Energy Economics 87 (2020) 104698

Contents lists available at ScienceDirect

Energy Economics

journal homepage: www.elsevier.com/locate/eneeco

Energy efficiency as a credence good: A review of informational barriers

to energy savings in the building sector
Louis-Gaëtan Giraudet
Ecole des Ponts ParisTech, Centre international de recherche sur l'environnement et le développement (CIRED), 45 bis, avenue de la Belle Gabrielle, 94736 Nogent-sur-Marne Cedex, France

a r t i c l e i n f o a b s t r a c t

Article history: Information problems have early been suspected to be the main barrier to energy-efficiency investment. I review
Received 11 March 2019 the vast yet piecemeal research that has been carried out since. Focusing on energy efficiency in buildings, I or-
Received in revised form 27 November 2019 ganize the review around the concept of credence good: just like that of auto repairs or taxi rides, the quality
Accepted 31 January 2020
of energy-efficiency measures is never fully revealed to the buyer; as a result, it is subject to multiple information
Available online 4 February 2020
asymmetries. My first contribution is to distinguish symmetric-information problems from information
Keywords:
asymmetries. The former arise when information is either incomplete or imperfect, but equally shared by
Energy-efficiency gap contracting parties; as non-market failures, these can be addressed by technological progress and insurance mar-
Imperfect information kets. My second contribution is to give structure to the information asymmetries associated with energy effi-
Adverse selection ciency by disentangling screening, signalling, moral hazard and price discrimination within a variety of
Principal-agent problems contractual relationships involving buyers and sellers, owners and renters, borrowers and lenders, and regulators
Home energy retrofit and policy stakeholders. I find evidence of information asymmetries to be compelling in utility-included rental
contracts, unclear in home sales and rentals, and scarce in retrofit contracting and financing. I conclude by
discussing the intricacies between informational and behavioural problems in energy-efficiency decisions.
© 2020 Elsevier B.V. All rights reserved.
JEL classification:
D82
D81
D86
H23
Q41

1. Introduction behaving in unexpected ways, a non-diligent installer, flawed engineer-

ing simulations, or simply bad luck with weather forecasts?
Energy-efficiency investments in residential and commercial build- Such a bewildering array of possible answers illustrates the
ings have uncertain returns. Long payback periods make them sensitive credence-good nature of energy efficiency (Sorrell, 2004). By definition,
to an array of contingencies. Their net present value depends on sto- the value of credence goods is never fully revealed to the buyer, even
chastic factors such as future energy prices and weather conditions. It long after purchase. Classical examples include medical treatments,
moreover depends on heterogeneous factors such as decision-makers' taxi rides or auto repairs. As illustrated above, energy efficiency shares
preferences (e.g., tolerance to cold, lighting habits) and constraint sets with these counterparts the following characteristics: sellers face het-
(e.g., physical properties of buildings, energy distribution infrastruc- erogeneous buyers; the quality of the product is not easily verifiable;
ture). To complicate matters further, many energy efficiency technolo- nor is it subject to complete liability rules. Altogether, these characteris-
gies require expert services, notably installation tasks, the quality of tics create a variety of information asymmetries, including adverse se-
which can be difficult to verify. Lastly, energy efficiency measures in- lection, moral hazard and price discrimination (Dulleck and
volve a number of stakeholders, all of whom having vested but not nec- Kerschbamer, 2006).
essarily aligned interests. This frequently includes, in addition to buyers Albeit pervasive in the economy, information asymmetries have
and sellers, users of energy-consuming assets and, as purchase prices specific implications in the context of energy efficiency. As we shall
can be substantial, credit suppliers. In this context, who's to blame if see in this essay, they can explain low uptake of energy efficiency, a
an insulation investment doesn't deliver as promised? The tenant long-standing paradox known as the energy efficiency gap (Jaffe and
Stavins, 1994). The problem has initially been identified through abnor-
mally high implicit discount rates in decisions to purchase energy-
efficient assets, suggesting that consumers discard supposedly profit-
E-mail address: [email protected]. able investment opportunities (Hausman, 1979; Train, 1985). More

https://doi.org/10.1016/j.eneco.2020.104698
0140-9883/© 2020 Elsevier B.V. All rights reserved.
2 L.-G. Giraudet / Energy Economics 87 (2020) 104698

recent studies document another manifestation of the energy-efficiency units. Evidence is growing of important information asymmetries in
gap, namely that energy savings measured after investment the supply of energy efficiency products (moral hazard, price discrimi-
underperform those predicted by engineering simulations before in- nation). More research is needed on information asymmetries in loan
vestment (Metcalf and Hassett, 1999; Fowlie et al., 2015). Three catego- markets, which are key to scaling up energy efficiency. Lastly, some in-
ries of economic problems are usually put forward to explain the formation asymmetries indirectly affect energy efficiency decisions by
energy-efficiency gap: market failures that truly impair socially desir- directly affecting the policies that target them. These have been mainly
able energy-efficiency investments; non-market failures that restrict in- studied in the context of subsidy programmes; much remains to be
vestment without affecting social welfare; and behavioural anomalies studied, in particular when it comes to energy performance contracts.
leading to individually irrational investment, with unclear implications The review focuses primarily, though not exclusively, on evidence
for social welfare (Gillingham et al., 2009; Allcott and Greenstone, gathered from revealed-preference studies conducted in the residential
2012; Gerarden et al., 2017). sector. It does not address in detail behavioural anomalies, a problem
Informational problems have been early pointed out as the main highly relevant to energy-efficiency decisions yet fairly well covered
cause of the energy-efficiency gap (Howarth and Andersson, 1993; elsewhere. Their relationships with information asymmetries, which
Huntington et al., 1994). As research has grown substantially since, are conceptually and empirically important, are nevertheless discussed
the contention can now be examined. I hereby take stock and review briefly at the end of the paper. Another discussion follows on what can
the vast yet piecemeal research into information in energy-efficiency be expected from rapidly developing information technologies in over-
decisions. A preliminary finding is that information problems are ill- coming barriers to energy efficiency.
characterized within the usual three-fold categorization. I sort this out The review proceeds as follows. Section 2 reviews symmetric-
by stressing the dichotomy between market and non-market failures, information problems. Section 3 introduces various types of information
which I restate as one between, respectively, asymmetric and symmet- asymmetries with the example of home energy retrofits. Section 4 de-
ric information. I thereby complement existing research that has fo- tails adverse-selection problems. Section 5 details principal-agent prob-
cused on behavioural anomalies as the main category of problems at lems. Section 6 discusses information problems that affect policy
the source of the energy-efficiency gap (Gillingham and Palmer, 2014; remedies to the energy efficiency gap. Section 7 puts the findings in per-
Allcott, 2016). My review is closest to Ramos et al. (2015), who also re- spective and Section 8 concludes.
view information problems, with the important difference that I place
less emphasis on behavioural problems and more emphasis on informa- 2. Symmetric-information problems
tion asymmetries. Another closely related review is that of Matisoff et al.
(2016), which covers some of the information asymmetries studied I define symmetric-information problems as information imperfec-
here, though within a broader framework encompassing non-energy tions or gaps identically faced by contracting parties. These are not mar-
impacts of building (e.g., health-related) and resource efficiency in ket failures in that no party extracts any informational rent from the
building design, construction, operation and deconstruction. Against other. Market outcomes can be improved by information technologies,
this background, my main contribution is to categorize problems that the development of which does not a priori require public support.2
are often referred to with inconsistent terminology1 and map the avail- Though symmetric-information problems are normal components of
able evidence on their significance. well-functioning markets, the trouble in energy-efficiency research is
My conclusions are two-fold. First, I find symmetric-information that they are often ignored when predicting economically efficient
problems to be important. This includes incomplete information – levels of energy efficiency. This consistently leads to overestimation of
e.g., infrequent billing of energy use, incomplete disclosure of product the energy-efficiency gap.
attributes, need for pre-retrofit audits – and imperfect information –
e.g., uncertainty about energy prices and weather conditions. These 2.1. Incomplete information
problems are frequently mistaken for information asymmetries or be-
havioural anomalies, which is a source of overestimation of the Incomplete information here denotes situations in which part of the
energy-efficiency gap. While technological progress and insurance mar- information needed to make a decision is missing, or costly to obtain
kets should suffice to overcome them, the effectiveness of these private with current technology. The problem can be identified by observing
solutions has not been examined. changes in adoption patterns when people are provided with more
Second, I find information asymmetries to be of a broader variety complete information about energy efficiency. As detailed below, in-
than previously thought. Assessment of their magnitude is however complete information affects several decision variables.
subject to methodological caveats. I disentangle screening, signalling,
moral hazard and price discrimination within a variety of contractual 2.1.1. Energy operating costs: evidence from infrequent billing
relationships involving buyers and sellers, owners and renters, bor- The cost of operating energy-consuming devices is usually not
rowers and lenders, and regulators and policy stakeholders. Information known in real-time.3 In the absence of consumption displays, which re-
asymmetries appear to be important in utility-included rental contracts main far from widespread, information is incomplete in at least two re-
(moral hazard, screening). Evidence is more mixed in building sales. spects: it is only provided occasionally when fuel tanks are filled or
Buyers are found to respond to energy efficiency, yet only a handful infrequently when electricity and natural-gas meters are monitored; it
analyses separate out the effect of energy-efficiency labels from that of covers a bundle of usages. How does more complete information affect
other observable energy-efficiency characteristics. Evidence is also in- market outcomes?
consistent in rental housing, where more energy efficient units seem A number of studies have evaluated experiments increasing the fre-
to rent with a premium, yet the market equilibrium is characterized quency of information through smart metering or in-home displays.
by a relatively poor energy efficiency, as compared to owner-occupied These so-called feedback interventions, extensively reviewed in
Abrahamse et al. (2005), Fischer (2008), Delmas et al. (2013), and
1
For instance, the notions of “landlord-tenant dilemma” and “split incentives” are com-
2
monly used to refer to any problem arising in rental housing. I characterize these problems Public support can be warranted if information technologies are subject to classical in-
into three categories: signalling in building rental, screening and moral hazard in utility- novation market failures. The question of whether this is the case in the context of energy
included contracts. Likewise, the notion of “credit constraint” is commonly used to refer efficiency is outside of the scope of this paper.
3
to any market failure affecting energy efficiency financing. I characterize these into screen- I focus here on information about energy quantities, which is more often missing than
ing and moral hazard in energy efficiency loans. Lastly, the notion of “free riding” is some- information about energy prices. Still, it is important to note that peak versus off-peak
times used in the context of subsidy programmes to refer to infra-marginal participants. I electricity prices and energy tariff menus are not always displayed transparently
stick to the latter designation. (e.g., Sexton et al., 1989).
 L.-G. Giraudet / Energy Economics 87 (2020) 104698 3

Buchanan et al. (2015), are generally implemented by electricity utili- Mandatory labels in place in the European Union and China are
ties. They initially produced mixed results. Abrahamse et al. (2005) framed within a discrete performance scale, thereby reconciling the ac-
find little impact based on 38 studies. Scepticism is shared by Buchanan curacy of EnergyGuide and the conciseness of Energy Star. Zhou and
et al. (2013) who even document cases where more frequent informa- Bukenya (2016) show in a discrete choice experiment that consumer's
tion increased energy use. Delmas et al. (2013) draw slightly more pos- mean willingness-to-pay for efficient air-conditioning systems in-
itive conclusions, estimating an average reduction in energy use of 7.4% creased when the performance was framed in a more segmented way.
from 156 studies. The effect is more pronounced at the high-utilization end of the distri-
More recent studies tend to confirm the negative yet modest effect bution. In the European Union, a similar experiment was conducted
of information frequency on energy use. Matsukawa (2004) finds a sig- by Andor et al. (2016), who find results similar to Houde (2018a),
nificant effect of electricity monitoring devices in a Japanese experi- namely that some people respond to information only, while others re-
ment. Houde et al. (2013) ran an experiment with 1500 employees spond to norms. In contrast, in an eye-tracking experiment, Waechter
from Google and found that participation in the feedback programme et al. (2015) find little impact of labels in decision-making.
yielded an average reduction in electricity use of 5.7%, persisting up to In case labels are not sufficient, sales agents can provide an addi-
four weeks. In a similar experiment involving 1500 Austrian house- tional information channel. A few studies have examined this conten-
holds, Schleich et al. (2013) find an average 4.5% reduction of electricity tion in field experiments. Anderson and Claxton (1982) found a
use attributable to getting feedback, however concentrated around the positive impact of sales staff support on label awareness, but no appar-
median of the distribution. Delmas and Lessem (2014), in an experi- ent impact on refrigerator choice in 18 department stores in Western
ment on UCLA campus find that real-time feedback was ineffective, Canada. Likewise, Kallbekken et al. (2013) find no statistical effect of
while publicly visible conservation ratings reduced electricity use by training of sales staff on the purchase of tumble driers and fridge-
20%, with more effect for above median energy users. Jessoe and freezer in six megastores in Norway. In a randomized controlled trial in-
Rapson (2014) find that informed households are three standard devia- volving 20000 agents in call centres of a large US retailer, Allcott and
tions more responsive to price variations than uninformed households Sweeney (2017) find that, unless combined with large rebates, informa-
and that this cannot be attributed to price salience. Sexton (2014) stud- tion and sales incentives alone have zero statistical effect on the sales of
ies the somewhat reverse experiment. The author finds that enrolment water heaters.
in automatic bill payment (which decreases the frequency with which
consumers receive information) increases electricity use by 6% to 7%. 2.1.3. Performance of tailored measures: evidence from energy audits
Chen et al. (2015) find evidence that consumers inaccurately estimate In large-scale projects such as home energy retrofits, which combine
energy use from appliances. Lastly, Tiefenbeck et al. (2016) finds a several measures and products within an idiosyncratic architectural lay-
large effect of 22% on showering. out, ex ante assessment of energy savings cannot be standardized. In-
Overall, the effect of more frequent information on energy use seems vestment appraisal requires customized audits which typically come
to be specific to individual preferences, to the point that rendering its at a cost of a few hundred dollars (Alberini and Towe, 2015; Palmer
sign ambiguous. It also seems to be more effective when targeting spe- et al., 2015).
cific energy services. Lastly, it vanishes when information frequency is Do audits produce accurate predictions? Available evidence points
reverted to normal. One difficulty for evaluation is that most experi- to a negative answer. The problem was first identified by Metcalf and
ments include other treatments such as tips or comparison with peers, Hassett (1999), who found that returns to insulation underperformed
which might confound identification of the purely informative effect. audit predictions. The result has recently been confirmed by other stud-
We will return to that point in Section 6. ies, such as Fowlie et al. (2015), Graff Zivin and Novan (2016) and
Giraudet et al. (2018). Graff Zivin and Novan (2016) find that 79% of
2.1.2. Performance of standardized products: Evidence from energy labels predicted savings are actually realized. Giraudet et al. (2018) find simi-
Next to information available while operating energy-consuming lar figures, on average, with ratios ranging from 31% to 352% depending
durables, information available at the time of purchase might also be in- on the measures considered. The discrepancies come from measure-
complete. For standardized products such as electrical appliances, infor- ment errors and complexities inherent in thermal simulation algo-
mation about product performance is generally produced by rithms (de Wilde, 2014; Hsu, 2014).5 They can also be due to market
normalized engineering calculations and displayed through labels. As- failures such as moral hazard, as we will see in Section 5.
suming that labels are trustworthy,4 how do consumers respond to The next question of interest is: how (possibly inaccurate) audits
the more complete information they convey? As we shall see, here too modify investment decisions? This can be directly assessed by observ-
it is difficult to disentangle information from behavioural effects. ing purchase behaviour. Early assessment of McDougall et al. (1982)
A few studies have examined the impact of the EnergyGuide label, a found little or no effect of audits on homeowner's conservation activi-
mandatory label implemented in 1979 in the United States reporting a ties. Frondel and Vance (2013), applying a mixed logit model in
cost figure based on average national usage and energy prices. Houde Germany find a mean positive effect, though with substantial heteroge-
(2018a) examines refrigerator purchases and finds that a fraction of neity, some people exhibiting negative responses to audits. Murphy
consumers respond to this piece of information in a privately rational (2014a) finds even more counter-intuitive results in the Netherlands,
way. Meanwhile, others over-respond, an effect the author attributes with a treated group not reacting to audits while non-treated individ-
to the coexisting Energy Star label, a voluntary label providing coarser uals make more energy-efficient investments. Palmer et al. (2013,
information. A third fraction of consumers do not respond to either 2015) find in a survey that the depth of an audit, as measured by the in-
label. Newell and Siikamäki (2014) also find that Energy Star leads to clusion of such items as energy bill assessment, blower door test or in-
cost-effective decisions, while over-reaction cannot be excluded. Davis frared imaging, is an important determinant of follow-up on audit
and Metcalf (2016) find a heterogeneous response to EnergyGuide in recommendations. Considine and Sapci (2016) estimate a significant
an online stated-choice experiment, with more relevant information but modest effect of audits on investment in a discrete-choice analysis
about local energy price leading to more rational decisions. of a programme conducted in Wyoming. In the commercial sector,
Anderson and Newell (2004), find that half of audits are followed up.
4 Comparable effects have been observed in Germany (Schleich, 2004)
At least two caveats apply here. First, the tests preceding label attribution could be
subject to falsification, just like the widely publicized Volkswagen case revealed in the au-
5
tomobile sector (U.S. EPA, 2015). To my knowledge, the issue has not been investigated in One source of error is the so-called “prebound effect” which arises when the baseline
appliances and other energy-consuming assets. Second, sellers can exploit labels to price- energy use against which savings are predicted is overestimated (Sunikka-Blank and
discriminate. We return to that point in Section 5.2. Galvin, 2012).
4 L.-G. Giraudet / Energy Economics 87 (2020) 104698

Fig. 1. Main stakeholders and contractual relationships in home energy retrofits.

and Sweden (Backlund and Thollander, 2015). In Italy, Barbetta et al. – for instance, whether or not to insulate – as opposed to discrete
(2015) find no effect of audits on either the number of investments or choices. In contrast, Ansar and Sparks (2009) follow Hassett and
the amounts invested in local public administrations. Metcalf's line and argue that incorporating technological change can
The effect of audits can also be assessed indirectly by examining var- produce high option values.
iation in energy use, under the assumption that it follows from unob- Whatever their size, option values, if unaccounted for, can be a
served investment. Taking this approach, Hirst and Goeltz (1985) source of overestimation of the energy efficiency gap. While energy per-
found that receiving a free audit induced significant but small energy formance contracts can provide some insurance against the problem,
savings. More recently, Alberini and Towe (2015) find that participating they are subject to some caveats that I discuss in Section 6.2.
in an audit programme yields 5% energy savings on average, an effect
commensurate with that estimated for rebates in the programme. 3. Asymmetric information: a framework
Altogether, it is difficult to disentangle the quality of audits from
their effect on investment. In addition, selection bias is an important Energy efficiency is subject to verifiability, liability and heterogene-
concern in the small-scale studies reviewed here. ity issues which together make the essence of credence goods and cre-
Results indicate that information relevant to energy-efficiency deci- ate information asymmetries – true market failures requiring public
sions is incomplete and that providing better information improves intervention (Dulleck and Kerschbamer, 2006). The problems are mag-
market outcomes. Yet the overall effect tends to be small and heteroge- nified by the high upfront costs and multiplicity of stakeholders in-
neous. The information gap is therefore probably modest. volved in energy-efficiency investments. To illustrate, let us consider a
measure that epitomizes these characteristics: home energy retrofit, in-
2.2. Imperfect information: option values cluding in particular insulation and improvements on weatherization
systems. As summarized in Fig. 1, the homeowner, who is central to
In addition to being incomplete, information about energy cost can the investment decision, may contract with four economic agents
be imperfect, in the sense that it bears some randomness. Energy prices (some of whom might be herself): a tenant whose utility bill may or
are volatile in the short to medium term; energy needs, in turn, are de- may not be included in the rent; a contractor selling and installing dura-
termined by intrinsically random factors such as the weather. Combined ble goods; a credit supplier; a subsequent buyer of the retrofitted
with the irreversible nature of energy efficiency improvements, such home.6
randomness creates option values (Dixit and Pindyck, 1994). These af- Each of these relationships can be subject to a variety of information
fect investment outcomes if decision-makers are risk-averse, which asymmetries. On top of these relationships, some stakeholders – owners
seems to be a valid assumption in the context of energy efficiency and sellers in particular – can be involved in a relationship with the reg-
(Farsi, 2010). Using calibrated simulations, Hassett and Metcalf (1993) ulator, or more generally a programme administrator, when a policy is
pointed out early that option values alone could fully explain the high implemented. I review below the evidence on the significance of infor-
hurdle rates observed in energy-efficiency decisions. Sanstad et al. mation asymmetries in all of these relationships. I use the standard ter-
(1995) objected that this was only valid for a narrow range of decisions minology of Mas-Colell et al. (1995), who classify information
in which delay is not costly – unlike, say, window replacement, which is
more expensive alone than if included in an earlier retrofit. Baker 6
For lack of empirical evidence, I do not include other actors who too can engage in
(2012) further restricts Hassett and Metcalf's result to binary decisions principal-agent relationships, such as energy suppliers, building certifiers and sales agents.
 L.-G. Giraudet / Energy Economics 87 (2020) 104698 5

asymmetries in two broad categories – adverse selection and principal- The hypothesis started being examined in the early 1980s. At the
agent problems – each encompassing subcategories – screening and sig- time, energy efficiency was measured by past billing data or coarse la-
nalling on the one hand, moral hazard and price discrimination on the bels describing the thermal integrity of the unit. Hedonic analyses
other. found evidence of capitalization of energy efficiency into home sale
For each relationship, I formulate a null hypothesis corresponding to prices (Johnson and Kaserman, 1983; Dinan and Miranowski, 1989;
the absence of any information asymmetry. I examine the validity of Laquatra, 1986; Nevin and Watson, 1998). This early literature was
each hypothesis and, if rejected, assess the significance of the implied however criticized for failing to appropriately take into account the
information asymmetry. The findings are summarized in a matrix fragmented and local nature of housing markets and the difficulties as-
interacting four types of information asymmetry and four types of rela- sociated with measuring costs and benefits of energy efficiency in hous-
tionship (Table 1). Further tables provide methodological detail at the ing (Laquatra et al., 2002).
study level. The topic attracted renewed interest in the early 2010s with the ad-
vent of energy performance certificates such as that promoted by the
4. Adverse selection European directive (hereafter EU-EPC) and the LEED and Energy Star la-
bels in the United States. Larger datasets and more modern hedonic
Adverse selection occurs when part of the relevant information is methods permitted more credible identification. Studying commercial
hidden to one party. Specifically, screening is imperfect whenever the buildings shortly after energy-efficiency labels became mandatory,
seller cannot observe buyers' types and signalling is imperfect when- Fuerst and MacAllister found that labelled buildings carried a price pre-
ever the seller is unable to convey the quality of its products to prospec- mium in the United States (Fuerst and McAllister, 2011a) but not in the
tive buyers. Either problem results in too little quality in the market – European Union (Fuerst and McAllister, 2011b). Brounen and Kok
products that are often referred to as ‘lemons.’ (2011) identified a price premium associated with the EU-EPC in the
Netherlands. Murphy (2014b) nuances the finding by surveying pur-
4.1. Imperfect signalling chasers, arguing that the EPC had little influence in sales negotiation.
Kahn and Kok (2014) find a premium associated with LEED, Energy
4.1.1. Building sales Star and other “green” labels in housing California. Hyland et al.
Perhaps the longest-studied information problem associated with (2013) and Stanley et al. (2015) find a similar premium in Ireland and
energy efficiency has to do with signalling in building sales. If the energy Dublin, respectively. Harjunen and Liski (2014) find that more efficient
efficiency of a building unit is perfectly observable, we expect the heating technologies such as electric and district heating are capitalized
following: in the Finnish housing market. Fuerst et al. (2015) find a significant ef-
fect of the EU-EPC in England. Myers (2019) finds evidence that changes
H1: Energy efficiency is capitalized into home prices
in relative fuel prices cause changes in relative housing prices in

Table 1
Summary of the evidence on information asymmetries in home energy retrofits.

BUYER/SELLER OWNER/USER BORROWER/LENDER REGULATOR/STAKEHOLDER

Adverse Imperfect H3: The rent premium for H4: Energy efficiency H9: Regulators optimally screen
selection screening including utilities covers effective investments carry lower out infra-marginal participants in
energy usage interest rates than do subsidy programmes
conventional ones
Rejected in 100% of the material Rejected in at least 90% of the
reviewed (3 studies, §4.2.1) Mixed: rejected in 50% of the material reviewed (10 studies,
material reviewed (2 studies, §6.1)
§4.2.2)

Imperfect H1: Energy efficiency is capitalized into H2: Energy efficiency is capitalized
signalling home prices into rents

Not rejected in 83% of the material Not rejected in 88% of the

reviewed (24 estimates out of 29). material reviewed (7 studies out
of 8).
Additional finding: 6 studies suggest
this is the case even in the absence of H2bis: Owner-occupied dwellings
energy efficiency labels are as energy efficient as are
rented ones
(§ 4.1.1)
Rejected in 100% of the material
reviewed (6 studies).

(§ 4.1.2)

Principal-agent Moral hazard H6: Contractors put identical effort into H5: Occupants use as much H7: Borrowers default less H10: Stakeholders do not respond
problems observable and unobservable energy energy under utility-included when borrowing to save energy to misplaced incentives stemming
efficiency tasks contracts as under individual Not rejected in 100% of the from poor policy design
billing material reviewed (2 studies, Rejected in one study
Partially rejected in one study finding §5.1.3) examining the role of auditors
evidence of moral hazard 20% of the Rejected in 100% of the material Qualification: studies subject in a subsidy programme (§6.1)
time (specifically on Fridays). reviewed (6 studies, §5.1.1) to selection issues

(§5.1.2.)

Price H8: Producers offer consumers their

discrimination optimal level of energy efficiency
Rejected in 100% of the material
reviewed (3 studies, §5.2)
6 L.-G. Giraudet / Energy Economics 87 (2020) 104698

Massachussetts in a way that is consistent with full capitalization of en- should be no difference in energy efficiency between owner-occupied
ergy savings. Lastly, Wahlström (2016) finds evidence of capitalization dwellings and rented ones. In other words, H2 implies the following
of the EU-EPC in Sweden. Like the responses discussed in corollary:
Section 2.1.2, capitalization sometimes exceed the present value of en-
ergy savings. This is the case in US office buildings (Eichholtz et al., H2bis: Owner-occupied dwellings are as energy efficient as are rented
2010, 2012) and homes in three US cities (Walls et al., 2017). ones
These studies together provide compelling evidence of full capitali-
zation of energy-efficiency labels. They are less conclusive, however, Research along this line tends to reject H2bis. Brechling and Smith
as to whether labels effectively fill an information gap. After all, the (1994) find lower ownership of energy-efficient assets in rented prop-
early studies reviewed above, despite their shortcomings, suggested erties as compared to owner-occupied ones in the United Kingdom.
that capitalization of energy efficiency was already effective prior to im- Scott (1997) finds similar results in Ireland. Davis (2012), using the
plementation of any label. Modern evaluations of labelling policies, in U.S. Residential Energy Consumption Survey (RECS), documents that
turn, do not compare situations with and without labels, which is the renters are significantly less likely to report having energy-efficient ap-
only way to determine whether labels operate by levelling the informa- pliances such as refrigerators, clothes washers and dishwashers.
tion shared by the buyer and the seller – thereby eliminating an infor- Gillingham et al. (2012), using the same database, report that owner-
mation asymmetry – or simply by restating information decision- occupied dwellings in California are 20% more likely to be insulated in
makers can anyway gather from observable features. the attic or ceiling than rented ones. Melvin (2018) extends the result
The most recent studies on the topic are beginning to fill this gap. to water heating, window thickness and weatherization. Myers (2015)
Exploiting a panel dataset in which some dwellings were sold multiple finds that energy price movements cause shifts in rents of energy-
times in Oslo, Norway, Olaussen et al. (2017) find that current EPCs ex- efficient units when rents include utilities, but not otherwise, suggest-
plain sale prices in a way consistent with the studies discussed above, ing the market does not convey information about energy use. In
but also explain the prices of transactions that occurred before imple- Europe, Krishnamurthy and Kriström (2015) report that owners are
mentation of the EU-EPC policy. Furthermore, the authors find no evi- more likely to have energy-efficient appliances, better insulation and
dence of a price premium after controlling for dwelling fixed effects. heat thermostats than tenants.
These results suggest that labels provide no information beyond what The evidence here is therefore paradoxical: Signalling issues do not
is observable without them. Similar conclusions are reached by seem to significantly affect rents, yet the resulting equilibrium is subop-
Fesselmeyer (2018) by exploiting price variation before and after certi- timal. This apparent inconsistency is arguably due to differences in
fication in Singapore.7 On a related issue, Frondel et al. (2020) exploit a methodological approaches to each fact. One way to reconcile these
shift from voluntary to mandatory disclosure of the EU-EPC in Germany findings would be to examine how the elasticity of rents paid by tenants
and find that it causes a contraction of the energy efficiency premium compares to that of implicit rents borne by owner-occupiers, both with
for owners who would not voluntarily disclose. This can be interpreted respect to energy efficiency. A lower elasticity (in absolute value) for
as evidence that sellers of low-efficiency dwellings did not voluntarily tenant's rents would reveal an inability of landlords to fully pass
engage in signalling. through investment costs onto rents, perhaps due to rigid rent regula-
To conclude, existing evidence fails to reject H1 and indicates that tions. Lastly, while labels seem to improve decisions, evidence is scarcer
the role of energy performance certificates in improving information as to whether they encourage landlords to implement energy-efficiency
is modest at best. Energy performance certificates may nevertheless improvements.
have other effects – behavioural in particular – which remain to be
estimated. 4.2. Imperfect screening

4.1.2. Building rental 4.2.1. Utility-included rental contracts

The question of capitalization similarly applies in rental markets, In many countries, rental contracts frequently include energy oper-
where by default we expect the following: ating costs. In the United States, for instance, approximately 60% of
housing rental contract include at least one energy or water utility
H2: Energy efficiency is capitalized into rents (Choi and Kim, 2012). How does a landlord offering utility-included
contracts adjust rents to the tenant's specific energy usage? Under per-
Existing studies tend to confirm H2 – for instance Fuerst and fect screening, we expect the following:
McAllister (2011a, 2011b) and Eichholtz et al. (2010, 2012) in US office
buildings, Kok and Jennen (2012) in commercial buildings in the H3: The rent premium for including utilities covers effective energy
Netherlands, Hyland et al. (2013) in the Irish residential sector. usage
Reichardt (2014) finds rent premia that exceed the value of savings on
operating expenses in the United States. Like in building sales, these The hypothesis has been relatively little-studied. Levinson and
studies are limited in their ability to disentangle the purely informative Niemann (2004), using RECS and the American Housing Survey
effect of labels from other potential effects. Indeed, Bala et al. (2014) (AHS), find that rents in utility-included rental apartments are higher
find that rents in Brussels in 2001 were positively correlated with the than in comparable metered apartments, but the difference is smaller
presence of observable features such as double glazing and wall insula- than the difference in energy operating costs observed between the
tion, which suggests that energy efficiency was already capitalized two types of apartments. This can be interpreted as a failure of the land-
without labels. In ongoing work, Dressler and Cornago (2018) address lord to screen tenants with high-intensity energy usage. Myers (2015),
this methodological gap by exploiting a shift from voluntary to manda- similarly using the ASH and exploiting variation in energy prices finds
tory certification in Brussels similar to that exploited by cf. infra Frondel that low-efficiency dwellings turn over faster than high-efficiency
et al. (2020) in housing sales. Their results provide suggestive evidence ones when tenants pay for energy, but not when utilities are included
of strategic non-compliance with mandatory disclosure in those units, in the rent. Again, this suggests that tenants are less likely to self-
the EPC of which is below average. select into the dwelling that best fits their preferences when they do
The evidence here suggests that energy efficiency is accurately sig- not pay the marginal cost of energy.
nalled in rental markets. Yet if it were the case, in equilibrium there These results together suggest that H3 is rejected, with utility-
included contracts favouring tenants with intensive energy usage and
7
Similar results are obtained in the Korean market for televisions (Park, 2017). pricing others out of the market. One way to address this market failure
 L.-G. Giraudet / Energy Economics 87 (2020) 104698 7

could be to ban such contracts – subject to some caveats I discuss in slightly more energy under such contracts. Maruejols and Young
Section 5.1.1. (2011) find similar effects in Canada. Gillingham et al. (2012) similarly
find that under such contracts, occupants in California are 16% more
4.2.2. Energy-efficiency loans likely to change the heating thermostat at night. Kahn et al. (2014)
In theory, adopting energy efficiency saves consumers money, find evidence of a better environmental performance in those commer-
thereby increasing their creditworthiness and reducing default risk. In cial buildings, the tenants of which face a positive marginal cost for elec-
a well-functioning credit market, we therefore expect the following: tricity. Myers (2015) finds that landlords are more likely to make cost
saving investments when they face the marginal cost of energy usage.
H4: Energy efficiency investments carry lower interest rates than do The most credible evidence to date is provided by Elinder et al. (2017)
conventional ones who compare energy use up to four years before and after an interven-
tion consisting in excluding utilities from rental contracts in Sweden.
Investigating this hypothesis in commercial mortgages in the United Compared to 1000 tenants in the control group, the 800 treated tenants
States, An and Pivo (2017) find better loan terms for buildings that were showed an immediate and permanent reduction in energy use by 25%.
certified green at loan origination than for other buildings which were Evidence of moral hazard in utility-included contracts is therefore
either non-green or certified green after loan origination. Though mod- compelling. All authors however underline that the effect is small in
est in magnitude, the effect is consistent with lenders efficiently using terms of excess energy use – which does not mean that welfare effects
green labels as a screening device. The analysis is however threatened are unimportant. Here again, banning utility-included contracts could
by selection issues, as the authors do not control for borrower avoid over-use of energy, a problem even more critical in the presence
characteristics. of uninternalized energy-use externalities. This recommendation is
Studying personal loans, which is the most commonly adopted tool however subject to some caveats. First, the benefits from the ban
to finance home energy retrofits, Giraudet et al. (2019) reach opposite must be weighed against the cost of installing individual meters. Sec-
findings in France. Using a unique dataset of posted interest rates col- ond, they must also be weighed against forgone benefits, as utility-
lected from credit institutions' websites, the authors circumvent the se- included contracts can enable landlords to attract certain tenants
lection issues faced by An and Pivo (2017) in that their data do not whose profile they particularly value (Choi and Kim, 2012). Third, ban-
depend on either borrower or loan contract characteristics. They find ning utility-included contracts can give rise to heating externalities in
that, on average, home retrofits are priced at a higher rate than other- multi-family housing. Indeed, because of heat transfers within build-
wise conventional investment (automobiles in particular), and even ings, especially along the vertical gradient, some occupants may ‘free
more so when home retrofits include energy efficiency improvements. ride’ by turning down their thermostat as they get heat input from an
A tentative explanation is that lenders use the loan project as a screen- adjacent dwelling; this in turn induces their neighbor to turn their ther-
ing device of the borrower's wealth in an attempt to extract surplus. mostat up in order to enjoy their desired level of comfort. By reducing
This can cause credit rationing and therefore too little investment in the marginal cost of changing thermostat settings to zero, utility-
home energy retrofits. included contracts remove these incentives. To the best of my knowl-
The evidence on whether screening problems affect energy effi- edge, however, such externalities have never been discussed nor
ciency loans is therefore mixed. quantified.

5. Principal-agent problems
5.1.2. Building retrofits
Principal-agent problems arise in situations where a principal hires An important relationship in the energy retrofit picture is that be-
an agent to perform a task. Moral hazard arises if the principal cannot tween a contractor – the agent – and an investor – the principal. In prac-
observe the agent's ex post actions. Price discrimination arises if a mul- tice, the quality of such retrofit works as attic insulation or duct sealing
tiproduct monopolist cannot observe the agents' types ex ante. Both cat- is hard to verify by non-experts, unless costly ex post audits involving
egories produce undesirable behaviours and they are likely to distort thermo-photography or blower-door tests are commissioned. The in-
the markets for energy efficiency. formational context is conducive to moral hazard in the form of
under-provision of quality in the installation tasks performed by con-
5.1. Moral hazard tractors. In the absence of the problem, we expect the following:

5.1.1. Utility-included rental contracts H6: Contractors put identical effort into observable and unobservable
Utility-included contracts can be seen as a relationship between a energy efficiency tasks
principal – the utility – and an agent – the occupant. This relationship
can be affected by incentive problems, which add up to the selection is-
sues studied in Section 4.2.1. If utilities could perfectly monitor occu- Using data from a utility-sponsored retrofit programme in Florida,
pants' behaviour and charge them accordingly, we would expect the Giraudet et al. (2018) find that energy-efficiency measures are subject
following: to day-of-the-week effects if they are deemed hard-to-observe, but
not otherwise. The day-of-the-week effect follows a specific pattern –
H5: Occupants use as much energy under utility-included contracts as energy savings are lower when the retrofit was completed on a Friday,
under individual billing as compared to other days of the week. The authors find that the prob-
lem can explain 65% of the discrepancy observed between predicted
Yet in practice, occupants' behaviour is hard to observe, either be- and realized savings. Interestingly, the fact that quality is underprovided
cause it cannot be monitored in real time or because it is pooled with on certain, but not all, days of the week suggests that some incentives
others' behaviour under shared billing in multi-family housing. Such sit- exist to provide quality, perhaps in the form of reputational returns.
uations are conducive to moral hazard and thus rejection of H5: just like Based on the only existing study, I conclude that evidence on H6 is
an insuree is expected to take little care of a product covered by an in- mixed.
surance contract, an energy user who does not face the marginal cost Moral hazard can be addressed by professional certifications – a
of energy is expected to over-use energy. public solution – or energy-savings insurance – a private one. While
H5 is rejected in all studies I am aware of that investigate it. Using the former incurs monitoring costs, the latter raises interesting incen-
RECS data, Levinson and Niemann (2004) find that US households use tive problems that are further discussed in Section 6.2.
8 L.-G. Giraudet / Energy Economics 87 (2020) 104698

5.1.3. Energy-efficiency loans water industry.9 Second, when energy efficiency investments are
As stated earlier, energy efficiency is supposed to reduce default risk targeted by public subsidy programmes, the regulator can address the
in associated loans (where the borrower can be seen as the agent and distortions due to price discrimination by carefully adjusting subsidy re-
the lender the principal). In well-functioning credit markets, we thus gimes, for instance by offering more generous subsidies for low-end
expect the following: products than for high-end ones (Nauleau et al., 2015). Despite restor-
ing efficiency in the provision of quality, this approach raise equity con-
H7: Borrowers default less when borrowing to save energy
cerns in that it compensates energy efficiency suppliers for the
deadweight loss they are causing.
Using US data from the Home Energy Rating System (HERS), Kaza
et al. (2014) found that more energy efficiency, as measured by 6. Policy-induced information asymmetries
ENERGY STAR ratings, is associated with lower default and prepayment
rates in residential mortgages. Applying a similar research design to So far I have examined information problems that directly affect en-
commercial mortgages, An and Pivo (2017) confirm that greener build- ergy efficiency decisions, assuming away further policy-induced effects.
ings are associated with lower default rates. The effect is more impor- In practice, however, energy efficiency decisions are targeted by an
tant than that identified by the authors in relation to loan terms (cf. array of policies meant to correct the multiple market failures they are
Section 4.2.2). Altogether, these results confirm H7, which can be associated with – chiefly energy-use externalities, but also the very in-
interpreted as efficient loan pricing, implying that information formation asymmetries that are at the core of this paper. These policies
asymmetries in energy-efficiency loans are not economically important. may in turn create information asymmetries between the regulator and
In home energy retrofits, an additional problem arises. Unlike other some stakeholders – programme participants and intermediaries such
assets of comparable purchase price, say a car, an energy retrofit cannot as retrofit contractors in particular. Let us review here the evidence on
be confiscated. Therefore, unless the retrofit is included in a mortgage, it how these information asymmetries affect policy effectiveness.
cannot serve as credit collateral. This might lead lenders to raise interest
rates in an effort to hedge against increased default risk (Palmer et al., 6.1. Imperfect screening in subsidy programmes
2012). The effect has not yet been empirically investigated.
The most common policy tool to promote energy efficiency is sub-
5.2. Price discrimination sidy programmes. Primarily motivated by reducing energy-use exter-
nalities, these programmes are either directly administered by
Energy efficiency is essentially a characteristic of energy-using assets governments (e.g., tax credits, zero-interest loans) or rolled out by elec-
and, as such, it can be provided in various degrees. In other words, the tric and gas utilities to comply with so-called energy saving obligations
seller of energy-using products typically sells a menu of products of dif- or demand-side management programmes.10
fering energy efficiency. If such markets function well, we expect the A regulator willing to minimize the cost of subsidy programmes
following: would proceed as follows:
H8: Producers offer consumers their optimal level of energy efficiency
H9: Regulators optimally screen out infra-marginal participants in sub-
sidy programmes
This hypothesis can however be rejected if price discrimination –
also known as monopolistic screening – occurs. This is typically the
In practice, programme administrators cannot observe the partici-
case in the presence of two market failures: imperfect competition
pants' willingness to invest without the subsidy; as a result, they may
and adverse selection. A multiproduct seller having market power but
give subsidies to households who would have invested anyway. Note
no ability to screen consumer's types has an incentive to under-
that the problem affects all subsidy programmes and has little to do
provide quality at the low-end of the product line so as to maintain
with the credence good nature of energy efficiency. Yet the evidence
high mark-ups on the sales of high-end products (Mussa and Rosen,
is compelling that it is significant in this context. With varying geo-
1978). If energy efficiency is the relevant dimension of quality, those
graphical scope and methodology, existing studies point to infra-
distortions result in too little energy efficiency at the bottom of the
marginal (or non-additional) participants typically accounting for 50%,
product line (Fischer, 2005; Nauleau et al., 2015).8 Houde (2018b) ex-
and not infrequently up to 90%, of total participants (Joskow and
ploits changes in the ENERGY STAR label in the US market for refrigera-
Marron, 1992; Hassett and Metcalf, 1995; Malm, 1996; Grösche and
tors and finds adjustments in the product line that are consistent with
Vance, 2009; Boomhower and Davis, 2014; Nauleau, 2014; Alberini
price discrimination. Spurlock (2013), exploiting simultaneous changes
et al., 2016; Rivers and Shiell, 2016; Houde and Aldy, 2017; Collins
in minimum energy efficiency standards and ENERGY STAR, reaches the
and Curtis, 2018).11 The welfare consequences of the problem are not
same conclusion for clothes washers. So do Cohen et al. (2017) using
clear when different options of varying energy efficiency are eligible in
variation in energy prices in the UK market for refrigerators. The scarce
the programme. Then, infra-marginal participants can opt for a higher
available evidence therefore suggests that H8 is rejected.
efficiency option than that they would have otherwise purchased. Al-
While there is no direct way to solve the problem, partial solutions
though the impact of subsidies has been less studied on the intensive
exist. First and foremost, anti-trust remains the main policy tool to
margin of investment than on the extensive one, preliminary evidence
limit the deadweight loss of imperfect competition. In this regard, it is
suggests it is small on the former (Rivers and Shiell, 2016; Houde and
worth mentioning that the French Anti-trust authority ruled against
Aldy, 2017). Overall, H9 is clearly rejected and screening issues
high concentration levels in the heating, air conditioning and hot
9
Conseil de la Concurrence, 2006. Décision no. 06-D-03 bis du 9 Mars 2006 Relative à
des Pratiques Mises en Oeuvre Dans le Secteur des Appareils de Chauffage, Sanitaires,
8
Improving energy efficiency normally means minimizing energy use while achieving a Plomberie, Climatisation.
10
given level of energy services. Yet the term is frequently used in the broader sense of sim- The relationship between the regulator and utilities in the context of energy saving
ply minimizing energy use, without necessarily holding energy service constant. This is obligations also raises interesting incentive problems (Lewis and Sappington, 1992; Wirl,
typically the case in transportation, where a small car is regarded as more energy- 1995) that are out of the scope of this paper.
11
efficient than a larger car. While a small car indeed allows one to cover more distance with The problem is generally difficult to empirically quantify, for lack of a counterfactual
the same amount of fuel, it also offers fewer services (e.g., limited capacity and comfort). If benchmark without subsidies. In this regard, the most credible approaches are those based
price discrimination operates along these other dimensions of energy services, it can lead on difference-in-differences (Alberini et al., 2016; Houde and Aldy, 2017) and regression
to too small cars, which, if energy efficiency is used in the broader sense, can be interpreted discontinuity (Boomhower and Davis, 2014) designs. The fact that less credible ap-
as an excess of it (Plourde and Bardis, 1999). proaches produce similar estimates confirms the significance of the screening problem.
Table 2
Evidence on imperfect signalling in building sales (contains 29 estimates from 21 studies).

H1 Energy efficiency is capitalized into home prices

Reference Geographical Sectoral scope Fuel type Energy Sample size Method Model Time period Hypothesis Certificate Comments
scope usage validated programme

Brounen and The Commercial and NA Heating 31,993 Heckman January Yes EU-EPC
Kok (2011) Netherlands Residential: two-step 2008–August
Apartment, estimation 2009
Detached, Duplex,
Semi-detached
Dinan and Des Moines, Residential: Natural gas, Heating 234 Hedonic January – Yes
Miranowski Iowa, US Detached dwellings other model June 1982
(1989)
Eichholtz US Commercial: Office NA NA 5993 GLS 2009 Yes Energy Star,
et al. (2012) buildings LEED
Eichholtz US Commercial: Office NA NA 199 in Diff-in-diff 2004–2007 Yes Energy Star, Green dwellings display a 16% sales premium.
et al. (2010) buildings treatment LEED
group, 1614
in control
group

L.-G. Giraudet / Energy Economics 87 (2020) 104698

Fesselmeyer Singapore Residential: NA NA 119,826 Diff-in-diff 2005–2016 Yes Green Mark The price premium is not caused by the energy label.
(2018) Apartments Scheme
Frondel et al. Germany Residential: NA NA 412,637 Diff-in-diff Fixed effects 2013–2015 Yes EU-EPC Upon disclosure of energy efficiency information,
(2020) Single-family and low-quality dwellings experience the largest price
multi-family reductions.
dwellings
Fuerst and US Commercial: Office NA NA 6157 Hedonic 1999–2008 Yes Energy Star, Sales premium of 25% and 26%, for LEED and Energy Star
McAllister buildings model LEED certified dwellings, respectively.
(2011a)
Fuerst and UK Commercial: Retail, NA NA 606 Hedonic April 2011 No EU-EPC Small sample size
McAllister office, industrial model
(2011b)
Fuerst et al. UK Residential: NA NA 16,653 Hedonic 1995–2012 No EU-EPC Rural, small sample, R sq. = 0.47
(2015) Detached dwellings model
in sparsely
populated areas
Fuerst et al. UK Residential: NA NA 68,354 Hedonic 1995–2012 Yes EU-EPC Densely populated areas
(2015) Detached dwellings model
in densely
populated areas
Fuerst et al. UK Residential: NA NA 106,793 Hedonic 1995–2012 Yes EU-EPC
(2015) Semi-detached model
dwellings
Fuerst et al. UK Residential: NA NA 115,658 Hedonic 1995–2012 Yes EU-EPC
(2015) Terraced dwellings model
Fuerst et al. UK Residential: NA NA 25,637 Hedonic 1995–2012 Yes EU-EPC
(2015) Apartments model
Harjunen and Finland Residential: District heating Heating 1868 2001–2012 Yes More energy efficient heating technologies are capitalized
Liski (2014) Single-family in house prices (district heating as opposed to electricity)
detached dwellings
Hyland et al. Ireland Residential: NA NA 15,060 Heckman January Yes BER Sales premium for A-rated dwellings is 9%.
(2013) Detached, selection 2008–March
semi-detached, 2012
terraced, apartment
Johnson and Tennessee, US Residential: Electricity, Heating 1317 IV, 2SLS Hedonic 1978 Yes
Kaserman Detached dwellings natural gas model

(continued on next page)

9
 10
Table 2 (continued)

H1 Energy efficiency is capitalized into home prices

Reference Geographical Sectoral scope Fuel type Energy Sample size Method Model Time period Hypothesis Certificate Comments
scope usage validated programme

(1983)
Kahn and Kok US Residential: NA Cooling 4321 in Hedonic 2007–2012 Yes Energy Star,
(2014) Single-family treatment model LEED,
dwellings group; GreenPoint
1,600,558 in
control
group
Laquatra Minnesota, US Residential: NA NA 81 IV Hedonicprice 1981–1982 Yes Low sample size.
(1986) Detached dwellings index Data subsequently used by Gilmer (1989) to calibrate a
search model
Murphy The Residential: NA NA 297 in Diff-in-diff, No EU-EPC Based on surveys, pre and post-purchase. The label had
(2014b) Netherlands Apartment, treatment Pearson's little impact in price negotiations on house sale whereas a
detached, 2 under 1 group; 1027 chi-squared third of the label owners stated that they applied for EPC
roof, corner, in control in order to increase the market value of their property.
terraced, other group
Myers (2019) Massachusetts, Residential: Heating oil, Heating 528,642 IV, 2SLS 1990–2011 Yes Relative fuel price has an effect on relative transaction

L.-G. Giraudet / Energy Economics 87 (2020) 104698

US Single-family natural gas price
dwellings
Nevin and US Residential: Electricity, piped Heating 14,400* IV Hedonic 1991–1996 Yes Laquatra et al. (2002) object that according to Nevin and
Watson Single-family gas model Watson's data, house values decrease with dropping
(1998) detached and utility bills
attached dwellings
Nevin and US Residential: Fuel oil Heating * IV Hedonic 1994–1996 Yes Laquatra et al. (2002) object that according to Nevin and
Watson Single-family model Watson's data, house values decrease with dropping
(1998) attached and utility bills.
detached dwellings
Nevin and US Residential: Fuel oil Heating * IV Hedonic 1991–1993 No Small sample size. Due to a sharp increase in fuel oil prices
Watson Single-family model following the invasion of Kuwait, a significant number of
(1998) detached dwellings households shifted to alternative heating fuels. The
authors suggest that the positive relation between the
heating expenditure and home values in that period could
be due to a large proportion of households reporting high
fuel expenditures following the spike, and property
values that reflect the future value after the planned
energy conversion works.
Olaussen et al. Oslo, Norway Residential: NA NA 4674 Fixed effects 2000–2014 No EU-EPC There price premium is not caused by the energy label.
(2017) Single-family
houses,
Apartments,
Townhouses,
Semi-detached
houses
Stanley et al. Dublin area, Residential: NA NA 2792 Hedonic January Yes BER
(2016) Ireland Apartments, regression 2009–June
Detached, Terraced, 2014
End-of-terrace,
Semi-detached
Wahlström Sweden Residential: Electricity, Heating, 77,000 OLS Hedonic 2009–2010 Yes EU-EPC The results indicate a strong willingness to pay for
(2016) Single-family district heating, cooling model housing attributes that reduce energy consumption. The
dwellings heat pump, EPC itself is not a variable in the regression.
biofuels,
oil-fired boiler,
other
 L.-G. Giraudet / Energy Economics 87 (2020) 104698
Walls et al. Austin, US Residential: NA NA 42,600 Hedonic 2008–2011 Yes Energy Star,
(2017) Single-family regression, local
dwellings multiple certifications
matching
procedures
Walls et al. Portland, US Residential: NA NA 117,828 Hedonic 2005–2011 Yes Energy Star, In Portland, the premium is exceeding the actual energy
(2017) Single-family regression, local saving.
dwellings multiple certifications
matching
procedures
Walls et al. the Research Residential: NA NA 16,041 Hedonic 2009–2011 Yes Energy Star,
(2017) Triangle region Single-family regression, local
in North dwellings multiple certifications
Carolina, US matching
procedures

11
12 L.-G. Giraudet / Energy Economics 87 (2020) 104698

Table 3
Evidence on imperfect signalling in building rental – rent premium (9 studies).

H2 Energy efficiency is capitalized into rents

Reference Geographical Sectoral scope Fuel Energy Sample Method Model Time period Hypothesis Certificate Comments
scope type usage size validated programme

Bala et al. Brussels Residential NA NA 507,868 Two-step estimation Hedonic 2001 Yes
(2014) region, procedure with model
Belgium locational fixed
effects
Dressler and Brussels Residential NA NA 13,586 IV Hedonic January Yes EU-EPC Owners with less
Cornago region, model 2010 – energy-efficient
(2018) Belgium September EPC ratings tend to
2014 hide this
information
Eichholtz US Commercial: NA NA 20,801 GLS Energy 2009 Yes
et al. Office buildings Star, LEED
(2013)
Eichholtz US Commercial: NA NA 7920 OLS Energy 2004–2007 Yes Green dwellings
et al. Office buildings Star, LEED display a 3% rental
(2010) premium.
Fuerst and US Commercial: NA NA 10,970 Hedonic 1999–2008 Yes Energy Rental premium of
McAllister Office buildings model Star, LEED 5% and 4%, for LEED
(2011a) and Energy Star
certified dwellings,
respectively.
Fuerst and UK Commercial NA NA 606 Hedonic April 2011 No EU-EPC Dependent
McAllister buildings: model variable: Market
(2011b) Retail, office, rent; Small sample
industrial size
Hyland et al. Ireland Residential: NA NA 20,825 Heckman selection January Yes BER Rent premium for
(2013) Detached, 2008–March A-rated dwellings
Semi-detached, 2012 is approximately
Terraced, 2%.
Apartment
Kok and the Commercial: NA NA 1100 Hedonic 2005–2010 Yes EU-EPC D to G rated office
Jennen Netherlands Office buildings model buildings receive
(2012) 6.5% lower rent
than A to C rated
buildings.
Reichardt Central and Office buildings NA NA 4217 Propensity-weighted Log-linear Yes Energy
(2014) Eastern US regression hedonic Star, LEED
model

therefore seem to significantly raise the public cost of energy efficiency Blonz (2019) identifies moral hazard by exploiting variation between
subsidy programmes. two forms of contract: one in which each task is performed by a differ-
Solutions start being discussed to address the problem. Globus- ent contractor and another one in which both are performed by one and
Harris (2018) proposes to introduce a waiting period between applica- the same contractor. The author finds that the latter contract induces
tion to the programme and investment. By exploiting the fact that, un- contractors to misreport non-eligible measures as eligible in an attempt
like marginal participants, infra-marginal ones are sensitive to time to reap the reward from subsequently installing it. The problem seri-
delays, the mechanism can help the regulator screen out infra- ously undermines the environmental effectiveness of the policy and an-
marginal participants. Its effectiveness is found to depend on the nihilates its social benefits. Interestingly, this study led the utility to ban
waiting period and the applicants' impatience. This solution remains ‘two-task’ contracts.
to be tested in the field. Another important energy efficiency policy where misplaced incen-
tives can be relevant is energy performance contracting. Under such
6.2. Moral hazard in subsidy programmes and energy performance contract, a retrofit contractor typically guarantees to the building man-
contracts ager that energy consumption after retrofit will not exceed a certain
level. The instrument is specifically meant to address informational bar-
Many policies can be seen as a principal-agent relationship between riers to energy efficiency improvements, in particular the moral hazard
the regulator – the principal – and some stakeholders – the agent, problem discussed in Section 5.1.2, which can be exacerbated by the in-
e.g., programme participants, contractors. In this context, incentive formation imperfections discussed in Section 2.2. Energy performance
problems can arise from poor policy design. Under perfect monitoring contracts are routinely offered by energy service companies in the in-
and verification, they should not affect policy outcomes, so that: dustrial and commercial sectors (Vine, 2005). They are much less com-
mon in the residential sector.
H10: Stakeholders do not respond to misplaced incentives stemming By addressing one form of moral hazard, energy performance con-
from poor policy design tracts may have the unintended effect of raising another moral hazard.
Just as the contractor's effort is unobservable, so too is the energy user's
How does this prediction fare in programmes that give rise to behaviour. In this informational context, having the contractor guaran-
misplaced incentives? One such programme is that studied by Blonz tee her energy consumption gives the energy user an incentive to over-
(2019) in California. In this subsidy programme, electric utilities hire consume – an incentive similar to that associated with utility-included
agents to perform two tasks: conduct an audit to determine the eligibil- contracts. Through calibrated simulations building up on their empirical
ity of an energy efficiency measure, and install the measure if eligible. analysis of a utility-sponsored retrofit programme, Giraudet et al.
Table 4
Evidence on imperfect signalling in building rental – market equilibrium (7 studies).

H2bis Owner-occupied dwellings are as energy-efficient as are rented ones

Reference Geographical scope Sectoral scope Fuel Energy usage Sample Method Model Time Hypothesis Certificate Comments
type size period validated programme

Brechling and UK Residential: NA NA 6395 Logit 1986 No Lower prevalence of loft insulation, wall
Smith (1994) Semi-detached, Reduced-Form insulation, double glazing
terraced
dwellings,

L.-G. Giraudet / Energy Economics 87 (2020) 104698

apartments
Davis (2012) US Residential NA Household Owners: Linear 2005 No Energy Star Tenants are less likely to report owning energy
appliances: 2979, probability efficient household appliances (5.6% for
refrigerators, renters: model refrigerators, 8.6% for dishwashers, 2.7% for
clothes 1219 clothes washers).
washers,
dishwashers,
lighting
Gillingham California, US Residential: Heating, cooling 20,933 Maximum Ordered 2003 No Owner-occupied dwellings are 20% more likely
et al. (2012) single-family likelihood probit model to be insulated in the attic or ceiling and 13% in
houses, exterior walls.
townhouses,
apartments
Krishnamurthy Australia, Canada, Chile, France, Residential NA NA 9432 Probit, Logit 2011 No Owners are 45% more likely to have
and Kriström Israel, South Korea, Japan, the model energy-efficient appliances and 2% more likely
(2015) Netherlands, Spain, Sweden, to have solar panels.
Switzerland
Melvin (2018) US Residential NA Heating, 11,498 Logit, Probit, 2009 No Rented dwellings less energy-efficient when it
cooling, other Linear comes to space-heating, water-heating,
Probability window thickness, insulation and
Model weatherization.
Myers (2015) Northeastern US Residential Heating Heating 6163 Fixed 2005–2009 No
oil, effects
natural regression
gas
Scott (1997) Ireland Residential NA NA 1200 Logistic November No
regression 1992

13
14 L.-G. Giraudet / Energy Economics 87 (2020) 104698

Table 5
Evidence on imperfect screening in utility-included contracts (3 studies).

H3 The rent premium for including utilities covers effective energy usage

Reference Geographical Sectoral Fuel type Energy Sample Method Model Time Hypothesis Certificate Comments
scope scope usage size period validated programme

Choi and Kim 21 Residential NA Heating, 4895 Logit Hedonic 2000–2002 No Utilities-included
(2012) metropolitan cooling, estimation model dwellings let out for
area in the US other higher rents than
comparable metered
dwellings
Levinson and 148 Residential: Natural gas, Heating 31,293 OLS 1985–1997 No
Niemann metropolitan Apartments fuel oil,
(2004) areas in the electricity,
US liquefied
propane gas
Myers (2015) Northeastern Residential Heating oil, Heating 6163 Fixed 2005–2009 No
US natural gas effects
regression

(2018) suggest that the deadweight loss associated with this form of incomplete, imperfect or asymmetric information – and behavioural
moral hazard reduces the efficiency of energy performance contracts, treatments. This is especially the case with energy-efficiency labels,
in particular relative to other forms of regulation such as certification which can serve either as a pure information provision addressing in-
of professional contractors. Note that this second moral hazard is only complete information, as a device levelling information between
relevant when the building manager and the energy user are one and contracting parties, or as a social norm provoking departures from indi-
the same person, which is typically the case in owner-occupied dwell- vidual rationality. In randomized experiments in the lightbulb market,
ings. This can explain why energy performance contracts are so scarce Allcott and Taubinsky (2015) provide information treatments and ob-
in the residential sector. serve how they affect consumers' willingness-to-pay for compact fluo-
The scant available evidence so far suggests that, when relevant, H10 rescent lightbulbs. The authors interpret the treatment as a “pure
is rejected. nudge” and assume that consumers' responses reveal the average mar-
ginal inattention bias. This study is an important first step highlighting
7. Discussion the importance of heterogeneity in consumer responses. In more recent
work, Astier (2018) proposes an interesting design to separate out in-
7.1. Information problems and behavioural anomalies formation provision from social norms. Online participants are first ran-
domly assigned to complete and incomplete information environments
Besides debate over the market-failure nature of barriers to energy then randomly assigned to different treatments: comparative feedback,
efficiency, an important research effort has been dedicated to behav- information only, and warning to outliers. The author finds that, while
ioural anomalies in energy-efficiency decisions in the past decade. Envi- feedbacks produce additional energy savings, complete information is
ronmental topics, and energy efficiency in particular, offer interesting a necessary condition for their effectiveness.
opportunities to test the predictions of the emerging field of behavioural
economics (Shogren and Taylor, 2008; Gillingham and Palmer, 2014; 7.2. What to expect from information technologies?
Allcott, 2016). Consumers indeed seem to value energy savings in a
way that is inconsistent with perfect rationality (Attari et al., 2010). Given the central role information technologies have come to play in
Much research along this line has focused on feedback experiments the economy, it seems natural to ponder on how they can support
with peer comparison, in which consumers are provided with informa- energy-efficiency improvements, which are subject to so many infor-
tion about how their energy use compares to that of their neighbours mation problems. The works reviewed here suggests that smart
(e.g., Allcott, 2011; Allcott and Rogers, 2014). Overall, such interven- metering and in-home displays of energy use can significantly improve
tions are found to strengthen conservation behaviours, however with market outcomes. So can emerging technologies such as thermo-
low persistence (Ayres et al., 2012; Delmas et al., 2013). This finding photography and other tests which enable verification of building per-
suggests that social norms influence an individual's behaviour, a feature formance. Nevertheless, the algorithms used to predict energy savings
not captured by the standard microeconomic model. still seem to lack accuracy. Another area for improvement is the devel-
As transpired throughout the review, however, behavioural anoma- opment of platforms facilitating search for retrofit contractors.
lies are difficult to separate out from information problems. Most empir- The question examined here echoes a broader reflection about
ical settings simultaneously involve informational barriers – whether recent breakthroughs in information technology mean the

Table 6
Evidence on imperfect screening in energy-efficiency loans (2 studies).

H4 Energy efficiency investments carry lower interest rates than do conventional ones

Reference Geographical Sectoral Fuel Energy Sample Method Model Time Hypothesis Certificate Comments
scope scope type usage size period validated programme

An and Pivo (2017) US Commercial NA NA 6304 Logit 2000–2013 Yes LEED, Energy efficiency investments
(Office model Energy Star display better loan terms if they
buildings) were certified green at loan
origin.
Giraudet et al. (2019) France Residential NA NA 240,962 OLS 2015–2016 No Based on posted interest-rate
data
 L.-G. Giraudet / Energy Economics 87 (2020) 104698 15

Table 7
Evidence on moral hazard in utility-included contracts (6 studies).

H5 Occupants use as much energy under utility-included contracts as under individual billing

Reference Geographical Sectoral Fuel type Energy Sample Method Model Time Hypothesis Certificate Comments
scope scope usage size period validated programme

Elinder Sweden Residential: NA NA 800 in Diff-in-diff January 1, No Tenants reduce their

et al. apartments treatment 2007 – energy consumption while
(2017) group, December switching from
1000 in 31, 2015 landlord-pay to
control tenant-pay regime.
group
Gillingham California, US Residential: NA Heating, 20,933 Maximum Ordered 2003 No Tenants in tenant-pay
et al. single-family cooling likelihood probit regimes more likely to
(2012) houses, model reduce heating during
townhouses, night.
apartments
Kahn et al. US Commercial Electricity NA 3521 Fixed 2009 No Energy consumption is
(2014) effects higher in buildings where
model the tenants are not
impacted by marginal cost
of energy consumption.
Levinson 148 Residential: Natural gas, Heating 31,293 OLS 1985–1997 No
and metropolitan Apartments fuel oil,
Niemann areas in the electricity,
(2004) US liquefied
propane
gas
Maruejols Canada Residential: Electricity, Space 4551 Least 2003 No
and multi-family natural gas, heating, squares
Young dwellings oil hot regression
(2011) water
Myers Northeastern Residential Heating oil, Heating 6163 Fixed 2005–2009 ?
(2015) US natural gas effects
regression

Table 8
Evidence on moral hazard in home energy retrofits (1 study).

H6 Retrofit contractors put identical effort into observable and unobservable energy efficiency tasks

Reference Geographical Sectoral Fuel type Energy Sample Method Model Time Hypothesis Certificate Comments
scope scope usage size period validated programme

Giraudet Florida, US Residential Electricity NA 2936 Diff-in-diff 2006–2012 No Moral hazard only occurs on Fridays. It
et al. dwellings and contributes to 70% of the discrepancy
(2018) natural between predicted and realized
gas savings.

end of information asymmetries (Cowen and Tabarrok, 2015). Prelimi- are particularly sensitive in the context of energy use, which infuses
nary research warrants healthy scepticism. For instance, internet mar- nearly every aspect of everyday life.
kets do not seem to reduce price dispersion, with some platforms
even engaging in obfuscation to compensate for increased competition 8. Conclusion
(Ellison and Ellison, 2005; Levin, 2011). In addition, internet ratings,
which are supposed to improve information, can be subject to manipu- Energy efficiency can be seen as a credence good, the performance of
lation (Luca and Zervas, 2016). Lastly, information technologies raise which is never fully revealed to the buyer. This characteristic is exacer-
privacy concerns which go far beyond economic inquiry. Those issues bated by the high upfront costs and multiplicity of stakeholders

Table 9
Evidence on moral hazard in energy-efficiency loans (2 studies).

H7 Borrowers default less when borrowing to save energy

Reference Geographical Sectoral scope Fuel Energy Sample size Method Model Time Hypothesis Certificate Comments
scope type usage period validated programme

An and US Commercial NA NA 6304 Logit model 2000–2013 Yes LEED, Energy efficiency
Pivo (Office buildings) Energy Star investments carry
(2017) lower default risk.
Kaza US Residential: NA NA 24,944 in Diff-in-diff Multinomial 2002–2010 Yes Energy Star Mortgages on Energy
et al. single-family treatment group, Logit model Star houses are
(2014) owner-occupied 46,118 in control one-third less likely
houses group to default.
16 L.-G. Giraudet / Energy Economics 87 (2020) 104698

Table 10
Evidence on price-discrimination in energy-efficiency product lines (3 studies).

H8 Producers offer consumers their optimal level of energy efficiency

Reference Geographical Sectoral scope Fuel type Energy Sample Method Model Time Hypothesis Certificate Comments
scope usage size period validated programme

Cohen UK Household Electricity Refrigerators 2421 2SLS Discrete-choice 2002–2007 No Consumers

et al. appliances: demand model, undervalue
(2017) refrigerators nested logit future energy
model costs by 35%.
Houde US Household Electricity Refrigerators 2752 Diff-in-diff January No Energy Star Products with
(2018b) appliances: 2007 – environmental
refrigerators December certifications are
2008 not only more
expensive due to
a higher cost for
producing
energy-efficient
technology but
producers also
charge a higher
premium for the
label itself.
Spurlock US Household Electricity Clothes 699 in Diff-in-diff 2003–2007 No Energy Star Clothes washer
(2013) appliances: washers treatment prices dropped
clotheswashers group, within-model
1415 in after 2004 and
control 2007 energy
group efficiency
standard
changes.

involved in building investments. As a result, building energy efficiency asymmetries have been understudied in the context of labour-
is subject to an array of information asymmetries, arguably more so intensive supply (moral hazard and price discrimination) and financing
than other well-studied credence goods such as medical treatments, (screening and moral hazard) of energy efficiency. Lastly, policy exper-
taxi rides or auto repair. iments could be conducted to test the promising solutions that are being
In this essay, I reviewed evidence of informational barriers to proposed to overcome the information asymmetries affecting energy ef-
energy-efficiency investment, with particular attention to whether ficiency policies.
they qualify as market failures – in the context studied here, information To conclude, it is worth noting that the procurement of retrofits to
asymmetries – or not – symmetric-information problems. I found that building contractors occurs very upstream in the production of energy
some information barriers are well documented, while others are either efficiency. Anticipating the moral hazard problem discussed in this
inaccurately characterized, or not clearly established, or simply paper (cf. Section 5.1.2), it might be difficult for a building manager to
overlooked. hire a diligent contractor. Such a screening problem is unexplored. Yet
I first noted that information relevant to operating energy- if confirmed, it might propagate in related transactions, such as building
consuming assets is incomplete and imperfect in many contexts, with rental and sales. Downstream, on the other hand, financing is somehow
unclear conclusions as to whether information provision improves mar- the recipient of all other information asymmetries. In the United States
ket outcomes. alone, the market for energy-efficiency finance is estimated to amount
I then reviewed ten possible sources of information asymmetries – to $100 billion annually (Freehling and Stickles, 2016). More research
that is, true market failures – and found different degrees of significance. is therefore needed into these two crucial topics – retrofit procurement
The longest-studied problems are those potentially occurring in and financing.
landlord-tenant relationships. While indistinctly referred to as “split in-
centives” in the literature, I classified them into three categories: signal- Author contribution
ling in rental buildings, moral hazard and screening in utility-included
rent contracts. I found evidence on the first problem to be inconsistent, Louis-Gaëtan Giraudet did all the analysis (including scoping, paper
as the capitalization of energy efficiency observed does not seem to be search, paper selection and paper interpretation) and the writing.
reflected in market equilibrium. In constrast, I found compelling evi- To complete the second revision, Pille-Riin Aja was hired as a re-
dence of moral hazard and screening problems in utility-included rent search assistant. She helped with putting together Tables 2 to 12.
contracts, which however does not seem to induce dramatic over-use
of energy. One implication is that banning utility-included contracts Acknowledgements
could improve social welfare, though this recommendation must be
weighed against forgone benefits that remain to be quantified. Another I gratefully acknowledge funding from the European Investment
much studied information asymmetry is signalling in building sales, the Bank Institute under the University Research Sponsorship Programme,
analysis of which has been facilitated by implementation of energy per- Grant EIBI/KnP/TT/ck (1-RGI-C311). I thank Pille-Riin Aja for excellent
formance certificates. Here, the conclusion is ambiguous. Prospective research assistance. I thank Nicolas Astier, Andreas Gerster, Ken Gilling-
buyers clearly respond to information labels, but two counterfactual ham, Sébastien Houde, Laurent Lamy, Philippe Quirion, François Salanié
benchmarks are often missing to ascertain that labels operate by elimi- for helpful comments on earlier drafts. I also thank an anonymous re-
nation of an information asymmetry: what occurs without labels (in viewer of the French Association of Environmental and Resource Econ-
order to identify information levelling), and what occurs with coarser omists for helpful comments on the draft that the Association came to
labels (in order to identify a social-norm effect). Information publish as a working paper.
Table 11
Evidence on imperfect screening in subsidy programmes (10 studies).

H9 Regulators optimally screen out infra-marginal participants in energy efficiency subsidy programmes

Reference Geographical Sectoral Fuel type Energy usage Sample Method Model Time Hypothesis Certificate Comments
scope scope size period validated programme

Alberini Maryland, US Residential Electricity NA 108,387 Diff-in-diff-in-diff 2008–2012 No Data displayed high levels of heat pump usage prior to its
et al. replacement which suggests that the incentives are used
(2016) for cases where the appliances would be replaced in the
absence of incentives.
Boomhower US Residential Electricity Appliances: 237,552 RD May No More than two-thirds of participants are inframarginal
and Davis refrigerators and 2009–April and approximately 50% of all participants would have

L.-G. Giraudet / Energy Economics 87 (2020) 104698

(2014) air-conditioners 2011 replaced their appliances without any subsidy offered.
Collins and Ireland Residential NA NA 28,454 OLS June No 7% of inframarginal participants
Curtis 2010 – July
(2018) 2016
Grösche and Western Residential: NA NA 21,28 Error components logit 2005 No Around 50% of inframarginal participants
Vance Germany single-family model
(2009) dwellings
Hassett and US Residential NA NA 74,792 Logit fixed effects Discrete choice model 1979–1981 ? Inframarginal participation is not directly quantified.
Metcalf Evidence that participants are not strategically delaying
(1995) conservation investment
Houde and US Residential Electricity Appliances: 37,150 Diff-in-diff January No 70% of inframarginal participants
Aldy refrigerators, 2008 –
(2017) clothes washers, November
dishwashers 2012
Joskow and US Residential Electricity 1991 No
Marron and
(1992) commercial
Malm US Residential Heating No 89% of participants are considered inframarginal although
(1996) the figure might be an overestimation given the lack of
exact definition of ‘highly energy-efficient’ among
respondents
Nauleau France Residential Electricity, 36,367 Random effect 2002–2011 No 40–85% of inframarginal participants after 2006
(2014) gas, fuel dichotomous logit
model
Rivers and Canada Residential Natural Heating 328,688 Nested logit 1 April No 50% of inframarginal participants in the subsidy and tax
Shiell gas estimation, weighted 2007–31 credit programmes
(2016) exogenous sample March
maximum likelihood 2011

17
18 L.-G. Giraudet / Energy Economics 87 (2020) 104698

Table 12
Evidence on moral hazard in subsidy programmes and energy performance contracts (2 studies).

H10 Stakeholders do not respond to misplaced incentives stemming from poor policy design

Reference Geographical Sectoral Fuel type Energy Sample Method Model Time Hypothesis Certificate Comments
scope scope usage size period validated programme

Blonz California, US Residential Electricity Appliances: 271,126 January No Contractors misreported data for
(2019) Refrigerators 2009 – 19% and 7.8% of households, for
September an average incentive of $123
2012 (‘integrated-task’) and $25
(‘separated-task’) per report,
respectively.
Giraudet Florida, US Residential NA NA 2936 Diff-in-diff 2006–2012 N/A Numerical simulations
et al. dwellings suggesting that consumer moral
(2018) hazard could significantly
undermine the efficiency of
energy savings insurance

References Brechling, V., Smith, S., 1994. Household energy efficiency in the UK. Fisc. Stud. 15, 44–56.
https://doi.org/10.1111/j.1475-5890.1994.tb00196.x.
Brounen, D., Kok, N., 2011. On the economics of energy labels in the housing market.
Abrahamse, W., Steg, L., Viek, C., Rothengatter, T., 2005. A review of intervention studies J. Environ. Econ. Manag. 62, 166–179. https://doi.org/10.1016/j.jeem.2010.11.006.
aimed at household energy conservation. J. Environ. Psychol. 25, 273–291. Buchanan, K., Russo, R., Anderson, B., 2015. The question of energy reduction: the prob-
Alberini, A., Towe, C., 2015. Information v. energy efficiency incentives: evidence from lem(s) with feedback. Energy Policy 77, 89–96. https://doi.org/10.1016/j.
residential electricity consumption in Maryland. Energy Economics, Frontiers in the enpol.2014.12.008.
Economics of Energy Efficiency 52 (Supplement 1), S30–S40. https://doi.org/ Chen, V.L., Delmas, M.A., Kaiser, W.J., Locke, S.L., 2015. What can we learn from high-
10.1016/j.eneco.2015.08.013. frequency appliance-level energy metering? Results from a field experiment. Energy
Alberini, A., Gans, W., Towe, C., 2016. Free riding, upsizing, and energy efficiency incen- Policy 77, 164–175. https://doi.org/10.1016/j.enpol.2014.11.021.
tives in Maryland homes. Energy J. 37. https://doi.org/10.5547/01956574.37.1.aalb. Choi, S.J., Kim, S., 2012. Why do landlords include utilities in rent? Evidence from the 2000
Allcott, H., 2011. Social norms and energy conservation. Journal of Public Economics, Spe- Housing Discrimination Study (HDS) and the 2002 American Housing Survey (AHS).
cial Issue: The Role of Firms in Tax Systems 95, 1082–1095. https://doi.org/10.1016/j. J. Hous. Econ. 21, 28–40. https://doi.org/10.1016/j.jhe.2012.02.001.
jpubeco.2011.03.003. Cohen, F., Glachant, M., Söderberg, M., 2017. Consumer myopia, imperfect competition
Allcott, H., 2016. Paternalism and energy efficiency: an overview. Annu. Rev. Econ. 8, and the energy efficiency gap: evidence from the UK refrigerator market. Eur. Econ.
145–176. https://doi.org/10.1146/annurev-economics-080315-015255. Rev. 93, 1–23. https://doi.org/10.1016/j.euroecorev.2017.01.004.
Allcott, H., Greenstone, M., 2012. Is there an energy efficiency gap? J. Econ. Perspect. 26, Collins, M., Curtis, J., 2018. Willingness-to-pay and free-riding in a national energy effi-
3–28. https://doi.org/10.1257/jep.26.1.3. ciency retrofit grant scheme. Energy Policy 118, 211–220. https://doi.org/10.1016/j.
Allcott, H., Rogers, T., 2014. The short-run and long-run effects of behavioural interven- enpol.2018.03.057.
tions: experimental evidence from energy conservation. Am. Econ. Rev. 104, Considine, T.J., Sapci, O., 2016. The effectiveness of home energy audits: a case study of
3003–3037. https://doi.org/10.1257/aer.104.10.3003. Jackson, Wyoming. Resour. Energy Econ. 44, 52–70. https://doi.org/10.1016/j.
Allcott, H., Sweeney, R.L., 2017. The role of sales agents in information disclosure: evi- reseneeco.2016.02.004.
dence from a field experiment. Manag. Sci. 63, 21–39. https://doi.org/10.1287/ Cowen, T., Tabarrok, A., 2015. The End of Asymmetric Information (Cato Unbound).
mnsc.2015.2327. Davis, L.W., 2012. Evaluating the slow adoption of energy efficient investments: are
Allcott, H., Taubinsky, D., 2015. Evaluating behaviourally motivated policy: experimental renters less likely to have energy efficient appliances? In: Fullerton, D., Wolfram, C.
evidence from the lightbulb market. Am. Econ. Rev. 105, 2501–2538. https://doi.org/ (Eds.), The Design and Implementation of U.S. Climate Policy. University of Chicago
10.1257/aer.20131564. Press, pp. 301–316
An, X., Pivo, G., 2017. Green Buildings in Commercial Mortgage Backed. The Effects of Davis, L.W., Metcalf, G.E., 2016. Does better information Lead to better choices? Evidence
LEED and Energy Star Certification on Default Risk and Loan Terms. Real Estate Eco- from energy-efficiency labels. J. Assoc. Environ. Resour. Econ. 3, 589–625. https://doi.
nomics, Securities https://doi.org/10.1111/1540-6229.12228. org/10.1086/686252.
Anderson, C.D., Claxton, J.D., 1982. Barriers to consumer choice of energy efficient prod- Delmas, M.A., Lessem, N., 2014. Saving power to conserve your reputation? The effective-
ucts. J. Consum. Res. 9, 163–170. ness of private versus public information. J. Environ. Econ. Manag. 67, 353–370.
Anderson, S.T., Newell, R.G., 2004. Information programs for technology adoption: the https://doi.org/10.1016/j.jeem.2013.12.009.
case of energy-efficiency audits. Resour. Energy Econ. 26, 27–50. https://doi.org/ Delmas, M.A., Fischlein, M., Asensio, O.I., 2013. Information strategies and energy conser-
10.1016/j.reseneeco.2003.07.001. vation behaviour: a meta-analysis of experimental studies from 1975 to 2012. Energy
Andor, M., Gerster, A., Sommer, S., 2016. Consumer Inattention, Heuristic Thinking and Policy 61, 729–739. https://doi.org/10.1016/j.enpol.2013.05.109.
the Role of Energy Labels (SSRN Scholarly Paper No. ID 2795579). Social Science Re- Dinan, T.M., Miranowski, J.A., 1989. Estimating the implicit price of energy efficiency im-
search Network, Rochester, NY. provements in the residential housing market: a hedonic approach. J. Urban Econ. 25,
Ansar, J., Sparks, R., 2009. The experience curve, option value, and the energy paradox. En- 52–67. https://doi.org/10.1016/0094-1190(89)90043-0.
ergy Policy 37, 1012–1020. https://doi.org/10.1016/j.enpol.2008.10.037. Dixit, A.K., Pindyck, R.S., 1994. Investment Under Uncertainty. Princeton University Press.
Attari, S.Z., DeKay, M.L., Davidson, C.I., Bruinde, W.B., 2010. Public perceptions of energy Dressler, L., Cornago, E., 2018. Incentives to (not) Disclose Energy Performance Informa-
consumption and savings. PNAS 107, 16054–16059. https://doi.org/10.1073/ tion in the Housing Market. https://ideas.repec.org/p/eca/wpaper/2013-278920.
pnas.1001509107. html.
Ayres, I., Raseman, S., Shih, A., 2012. Evidence from Two Large Field Experiments that Peer Dulleck, U., Kerschbamer, R., 2006. On doctors, mechanics, and computer specialists: the
Comparison Feedback Can Reduce Residential Energy Usage. J Law Econ Organ economics of credence goods. J. Econ. Lit. 44, 5–42.
ews020. doi:https://doi.org/10.1093/jleo/ews020. Eichholtz, P., Kok, N., Quigley, J.M., 2010. Doing well by doing good? Green Office Build-
Astier, N., 2018. Comparative feedbacks under incomplete information. Resour. Energy ings. The American Economic Review 100, 2492–2509.
Econ. 54, 90–108. https://doi.org/10.1016/j.reseneeco.2018.07.002. Eichholtz, P., Kok, N., Quigley, J.M., 2012. The economics of green building. Rev. Econ. Stat.
Backlund, S., Thollander, P., 2015. Impact after three years of the Swedish energy audit 95, 50–63. https://doi.org/10.1162/REST_a_00291.
program. Energy 82, 54–60. https://doi.org/10.1016/j.energy.2014.12.068. Elinder, M., Escobar, S., Petré, I., 2017. Consequences of a price incentive on free riding and
Baker, E., 2012. Option value and the diffusion of energy efficient products. Energy J. 33, electric energy consumption. PNAS 201615290. https://doi.org/10.1073/
49–59. pnas.1615290114.
Bala, A.P., Peeters, D., Thomas, I., 2014. Spatial issues on a hedonic estimation of rents in Ellison, G., Ellison, S.F., 2005. Lessons about markets from the Internet. J. Econ. Perspect.
Brussels. J. Hous. Econ. 25, 104–123. https://doi.org/10.1016/j.jhe.2014.05.002. 19, 139–158.
Barbetta, G.P., Canino, P., Cima, S., 2015. The impact of energy audits on energy efficiency EPA, U.S., 2015. Notice of Violation: Volkswagen Clean Air Act Partial Settlements.
investment of public owners. Evidence from Italy. Energy 93 (Part 1), 1199–1209. Farsi, M., 2010. Risk aversion and willingness to pay for energy efficient systems in rental
https://doi.org/10.1016/j.energy.2015.09.117. apartments. Energy Policy, The Role of Trust in Managing Uncertainties in the Transi-
Blonz, J.A., 2019. The Welfare Costs of Misaligned Incentives: energy Inefficiency and the tion to a Sustainable Energy Economy, Special Section with Regular Papers 38,
Principal-Agent Problem (SSRN Scholarly Paper No. ID 3473057). Social Science Re- 3078–3088. https://doi.org/10.1016/j.enpol.2010.01.048.
search Network, Rochester, NY. Fesselmeyer, E., 2018. The value of green certification in the Singapore housing market.
Boomhower, J., Davis, L.W., 2014. A credible approach for measuring inframarginal partic- Econ. Lett. 163, 36–39. https://doi.org/10.1016/j.econlet.2017.11.033.
ipation in energy efficiency programs. J. Public Econ. 113, 67–79. https://doi.org/ Fischer, C., 2005. On the importance of the supply side in demand-side management. En-
10.1016/j.jpubeco.2014.03.009. ergy Econ. 27, 165–180. https://doi.org/10.1016/j.eneco.2004.11.001.
 L.-G. Giraudet / Energy Economics 87 (2020) 104698 19

Fischer, C., 2008. Feedback on household electricity consumption: a tool for saving en- Johnson, R.C., Kaserman, D.L., 1983. Housing market capitalization of energy-saving dura-
ergy? Energy Efficiency 1, 79–104. https://doi.org/10.1007/s12053-008-9009-7. ble good investments. Econ. Inq. 21, 374–386. https://doi.org/10.1111/j.1465-
Fowlie, M., Greenstone, M., Wolfram, C., 2015. Do Energy Efficiency Investments Deliver? 7295.1983.tb00639.x.
Evidence from the Weatherization Assistance Program (Working Paper No. 21331). Joskow, P.L., Marron, D.B., 1992. What does a Negawatt really cost? Evidence from Utility
National Bureau of Economic Research. Conservation Programs. The Energy Journal 13, 41–74.
Freehling, J., Stickles, B., 2016. Energy Efficiency Finance: A Market Reassessment|ACEEE Kahn, M.E., Kok, N., 2014. The capitalization of green labels in the California housing
(ACEEE White Paper). market. Regional Science and Urban Economics, SI: Tribute to John Quigley 47,
Frondel, M., Vance, C., 2013. Heterogeneity in the effect of home energy audits: theory 25–34.
and evidence. Environ. Resour. Econ. 55, 407–418. https://doi.org/10.1007/s10640- Kahn, M.E., Kok, N., Quigley, J.M., 2014. Carbon emissions from the commercial building
013-9632-4. sector: the role of climate, quality, and incentives. J. Public Econ. 113, 1–12. https://
Frondel, M., Gerster, A., Vance, C., 2020. The Power of Mandatory Quality Disclosure: Ev- doi.org/10.1016/j.jpubeco.2014.03.003.
idence from the German Housing Market. J. Assoc. Environ. Resour. Econ. 7, 181–208. Kallbekken, S., Sælen, H., Hermansen, E.A.T., 2013. Bridging the energy efficiency gap: a
https://doi.org/10.1086/705786. field experiment on lifetime energy costs and household appliances. J. Consum. Policy
Fuerst, F., McAllister, P., 2011a. Green noise or green value? Measuring the effects of en- 36, 1–16.
vironmental certification on office values. Real Estate Econ. 39, 45–69. Kaza, N., Quercia, R.G., Tian, C.Y., 2014. Home energy efficiency and mortgage risks. City-
Fuerst, F., McAllister, P., 2011b. The impact of energy performance certificates on the scape: A Journal of Policy Development and Research 16, 279–298.
rental and capital values of commercial property assets. Energy Policy, Sustainability Kok, N., Jennen, M., 2012. The impact of energy labels and accessibility on office rents. En-
of biofuels 39, 6608–6614. https://doi.org/10.1016/j.enpol.2011.08.005. ergy Policy 46, 489–497. https://doi.org/10.1016/j.enpol.2012.04.015.
Fuerst, F., McAllister, P., Nanda, A., Wyatt, P., 2015. Does energy efficiency matter to Krishnamurthy, C.K.B., Kriström, B., 2015. How large is the owner-renter divide in energy
home-buyers? An investigation of EPC ratings and transaction prices in England. En- efficient technology? Evidence from an OECD cross-section. Energy J. 36. https://doi.
ergy Econ. 48, 145–156. https://doi.org/10.1016/j.eneco.2014.12.012. org/10.5547/01956574.36.4.ckri.
Gerarden, T.D., Newell, R.G., Stavins, R.N., 2017. Assessing the energy-efficiency gap. Laquatra, J., 1986. Housing market capitalization of thermal integrity. Energy Econ. 8,
J. Econ. Lit. 55, 1486–1525. https://doi.org/10.1257/jel.20161360. 134–138. https://doi.org/10.1016/0140-9883(86)90011-3.
Gillingham, K., Palmer, K., 2014. Bridging the energy efficiency gap: policy insights from Laquatra, J., Dacquisto, D.J., Emrath, P., Laitner, J.A., 2002. Housing Market Capitalization of
economic theory and empirical evidence. Rev Environ Econ Policy ret021. https:// Energy Efficiency Revisited — ECEEE [WWW Document]. URL. http://www.eceee.org/
doi.org/10.1093/reep/ret021. library/conference_proceedings/ACEEE_buildings/2002/Panel_8/p8_12 (accessed
Gillingham, K., Newell, R.G., Palmer, K., 2009. Energy efficiency economics and policy. 2.29.16).
Ann. Rev. Resour. Econ. 1, 597–620. https://doi.org/10.1146/annurev. Levin, J.D., 2011. The Economics of Internet Markets (Working Paper No. 16852). National
resource.102308.124234. Bureau of Economic Research.
Gillingham, K., Harding, M., Rapson, D., 2012. Split incentives in residential energy con- Levinson, A., Niemann, S., 2004. Energy use by apartment tenants when landlords pay for
sumption. Energy J. 33, 37–62. utilities. Resour. Energy Econ. 26, 51–75. https://doi.org/10.1016/S0928-7655(03)
Gilmer, R.W., 1989. Energy labels and economic search. Energy Econ. 11, 213–218. 00047-2.
https://doi.org/10.1016/0140-9883(89)90026-1. Lewis, T.R., Sappington, D.E.M., 1992. Incentives for conservation and quality-
Giraudet, L.-G., Houde, S., Maher, J., 2018. Moral Hazard and the energy efficiency gap: improvement by public utilities. Am. Econ. Rev. 82, 1321–1340.
theory and evidence. J. Assoc. Environ. Resour. Econ. 5, 755–790. https://doi.org/ Luca, M., Zervas, G., 2016. Fake it till you make it: reputation, competition, and yelp re-
10.1086/698446. view fraud. Manag. Sci. 62, 3412–3427. https://doi.org/10.1287/mnsc.2015.2304.
Giraudet, L.-G., Petronevich, A., Faucheux, L., 2019. How do lenders price energy effi- Malm, E., 1996. An actions-based estimate of the free rider fraction in electric utility DSM
ciency? Evidence from personal consumption loans. Banque de France Working programs. Energy J. 17. https://doi.org/10.5547/ISSN0195-6574-EJ-Vol17-No3-3.
Paper 716. https://publications.banque-france.fr/sites/default/files/medias/docu- Maruejols, L., Young, D., 2011. Split incentives and energy efficiency in Canadian multi-
ments/wp716.pdf1. family dwellings. Energy Policy 39, 3655–3668. https://doi.org/10.1016/j.
Globus-Harris, I., 2018. Waiting Periods as a Screening Mechanism for Environmental enpol.2011.03.072.
Subsidies. Working paper. https://drive.google.com/file/d/ Mas-Colell, A., Whinston, M.D., Green, J.R., 1995. Microeconomic Theory (OUP USA).
1si61qoSSqrlwMf1xhJyDlQ_vnI9Kp0Ss/view?usp=sharing&usp=embed_facebook. Matisoff, D.C., Noonan, D.S., Flowers, M.E., 2016. Policy monitor—green buildings: eco-
Graff Zivin, J., Novan, K., 2016. Upgrading efficiency and behaviour: electricity savings nomics and policies. Rev. Environ. Econ. Policy 10, 329–346. https://doi.org/
from residential weatherization programs. Energy J. 37. https://doi.org/10.5547/ 10.1093/reep/rew009.
01956574.37.4.jziv. Matsukawa, I., 2004. The effects of information on residential demand for electricity. En-
Grösche, P., Vance, C., 2009. Willingness to pay for energy conservation and free-ridership ergy J. 25. https://doi.org/10.5547/ISSN0195-6574-EJ-Vol25-No1-1.
on subsidization: evidence from Germany. Energy J. 30, 135–154. McDougall, G.H.G., Claxton, J.D., Ritchie, J.R.B., 1982. Residential home audits: an empirical
Harjunen, O., Liski, M., 2014. Not so Myopic Consumers - Evidence on Capitalization of En- analysis of the Enersave program. J. Environ. Syst. 12, 265–278. https://doi.org/
ergy Technologies in a Housing Market (SSRN Scholarly Paper No. ID 2507740). Social 10.2190/8H0K-D3QC-TGY8-9YCJ.
Science Research Network, Rochester, NY. Melvin, J., 2018. The split incentives energy efficiency problem: evidence of underinvest-
Hassett, K.A., Metcalf, G.E., 1993. Energy conservation investment. Energy Policy 21, ment by landlords. Energy Policy 115, 342–352. https://doi.org/10.1016/j.
710–716. https://doi.org/10.1016/0301-4215(93)90294-P. enpol.2017.11.069.
Hassett, K.A., Metcalf, G.E., 1995. Energy tax credits and residential conservation invest- Metcalf, G.E., Hassett, K.A., 1999. Measuring the energy savings from home improvement
ment: evidence from panel data. J. Public Econ. 57, 201–217. https://doi.org/ investments: evidence from monthly billing data. Rev. Econ. Stat. 81, 516–528.
10.1016/0047-2727(94)01452-T. https://doi.org/10.1162/003465399558274.
Hausman, J.A., 1979. Individual discount rates and the purchase and utilization of energy- Murphy, L., 2014a. The influence of energy audits on the energy efficiency investments of
using durables. Bell J. Econ. 10, 33–54. https://doi.org/10.2307/3003318. private owner-occupied households in the Netherlands. Energy Policy 65, 398–407.
Hirst, E., Goeltz, R., 1985. Estimating energy savings due to conservation programmes: the https://doi.org/10.1016/j.enpol.2013.10.016.
BPA residential weatherization pilot programme. Energy Econ. 7, 20–28. https://doi. Murphy, L., 2014b. The influence of the energy performance certificate: the Dutch case.
org/10.1016/0140-9883(85)90035-0. Energy Policy 67, 664–672. https://doi.org/10.1016/j.enpol.2013.11.054.
Houde, S., 2018a. How consumers respond to product certification and the value of en- Mussa, M., Rosen, S., 1978. Monopoly and product quality. J. Econ. Theory 18, 301–317.
ergy information. RAND J. Econ. 49, 453–477. https://doi.org/10.1111/1756- https://doi.org/10.1016/0022-0531(78)90085-6.
2171.12231. Myers, E., 2015. Asymmetric Information in Residential Rental Markets: Implications for
Houde, S., 2018b. Bunching with the Stars: How Firms Respond to Environmental Certifi- the Energy Efficiency Gap (No. 246R), EI @ Haas WP.
cation. E2e Working Paper 37. https://n.ethz.ch/~shoude/BunchingStars_Houde_ Myers, E., 2019. Are home buyers inattentive? Evidence from capitalization of energy
E2eWP037.pdf. costs. Am. Econ. J. Econ. Pol. 11, 165–188. https://doi.org/10.1257/pol.20170481.
Houde, S., Aldy, J.E., 2017. Consumers’ response to state energy efficient appliance rebate Nauleau, M.-L., 2014. Free-riding on tax credits for home insulation in France: an econo-
programs. Am. Econ. J. Econ. Pol. 9, 227–255. https://doi.org/10.1257/pol.20140383. metric assessment using panel data. Energy Econ. 46, 78–92. https://doi.org/10.1016/
Houde, S., Todd, A., Sudarshan, A., Flora, J., Armel, C., 2013. Real-time feedback and elec- j.eneco.2014.08.011.
tricity consumption: a field experiment assessing the potential for savings and persis- Nauleau, M.-L., Giraudet, L.-G., Quirion, P., 2015. Energy efficiency subsidies with price-
tence. Energy J. 34. quality discrimination. Energy Economics, Frontiers in the Economics of Energy Effi-
Howarth, R.B., Andersson, B., 1993. Market barriers to energy efficiency. Energy Econ. 15, ciency 52 (Supplement 1), S53–S62. https://doi.org/10.1016/j.eneco.2015.08.024.
262–272. https://doi.org/10.1016/0140-9883(93)90016-K. Nevin, R., Watson, G., 1998. Evidence of rational market valuations for home energy effi-
Hsu, D., 2014. How much information disclosure of building energy performance is nec- ciency. Appraisal Journal 66, 401–409.
essary? Energy Policy 64, 263–272. https://doi.org/10.1016/j.enpol.2013.08.094. Newell, R.G., Siikamäki, J., 2014. Nudging energy efficiency behaviour: the role of informa-
Huntington, H., Schipper, L., Sanstad, A.H., 1994. Editors’ introduction. Energy Policy 22, tion labels. J. Assoc. Environ. Resour. Econ. 1, 555–598. https://doi.org/10.1086/
795–797. https://doi.org/10.1016/0301-4215(94)90136-8. 679281.
Hyland, M., Lyons, R.C., Lyons, S., 2013. The value of domestic building energy efficiency — Olaussen, J.O., Oust, A., Solstad, J.T., 2017. Energy performance certificates – informing the
evidence from Ireland. Energy Econ. 40, 943–952. https://doi.org/10.1016/j. informed or the indifferent? Energy Policy 111, 246–254. https://doi.org/10.1016/j.
eneco.2013.07.020. enpol.2017.09.029.
Jaffe, A.B., Stavins, R.N., 1994. The energy-efficiency gap: what does it mean? Energy Pol- Palmer, K., Walls, M., Gerarden, T., 2012. Borrowing to Save Energy: An Assessment of
icy 22, 804–810. https://doi.org/10.1016/0301-4215(94)90138-4. Energy-Efficiency Financing Programs (Resources for the Future Discussion Paper).
Jessoe, K., Rapson, D., 2014. Knowledge is (less) power: experimental evidence from res- Palmer, K., Walls, M., Gordon, H., Gerarden, T., 2013. Assessing the energy-efficiency infor-
idential energy use. Am. Econ. Rev. 104, 1417–1438. https://doi.org/10.1257/ mation gap: results from a survey of home energy auditors. Energy Efficiency 6,
aer.104.4.1417. 271–292. https://doi.org/10.1007/s12053-012-9178-2.
20 L.-G. Giraudet / Energy Economics 87 (2020) 104698

Palmer, K.L., Walls, M.A., O’Keeffe, L., 2015. Putting Information into Action: What Ex- Sorrell, S., 2004. Understanding barriers to energy efficiency. In: Sorrell, S., O’Malley, E.,
plains Follow-up on Home Energy Audits? (Resources for the Future) Schleich, J., Scott, S. (Eds.), The Economics of Energy Efficiency: Barriers to Cost-
Park, J.Y., 2017. Is there a price premium for energy efficiency labels? Evidence from the effective Investment. Edward Elgar, Cheltenham, pp. 25–93.
introduction of a label in Korea. Energy Econ. 62, 240–247. https://doi.org/10.1016/j. Spurlock, A., 2013. Appliance Efficiency Standards and Price Discrimination [WWW
eneco.2016.12.004. Document].
Plourde, C., Bardis, V., 1999. Fuel economy standards in a model of automobile quality. En- Stanley, S., Lyons, R.C., Lyons, S., 2015. The price effect of building energy ratings in the
ergy Econ. 21, 309–319. https://doi.org/10.1016/S0140-9883(99)00012-2. Dublin residential market. Energy Efficiency, 1–11 https://doi.org/10.1007/s12053-
Ramos, A., Gago, A., Labandeira, X., Linares, P., 2015. The role of information for energy ef- 015-9396-5.
ficiency in the residential sector. Energy Economics, Frontiers in the Economics of En- Sunikka-Blank, M., Galvin, R., 2012. Introducing the prebound effect: the gap between
ergy Efficiency 52 (Supplement 1), S17–S29. https://doi.org/10.1016/j. performance and actual energy consumption. Building Research & Information 40,
eneco.2015.08.022. 260–273. https://doi.org/10.1080/09613218.2012.690952.
Reichardt, A., 2014. Operating expenses and the rent premium of energy Star and LEED Tiefenbeck, V., Goette, L., Degen, K., Tasic, V., Fleisch, E., Lalive, R., Staake, T., 2016. Over-
certified buildings in the central and eastern U.S. J Real Estate Finan Econ 49, coming salience bias: how real-time feedback fosters resource conservation. Manage-
413–433. ment Science https://doi.org/10.1287/mnsc.2016.2646.
Rivers, N., Shiell, L., 2016. Free-Riding on Energy Efficiency Subsidies: the Case of Natural Train, K., 1985. Discount rates in consumers’ energy-related decisions: a review of the lit-
Gas Furnaces in Canada. Energy J. 37. https://doi.org/10.5547/01956574.37.4.nriv. erature. Energy 10, 1243–1253. https://doi.org/10.1016/0360-5442(85)90135-5.
Sanstad, A.H., Blumstein, C., Stoft, S.E., 1995. How high are option values in energy- Vine, E., 2005. An international survey of the energy service company (ESCO) industry.
efficiency investments? Energy Policy 23, 739–743. https://doi.org/10.1016/0301- Energy Policy 33, 691–704. https://doi.org/10.1016/j.enpol.2003.09.014.
4215(95)00065-Q. Waechter, S., Sütterlin, B., Siegrist, M., 2015. Desired and undesired effects of energy la-
Schleich, J., 2004. Do energy audits help reduce barriers to energy efficiency? An empirical bels—an eye-tracking study. PLoS One 10, e0134132. https://doi.org/10.1371/jour-
analysis for Germany. International Journal of Energy Technology and Policy 2, nal.pone.0134132.
226–239. Wahlström, M.H., 2016. Doing good but not that well? A dilemma for energy conserving
Schleich, J., Klobasa, M., Gölz, S., Brunner, M., 2013. Effects of feedback on residential elec- homeowners. Energy Econ. 60, 197–205. https://doi.org/10.1016/j.
tricity demand—findings from a field trial in Austria. Energy Policy 61, 1097–1106. eneco.2016.09.025.
https://doi.org/10.1016/j.enpol.2013.05.012. Walls, M., Gerarden, T., Palmer, K., Bak, X.F., 2017. Is energy efficiency capitalized into
Scott, S., 1997. Household energy efficiency in Ireland: a replication study of ownership of home prices? Evidence from three U.S. cities. J. Environ. Econ. Manag. 82, 104–124.
energy saving items. Energy Econ. 19, 187–208. https://doi.org/10.1016/S0140-9883 https://doi.org/10.1016/j.jeem.2016.11.006.
(96)01000-6. de Wilde, P., 2014. The gap between predicted and measured energy performance of
Sexton, S., 2014. Automatic bill payment and salience effects: evidence from electricity buildings: a framework for investigation. Autom. Constr. 41, 40–49. https://doi.org/
consumption. Rev. Econ. Stat. 97, 229–241. https://doi.org/10.1162/REST_a_00465. 10.1016/j.autcon.2014.02.009.
Sexton, R.J., Sexton, T.A., Wann, J.J.-W., Kling, C.L., 1989. The conservation and welfare ef- Wirl, F., 1995. Impact of regulation on demand side conservation programs. J. Regul. Econ.
fects of information in a time-of-day pricing experiment. Land Econ. 65, 272–279. 7, 43–62. https://doi.org/10.1007/BF01062779.
https://doi.org/10.2307/3146671. Zhou, H., Bukenya, J.O., 2016. Information inefficiency and willingness-to-pay for energy-
Shogren, J.F., Taylor, L.O., 2008. On Behavioural-environmental economics. Rev. Environ. efficient technology: a stated preference approach for China Energy Label. Energy
Econ. Policy 2, 26–44. https://doi.org/10.1093/reep/rem027. Policy 91, 12–21. https://doi.org/10.1016/j.enpol.2015.12.040.