UTS Business School PhD in Economics (Behavioural or Experimen-
tal Economics research field)

(See also UTS Behavioural Laboratory)

Austria

University of Vienna Faculty of Business, Economics, PhD in Economics

and Statistics
MSc in Economics

(see also Vienna Center for

Experimental Economics)

Sigmund Freud University Master in Psychology (Business & Economic

Psychology specialization)

Canada

University of British Columbia UBC Sauder School of Business PhD in Marketing and Behavioural Science

University of Saskatchewan Interdisciplinary PhD in Applied Economics (Research area in

Behavioural and Experimental Economics)

(See also Experimental Decision Laboratory)

University of Toronto Rotman School of Management MBAs and PhDs in Marketing and
Business Economics

(see also Behavioural Economics in Action)

Denmark

University of Copenhagen Department of Economics MSc and PhD in Economics

(See also Centre for Experimental Economics)

Finland

Oulu University in Finland Business School Master’s program in Economics

France

Paris School of Economics School of Economics Masters and PhDs in Economics

(see also Parisian Experimental Economics

Laboratory)

Toulouse School of Economics PhD in Economics (Behavioral and Experimen-

tal Economics specialization)

University Bourgogne Franche-Comté MSc in Behavioral and Digital Economics for

(UBFC) Effective Management

University of Paris Panthéon-Sorbonne Master in Economics & Psychology

/ University Paris Descrates

Behavioral Economics Guide 2021 225

 Postgraduate Programs

India

Christ University Master of Science in Behavioral Science

Ireland

Trinity College Dublin MSc Applied Behaviour Analysis

University College Dublin School of Economics MSc Behavioural Economics

Israel

Hebrew University of Jerusalem The Federmann Center for the PhDs at the Federman Center for the Study of
Study of Rationality Rationality (interdisciplinary)

IDC Herzliya Raphael Recanati International MA Behavioral Economics

School

Italy

Bocconi University in Milan Bocconi Experimental Laboratory for the Social

Sciences

Catholic University of the Sacred Heart, PhD School in Economics and PhD Economics and Finance
Milan Finance

LUISS (Libera Università Internazionale LUISS School Of European Polit- Master in Behavioral Science and
degli Studi Sociali Guido Carli) ical Economy Administration

University of Chieti-Pescara School of Advanced Studies PhD in Business and Behavioural Sciences

Master in Behavioral Economics & Neuromar-

keting

University of Trento Department of Economics and Master in Behavioural and Applied Economics
Management

Doctoral School of Sciences PhD in Economics and Management (Behav-

ioural Economics)

Norway
Norwegian School of Economics PhD in Business and Management Science

(see also the Choice Lab)

Portugal

Universidade Catolica Portuguesa Master in Psychology in Business and

Economics

Romania

Behavioral Economics Guide 2021 226

 Postgraduate Programs

University of Bucharest Faculty of Business and Master in Behavioural Economics

Administration & Faculty of
Psychology

Russia

National Research University Higher Master in Applied Social Psychology

School of Economics

Singapore

National University of Singapore NUS Business School MBA and PhDs in Management,
Decision Sciences and Economics

(see also Centre for Behavioural

Economics)

South Africa

University of Cape Town School of Economics Masters and PhD in Economics

(see also Research Unit in Behavioural Eco-

nomics and Neuroeconomics)

Spain

University of Barcelona Faculty of Psychology Master’s in Research in Behaviour and Cogni-

tion

Sweden

University of Gothenburg School of Business, Economics, PhD in Economics (Behavioural

and Law Economics concentration)

(see also Behavioural and Experimental Eco-

nomics Group)

Switzerland

Conférence Universitaire de Suisse PhD in Behavioral Economics and Experimental

Occidentale Research

University of Zurich (Zurich Graduate Department of Economics PhD in Economics and

School of Economics) Neuroeconomics

(see also Laboratory for Experimental and

Behavioral Economics)

Behavioral Economics Guide 2021 227

 Other
Resources
For the most up-to-date behavioral
science resources, including jobs, events,
popular books, and scholarly journals,
please visit

BehavioralEconomics.com
APPENDIX
 Author Profiles

John List (Introduction) Policy Program and the faculty chair of the executive
program “Global Leadership and Public Policy for
the 21st Century” for the World Economic Forum’s
Young Global Leaders. She serves on the boards, ad-
John A. List is Kenneth C. visory boards or as a patron of Credit Suisse Group,
Griffin Distinguished Service Applied, Edge, genEquality, TaketheLeadWomen,
Professor in Economics at the We Shape Tech, Women in Banking and Finance,
University of Chicago. His re- and the UK Government’s Equalities Office as well
search focuses on combining as numerous academic journals. She was named
field experiments with economic theory to deepen one of the Most Influential People in Gender Policy
our understanding of the economic science. In the by apolitical in 2018 and 2019, a Leading Thinker
early 1990s, List pioneered field experiments as a of Victoria, Australia, 2016-2019, and has received
methodology for testing behavioral theories and an honorary degree from the University of Lucerne,
learning about behavioral principles that are shared Switzerland, in 2016. She is married and the mother
across different domains. He was elected a Mem- of two children.
ber of the American Academy of Arts and Sciences
in 2011, and a Fellow of the Econometric Society Siri Chilazi (Guest editorial)
in 2015. List received the 2010 Kenneth Galbraith
Award, the 2008 Arrow Prize for Senior Economists
for his research in behavioral economics in the field,
and was the 2012 Yrjo Jahnsson Lecture Prize recip-
ient. He is a current editor of the Journal of Political Siri Chilazi is a Research Fel-
Economy. low at the Women and Public
Policy Program at Harvard
Iris Bohnet (Guest editorial) Kennedy School. Her life’s
work is to advance gender equality in the workplace
through research and research translation. As an
academic researcher, Siri specializes in identifying,
testing, and documenting specific interventions
Iris Bohnet, the Albert Pratt that work to close gender gaps and de-bias struc-
Professor of Business and tures and processes in organizations. As an advisor
Government, is the Academ- and speaker, Siri brings these evidence-based in-
ic Dean of Harvard Kennedy sights to practitioners and frequently collaborates
School. She is a behavioral economist, combining with organizations including Fortune 500 compa-
insights from economics and psychology to improve nies, top professional services firms, start-ups, and
decision-making in organizations and society, of- leading academic institutions. She has presented
ten with a gender or cross-cultural perspective. Her at numerous conferences around the world, and
most recent research examines behavioral design to her work regularly appears in leading media outlets
de-bias how we live, learn and work. She is the au- including Harvard Business Review, The New York
thor of the award-winning book What Works: Gender Times, BBC, Fast Company, and Forbes. Siri has an
Equality by Design, and advises governments and MBA from Harvard Business School, a Master in
companies on the topic around the world. Professor Public Policy from Harvard Kennedy School, and a
Bohnet is the co-director of the Women and Public BA in Chemistry and Physics from Harvard College.

Behavioral Economics Guide 2021 230

 Author Profiles

Alain Samson (Editor) sity, a MiM from the UCL School of Management,
and a Master of Science in Behavioural Decision
Making for Finance from the University of Stirling.
Collaborating at Behavioral Science Solutions since
Alain Samson is the editor 2015, he has a special passion and enthusiasm for
of the Behavioral Economics long-term decision making, investing, finance and
Guide, founder of Behavio- sustainability.
ralEconomics.com and Chief
Science Officer at Syntoniq. In
the past, he has worked as a consultant, researcher
and scientific advisor. His experience spans mul-
tiple sectors, including finance, consumer goods,
media, higher education, energy and government.
Alain studied at UC Berkeley, the University of
Michigan and the London School of Economics,
where he obtained a PhD in Social Psychology. His
scholarly interests have been eclectic, including
culture and cognition, social perception, consum-
er psychology and behavioral economics. He has
published articles in scholarly journals in the fields
of management, consumer behavior and economic
psychology. He is the author of Consumed, a Psy-
chology Today online popular science column about
behavioral science.

[email protected]

Andreas Haberl (Editorial Assistant)

Andreas Haberl is the edito-

rial assistant of the Behavioral
Economics Guide. Past expe-
riences include being an Argo
Scholar in Public Diplomacy at
the Embassy of Spain in Australia, doing an intern-
ship as research assistant at US behavioral Wealth-
Tech Syntoniq Inc and being Behavioral Design
Analyst at Monaco-based Panthera Solutions sarl.
He recently started a new role at Fisher Investments
Europe in London.
Andreas completed four postgraduate degrees in
Spain, England and Scotland. He holds a Dual-Li-
centiate (Premio Extraordinario) in Political Science
and Adverising from CEU Cardinal Herrera Univer-

Behavioral Economics Guide 2021 231

 Contributing Organizations

Think Forward Initiative several lines of research, one of them within the
collaboration with the Welfare Economics group of
Empower People to Make Better Financial Decisions the UNED. Currently, Behavior & Law is offering a
Master’s degree in Behavioral Economics, in col-
The Think Forward Initiative (TFI) is based on the laboration with the Madrid University UDIMA.
belief that our society is better off when people make
sound financial decisions. TFI’s Research Hub does www.behaviorandlaw.com
cutting-edge, data-driven research in social and
behavioural sciences to learn more about people’s
decision-making. TFI’s Accelerator Hub translates The Behavioral Insights Team
research insights into innovations. It scouts and
selects start-ups and supports them to scale fast- Starting life inside the UK Government the Be-
er and helps people to change their behaviour and havioral Insights Team (BIT) has evolved to become
improve their financial well-being. Lastly, TFI’s a world-leading social purpose company with over
Community Hub promotes the activities of the oth- 200 staff and with offices in London, Manchester,
er two hubs through our network of TFI community New York, Paris, Sydney, Toronto, Wellington and
members, and it launches campaigns to ensure our Singapore. Our mission is simple: we generate and
work reaches the people that need it the most. apply behavioural insights to inform policy, im-
prove public services and deliver positive results for
www.thinkforwardinitiative.com people and communities.
We work closely with central government, regu-
lators, national agencies, local government, local
Behavior & Law public services, voluntary sector agencies and so-
cially responsible businesses.
Behavior & Law is a company established in Flor- Our focus is on the practical, sometimes small
ida (US) and Madrid (Spain), dedicated to research, things, that make a big difference. It can be a single
scientific dissemination and teaching in behavioral sentence in a letter or a text message from a friend.
sciences and forensic sciences. Since its foundation Whether we’re working to encourage healthier
in 2008, it has specialized in the application of these eating, recycling or volunteering, we bring un-
sciences to the field of public and private security. rivalled expertise in behavioural science, a deep
In the area of public security, it has stood out for knowledge of public policy and the pragmatism to
its collaboration with police forces from different make change happen.
countries (Mexico, Colombia, Ecuador, USA, etc.),
obtaining various national and international ac- www.bi.team
knowledgements. Regarding private security, it has
stood out for the creation of the SAVE meta-proto-
col for fraud management, a method for training Dectech
teams within private companies to fight internal
and external forms of fraud. In recent years, large Dectech strives to provide the most accurate and
insurance and financial companies have been best value forecasts available on how people will
trained in this method. behave in new situations. Founded in 2002, we’ve
Behavior & Law has been intensifying its work conducted more than 400 studies involving over
in behavioral economics, currently focusing on three million participants. We hold that people

Behavioral Economics Guide 2021 232

 Contributing Organization Profiles

make very different decisions depending on their research centers and the Geisinger Commonwealth
context and often struggle to self-report their be- School of Medicine. With nearly 24,000 employees
liefs and motives. So, we developed Behaviourlab, a and more than 1,600 employed physicians, Geising-
randomised controlled trial approach that immers- er boosts its hometown economies in Pennsylvania
es participants in a replica of the real-world deci- by billions of dollars annually.
sion environment. Over the years we’ve shown how
Behaviourlab can provide higher accuracy forecasts www.geisinger.org
and more actionable insights.

www.dectech.co.uk GetReskilled

GetReskilled is a private education provider for

Frontier Economics the Pharmaceutical manufacturing industry, with
online programmes for experienced workers, fo-
Frontier Economics is a consulting firm with over cussed on the knowledge and skills necessary to
200 economists across London, Berlin, Brussels, manufacture safe medicines and vaccines.
Cologne, Dublin, Madrid and Paris. We specialise Their strong research team are focussed on devel-
in competition, regulation and strategy, across all oping a better understanding of the decision mak-
major sectors and areas of economic analysis. ing of experienced workers in the areas of career
Our clients benefit from objective advice, clearly change and lifelong learning, and how this behav-
expressed, that helps to inform key decisions. To iour aligns with current best practice in Behavioural
get to the heart of what matters, you need both Science. This research has particular relevance for
analytical expertise and creative problem solving. Government back to work initiatives.
Frontier Economics combines both to take on some Their work has been presented at major confer-
of the biggest questions facing business and society. ences including the World Conference on Online
We combine our expertise in economics with be- Learning in Dublin 2019 (WCOL), the International
havioural sciences to develop a richer picture of the Multi-Conference on Society, Cybernetics and In-
present, helping us to advise our clients on the right formatics 2020 (IMSCI), the International Confer-
decisions for them, for future success. We have one ence on Education Research and Innovation 2020
of the largest economic regulation practices in Eu- (ICERI), the International Technology, Education
rope - our behavioural economics work supports and Development Conference 2021 (INTED) and at
wider engagement with regulators and helps devel- the BEST Conference on Human Behaviour and De-
op regulatory policy. Our work on customer strategy cision Making 2021.
centres around understanding the actual behaviours
of our clients’ customers to help develop innovative www.getreskilled.com
customer-based solutions.

www.frontier-economics.com Gorilla Experiment Builder

We make powerful, flexible and intuitive software

Geisinger for pioneering behavioural science students, re-
searchers and practitioners to help them run novel
Geisinger is committed to making better health behavioural experiments on humans quickly, easily
easier for the more than 1 million people it serves. and cheaply.
Founded more than 100 years ago by Abigail Academics use our software to make discoveries
Geisinger, the system now includes nine hospital about all aspects of the human mind (for instance
campuses, a 550,000-member health plan, two memory, attention, language and emotions).

Behavioral Economics Guide 2021 233

 Contributing Organization Profiles

Students use our software to learn how to conduct planet with prosperous people. A planet free from
online research methods and prepare for careers the threat of climate crisis. People with basic human
that value digital experimentation (such as market rights, decent work, good labour conditions and, ul-
research and advertising). timately, good financial health. As a partner it is also
Practitioners use our software to design and run strongly involved in the Think Forward Initiative,
behavioural experiments that provide their clients a multi-year movement bringing together experts
with behavioural insights. The outcomes of these representing governments, academics, consumers,
experiments deliver value to clients in a wide range and the financial and technology sectors with the
of scenarios. aim of developing tools that can help people make
Providing behavioural scientists with the tools conscious and informed choices about money.
needed accelerate their research will liberate this
community to creatively use their knowledge to https://www.ing.com/Sustainability/Sustainabili-
discover behavioural insights that address a wide ty-direction.htm
range of challenges in society.

www.gorilla.sc National Acoustic Laboratories

Established over 70 years ago, the National

IAG Acoustic Laboratories (NAL), the research division
of Hearing Australia, is a world-leader in hearing
IAG is the parent company of a general insurance research and evidence-based innovation. NAL’s fo-
group (the Group) with controlled operations in cus is applied research aimed at improving hearing
Australia and New Zealand. health and transforming the lives of people with
The Group’s businesses underwrite over $12 bil- hearing difficulties.
lion of premium per annum, selling insurance un- NAL’s Behavioural Sciences department inves-
der many leading brands, including: NRMA Insur- tigates the human element at the heart of hearing
ance, CGU, SGIO, SGIC, Swann Insurance and WFI health. Through the application of knowledge from
(Australia); and NZI, State, AMI and Lumley (New the field of behavioural science, they seek to pro-
Zealand). IAG also has an interest in a general in- mote better hearing health by empowering individ-
surance joint venture in Malaysia. uals to take action - to prevent hearing loss and/or
IAG’s applied behavioural science capability engage with hearing rehabilitation
strives to improve customer outcomes and is pur- Most recently, the team has been focused on re-
pose-led, to make the world a safer place. search within the clinical setting, seeking to identify
and address cognitive biases that create barriers for
www.iag.com.au optimal hearing health. This work pays particular
attention to how information is presented and in-
terpreted within the audiologist-patient dynamic,
ING and how nudges for both parties can create envi-
ronments for more informed decision-making and
ING is a global financial institution with a strong improved outcomes.
European base, offering banking services through
its operating company ING Bank. The purpose of www.nal.gov.au
ING Bank is empowering people to stay a step ahead
in life and in business. ING Bank’s more than 57,000
employees offer retail and wholesale banking ser- OECD
vices to customers in over 40 countries.
Global Sustainability helps to create a healthy The Organisation for Economic Co-operation and

Behavioral Economics Guide 2021 234

 Contributing Organization Profiles

Development (OECD) is an international organisa- goals, while investing their funds in ways that will
tion that works to promote policies to improve the create a positive future for them, their families,
economics and social well-being of people around their communities and broader society. In this way,
the world. Together with governments, policy OML significantly contributes to improving peo-
makers and citizens, the OECD works on establish- ples’ lives, while ensuring a sustainable future for
ing evidence-based international standards and the group.
finding solutions to a range of social, economic and
environmental challenges. The OECD’s Executive www.oldmutual.com
Directorate is responsible for managing and co-
ordinating the corporate services of the organisa-
tion, including, for example, financial and human Quadrangle Consulting Services
resources management, information technology
and internal communication services as well as Quadrangle Consulting Services is a boutique
operational services including building infrastruc- consulting firm that offers a wide range of talent
ture and logistics. It has been applying innovative and customer solutions built on theories and mod-
approaches in management including the use of els from behavioral sciences. Our team of social
behavioural insights within human resource man- and organization psychologists work with over 200
agement, digital security and health and safety, as cross-sector clients. Using our skills in psychomet-
well as enhancing and debiasing decision-making rics, statistics and research, we deliver comprehen-
to make behaviour change. sive outcomes that are reliable and predictive.
We offer customer and talent profiling research
www.oecd.org & services, bespoke behavioral assessment, nudge
based communication design, as well as consulta-
tion in customer & people process design. We rou-
Old Mutual tinely develop assessment instruments and surveys
to measure behavior and create predictive models
Founded in Cape Town, South Africa in 1845, Old around retention and performance for large scale
Mutual Limited (OML) is a premium African finan- implementation.
cial services group that offers a broad spectrum of Our solutions are tailored to the demographic and
financial solutions to retail and corporate custom- economic diversity of talent and customer segments
ers. OML operates in 14 countries, and employs over in the Indian Subcontinent.
30,000 people across Africa and Asia.
The key services of the group include: www.quadrangleconsulting.org
1. Providing life assurance-based solutions and
short term insurance.
2. Growing customers’ savings and wealth Rare
through active and direct asset management, and
using multi-managers to select funds for cus- Rare is the leading behavior change organization
tomer investments. in the environmental field, recognizing that con-
3. Supplying loans and debt consolidation ser- servation ultimately comes down to people. For over
vices tailored to the individual needs of our cus- 45 years and across 60 countries, Rare has support-
tomers. ed individuals, their communities, and their local
4. Offering a low-cost transactional account leaders to adopt behaviors that benefit both people
linked to a unit trust savings account, known as and nature. The Center for Behavior & the Environ-
the Money Account. ment translates science into practice and leverages
OML’s purpose is to help its customers thrive by the best behavioral insights and design thinking
enabling them to achieve their lifetime financial approaches to tackle some of the most challenging

Behavioral Economics Guide 2021 235

 Contributing Organization Profiles

environmental issues. Through partnerships with strategy in both B2C and B2B. We have been award-
leading academic institutions and with environ- ed numerous prizes for developing research meth-
mental practitioners, we are translating the science odology (for example, ‘ESOMAR Congress Award’
of human behavior into practical solutions for con- 2012 for the ‘Best Methodological Paper’) as well
servationists worldwide. as for proving that the application of behavioral
economics has a significant impact on our clients’
behavior.rare.org return on investment (for example, ‘ESOMAR Re-
search Effectiveness Award’ 2012 and 2013 for the
most effective market research projects). Founded
UK Finance in 1999, Vocatus is headquartered in Munich, Ger-
many, and serves clients all over the world.
UK Finance is the collective voice for the bank-
ing and finance industry. Representing nearly 300 www.vocatus.com
firms, through research, policy expertise, thought
leadership and advocacy it acts to enhance compet-
itiveness, customer engagement and innovation,
promoting a safe, transparent and innovative sec-
tor. Its Academies and discussion forums promote
members’ knowledgeable engagement with many
stakeholder interests.
UK Finance’s Conduct and Culture Academy (CCA)
promotes effective use of behavioural science among
financial service providers. Co-founded by Dr Roger
Miles, it brings together senior finance practition-
ers to socialise human-factor risk insights and
approaches, to lead constructive engagement with
regulators, and to identify and report on significant
components of ‘purposeful culture’. More than 100
conduct leaders – and several conduct regulators –
have graduated the Academy’s core course. 4000+
financial business unit leaders in more 20 countries
have vigorously supported UK Finance’s Conduct
Focus participative workshops, awarding these
consistently 5* approval ratings.
Anthologies of CCA-related research include Con-
duct Risk Management: A Behavioural Approach
(2017) and Culture Audit in Financial Services (2021)
(both Kogan Page).

www.ukfinance.org.uk

Vocatus

Vocatus is a market research and consulting com-

pany specializing in the application of behavioral
economics to portfolio, product, sales and pricing

Behavioral Economics Guide 2021 236