Journal of Management Information Systems

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/mmis20

How Firms Make Information Technology

Investment Decisions: Toward a Behavioral Agency
Theory

John Qi Dong, Prasanna P. Karhade, Arun Rai & Sean Xin Xu

To cite this article: John Qi Dong, Prasanna P. Karhade, Arun Rai & Sean Xin Xu (2021) How
Firms Make Information Technology Investment Decisions: Toward a Behavioral Agency Theory,
Journal of Management Information Systems, 38:1, 29-58, DOI: 10.1080/07421222.2021.1870382

To link to this article: https://doi.org/10.1080/07421222.2021.1870382

Published online: 02 Apr 2021.

Submit your article to this journal

Article views: 338

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=mmis20
JOURNAL OF MANAGEMENT INFORMATION SYSTEMS
2021, VOL. 38, NO. 1, 29–58
https://doi.org/10.1080/07421222.2021.1870382

How Firms Make Information Technology Investment

Decisions: Toward a Behavioral Agency Theory
a b c d
John Qi Dong , Prasanna P. Karhade , Arun Rai , and Sean Xin Xu
a
Trinity Business School, Trinity College Dublin, University of Dublin, Dublin, Ireland; bDepartment of
Information Technology Management, Shidler College of Business, University of Hawaii at Mānoa, Honolulu,
HI, USA; cCenter for Digital Innovation & Department of Computer Information Systems, J. Mack Robinson
College of Business, Georgia State University, Atlanta, GA, USA; dSchool of Economics and Management,
Tsinghua University, Beijing, China

ABSTRACT KEYWORDS
While prior research has established that information technology (IT) IT investment; technological
investment has a significant impact on firm performance, relatively choices; performance
few studies have provided insights into the antecedents of IT invest­ shortfalls; corporate
governance; digital
ment decisions. By integrating the behavioral theory of the firm and
innovation; behavioral
agency theory, we propose a behavioral agency theory to explain theory of the firm; agency
performance shortfalls and corporate governance, which monitors theory; behavioral agency
and controls managers’ tendency of overinvestment or underinvest­ theory
ment in IT, as key drivers that jointly determine IT investment. As
such, IT investment facilitates a firm’s problemistic search that gen­
erates innovation in response to performance gaps. We further exam­
ine the role of innovation outputs as a mediating mechanism linking
IT investment to firm performance. Our econometric analysis of
a large-scale panel dataset provides empirical evidence corroborating
our theory. Overall, this study contributes a behavioral agency theory
to deepen our understanding about performance drivers and out­
comes of IT investment decisions.

Introduction
We develop a behavioral agency theory to examine key drivers and outcomes of informa­
tion technology (IT) investment decisions. We are motivated to do so as firms have been
allocating substantial financial resources to IT making it important to understand how
firms decide the extent of investment in IT [24]. In light of the remarkable amount of
investment in IT we have witnessed in the past decades, substantial research has been
directed at investigating the business value of IT investment. Prior research, however, has
predominately focused on the impact of IT investment on firm performance and has
almost taken the level of IT investment as a “given.” Research on key drivers of IT
investment decisions is sparse and is even labeled as a “blind spot” [3].
It is important to fill this gap because, theoretically, how firms make decisions to
allocate substantial financial resources to various investments (e.g., IT, R&D, etc.) can
influence their performance [31, 99]. Prior studies have found that the performance
impact of IT investment, while positive on average, varies notably across firms [75, 81].
On the one hand, rapid advances in IT may have pushed some firms to chase new

CONTACT Arun Rai [email protected] Center for Digital Innovation & Department of Computer Information
Systems, J. Mack Robinson College of Business, Georgia State University, Atlanta, GA 30303, USA
© 2021 Taylor & Francis Group, LLC
30 DONG ET AL.

technologies [5] and even overspend on them [4, 35, 57]. On the other hand, some firms
fail to keep up with technological progress and underinvest in IT [94, 111]. If firms’ IT
investment decisions are not efficient, it is important to ask how IT investment decisions
are determined; in what circumstances are firms likely to overinvest or underinvest in IT;
what mechanisms can mitigate IT overinvestment or underinvestment tendency; and,
through what mechanisms does the IT investment made by firms in turn affect long-
term performance. Answers to these related questions are important for research and
practice on business value of IT.
We seek to answer these questions by proposing a behavioral agency theory to under­
stand firms’ decisions with regard to the level of IT investment. Our behavioral agency
perspective has two theoretical anchors [46, 92]. The first one is the behavioral theory of
the firm (BTOF), which suggests that performance shortfalls (i.e., declines in performance
relative to aspiration level — the reference point defining success and failure) motivate
firms to search for solutions to address performance gaps [6, 28]. The search processes
directed toward solutions to specific performance problems is called problemistic search
[95]. Recently, problemistic search has attracted attention in the information systems (IS)
field. For example, Karhade and Dong focus on identifying specific problem-solving
activities that a firm employs to characterize problemistic search [67]. In contrast, our
emphasis is not on discerning search activities to describe problemistic search, but on
problemistic search as an overarching theoretical mechanism to explain the causal rela­
tionship between performance shortfalls and IT investment, which, in turn, generates
innovation that improves performance of the firm.
BTOF, however, has largely omitted the incentive issues that decision makers of firms
may have. Agency theory, as the second cornerstone in our theorizing, suggests that
corporate governance is necessary to address the agency problems of decision makers of
a firm, as their incentives may not always be aligning with the interests of firm owners [40,
63]. As firms’ IT investment is large in recent years—IT spending was projected at $3.5
trillion worldwide, about 8 percent of company revenue in 2019 [59]—the corporate board
of directors plays an important role in overseeing IT investment. For example, industry
press shows that 35 percent of directors spent 6-10 percent of annual board hours
discussing IT-related topics [97], and the allocation of technology investment was an
important topic for board discussions [30]. In particular, 50 percent directors on board
elaborated returns on IT investment in comparison to predetermined targets in board
meetings and used these measures to justify IT budgets [27]. In the same line, the
academic literature has suggested that directors discuss and oversee IT investment [e.g.,
12, 87], and documented empirical evidence on the influences of an independent, active
board on IT investment decisions [e.g., 24, 58]. As a result, in the context of IT invest­
ment, overinvestment and underinvestment are possible if strong corporate governance is
absent and decision makers are not overseen [57].
By integrating insights from BTOF and agency theory, we propose that performance
shortfalls and IT investment tendency (that is, a firm’s overall gesture of overinvestment in
IT relative to underinvestment in IT due to agency problems) jointly determine a firm’s
level of IT investment, conditional on corporate governance. On the one hand, when
corporate governance is strong and decision makers are closely monitored, overinvest­
ment or underinvestment in IT is less likely to occur when the firm increases IT invest­
ment in response to performance shortfalls. On the other hand, when corporate
 JOURNAL OF MANAGEMENT INFORMATION SYSTEMS 31

governance is weak and decision makers are not effectively monitored, overinvestment or
underinvestment in IT is more likely to exist when the firm increases IT investment in
response to performance shortfalls. Our efforts to integrate two complementary theoretical
perspectives can contribute to a more holistic picture of how IT investment decisions are
made by firms.
In addition, we draw on BTOF to close the loop of IT investment and firm performance
by linking a firm’s IT investment to its innovation outputs and subsequently to its long-
term performance. Given the rich evidence supporting the benefits from IT investment to
facilitate innovation [e.g., 10, 45, 73, 98, 118], we conceptualize IT investment as a key
resource investment in support of problemistic search leading to innovation outputs,
which is a key intermediate outcome that channels the performance impact of IT invest­
ment. In particular, Karhade and Dong highlight that IT investment enables innovation
while inducing dynamic adjustment costs [66], calling for a more comprehensive under­
standing of the overall impacts of IT investment on innovation and performance [66].
While the burgeoning literatures on digital innovation and business value of IT shows
direct links from IT investment to innovation and performance respectively, we integrate
separate research streams by showing the mediating role of innovation in the performance
impact of IT investment. We collected a large-scale panel dataset from 1833 firms in 2001-
2005 to test a parsimonious set of integrative hypotheses about the moderating roles of
a tendency to overinvest/underinvest in IT and corporate governance, and the mediating
role of innovation outputs to discover evidence for our theorized claims.
Our contributions to the IS literature are twofold. First, we propose that in isolation,
BTOF alone is not sufficient in explaining a firm’s IT investment decisions as it assumes
quasi-resolution of conflicting interests among decision makers within the firm; that is,
decision makers’ conflicting interests can be at least partly reconciled and unified [28].
However, from a BTOF lens it is unclear as to how such reconciliation is achieved. By
integrating BTOF with agency theory, we theorize that while IT investment in response to
performance shortfalls represents the boundedly rational choice for managers [3, 101], the
degree to which their conflicting interests with firm owners is at least partly resolved by
corporate governance leading to IT investment decisions that comply with or deviate from
that boundedly rational choice. Our empirical evidence suggests the high prevalence of
overinvestment in IT when firms respond to performance shortfalls and surface the role of
corporate governance in controlling an overinvestment tendency, demonstrating the
utility of the integrated behavioral agency perspective. Second, we extend prior research
on digital innovation and business value of IT by showing IT investment leads to
innovation outputs that in turn improve long-term performance. By doing so, we provide
empirical evidence in support of a more holistic picture of business value of IT in which
innovation outcomes channel the performance impact of IT investment by addressing
performance shortfalls.

Theoretical Background
What We Have Known: A Behavioral Theory of IT Investment Decisions
Our theoretical foundation is BTOF, developed by Cyert and March [28]. BTOF chal­
lenges the assumptions of neoclassical economics that managers have perfect information
32 DONG ET AL.

and maximize performance, and proposes that they aim at satisficing rather than optimiz­
ing in decision making. BTOF builds on the idea of bounded rationality developed by
Simon [105]. Specifically, boundedly rational decision makers of a firm seek satisfaction by
comparing firm performance with a “good-enough” reference point, that is, aspiration
level. Dissatisfaction stems from performance shortfalls when firm performance decreases
relative to aspiration level, which, in turn, drives firms to invest more resources for
problemistic search to address performance gaps. For example, undesirable products can
reduce firm performance, leading to performance shortfalls when performance decreases
relative to aspiration level. To close this performance gap, the firm can invest in R&D and
IT as resource inputs to problem-solving activities, which develop innovation that
improves firm performance [10, 73, 98]. This underlying logic explains how organizational
decisions are made and suggests a negative relationship between increases in performance
relative to aspiration level and resource investments. In the context of IT investment
decisions, Salge, Kohli, and Barrett show that declines in performance relative to aspira­
tion level lead to increasing IT investment based on a sample of 153 U.K. hospitals [101].
More recently, Anand, Sharma, and Kohli demonstrate that seven U.S. hospitals increas­
ingly use business intelligence and analytics in response to performance shortfalls [3].
BTOF examines organizational decision making based on the concept of satisficing. It
argues that decision makers of a firm do not aim to optimize but seek results that “satisfy
and suffice” [105]. Their satisfaction with firm performance depends on aspiration level at
the organizational level, defined as the organizational goal or threshold they use to
determine the boundary between success and failure when evaluating firm performance.
As firm performance decreases relative to aspiration level, managers become increasingly
dissatisfied with performance, which triggers search processes for solutions to address
performance shortfalls. Following the aforementioned discussion, performance shortfalls
are often operationalized, reversely, as firm performance minus aspiration level [e.g., 18,
48]. Thus, the higher is the performance relative to aspiration level, the smaller are the
performance shortfalls.
As managers are boundedly rational and the innovative solutions to address perfor­
mance shortfalls might not be easily apparent, firms need to engage in a search for
solutions. In other words, failures to obtain satisfactory performance relative to aspiration
level or performance shortfalls are triggers of problemistic search [95]. In BTOF, proble­
mistic search is defined as problem-solving activities that are directed toward solutions to
specific problems [28, 48]. As behavioral consequences of problemistic search, innovation
resulting from problemistic search may resolve performance shortfalls and improve long-
term performance [48, 95]. Key investments, such as R&D and IT investments [10, 73],
which are uncertain in return but render upside potential to create innovative solutions,
have been considered as organizational changes due to problemistic search [53, 95].
Drawing on prior literature that conceptualizes IT investment as a key resource invest­
ment for problemistic search [e.g., 67, 101], we view IT investment as a resource input to
the problemistic search.

What We Need to Know: A Behavioral Agency Theory of IT Investment Decisions

To achieve a unified aspiration at the organizational level based on diverse goals among
decision makers of a firm, BTOF assumes quasi-resolution of conflicting interests for the
 JOURNAL OF MANAGEMENT INFORMATION SYSTEMS 33

decision makers [28]. However, agency theory suggests that it is not always possible to
resolve conflicting interests between decision makers and firm owners, leading to agency
problems in firm decisions [40, 63]. In other words, decision makers of a firm (especially
a public corporation with separate administration and ownership) tend to make decisions
for their own interests, rather than for the interests of firm owners. Agency theory
maintains that interests of firm owners may differ from that of managers who make
investment decisions, and that the divergent goals of these two kinds of decision makers
could lead to agency problems with managers behaving in a manner inconsistent with the
interests of the firm owners [40, 63].
In the presence of agency problems, firms are likely to overinvest or underinvest in IT
[24, 58, 77]. Managers may tend to overinvest in IT because of the “empire-building”
incentive, motivating them to invest unnecessarily and excessively in, for instance, IT
labor to increase their managerial power [e.g., 13, 104, 107]. Managers may also have
a tendency of underinvestment in IT, because it allows them to exert less effort and shirk
[e.g., 32, 84, 90].
From an empirical viewpoint, recent IS research uses the industry average IT invest­
ment as the benchmark and suggests that some firms are likely to overinvest in IT when
their investment is above the benchmark whereas other firms are likely to underinvest
when their investment is below the benchmark [57]. Firms may arguably have distinct IT
needs [101, 108], and, therefore, deviations from the average of an industry may not
always be overinvestment or underinvestment in IT. However, the industry average IT
investment is a meaningful benchmark to reflect the tendency of firms’ IT investment [82].
We follow this approach to conceptualize IT investment tendency, indicating if a firm’s IT
investment is above the industry average IT investment (a tendency to overinvest in IT
= 1) or below the industry average IT investment (a tendency to underinvest in IT = 0).1
The logic is that a firm’s IT investment far behind competitors’ investment can rarely be
an overinvestment. Similarly, it is less likely that a firm can underinvest in IT if its
investment is much more than the investment made by other firms in the same industry.
Our conceptualization of a firm’s investment tendency specific to IT investment is in the
same line with the agency theory literature, which suggests that agency problems can drive
decision makers to either overinvest or underinvest financial resources for their self-
interests instead of the interests of firm owners [e.g., 32, 84, 89].
When there is an overinvestment or underinvestment tendency, corporate governance
is needed to monitor the decisions made by managers in a firm, entailing different
arrangements such as organizing a large board of directors, including external directors
who are more independent, avoiding duality of CEO and the chairperson of the board,
and requiring attendance of directors in board meetings [11, 16, 62]. When a firm exhibits
a tendency to overinvest in IT (e.g., above the industry average IT investment), corporate
governance is likely to reduce overinvestment leading to a smaller increase of IT invest­
ment in response to performance shortfalls. When a firm tends to underinvest in IT (e.g.,
below the industry average IT investment), corporate governance is likely to mitigate
underinvestment resulting in a greater increase of IT investment in response to perfor­
mance shortfalls.
In sum, it is important to relax the assumption of BTOF and consider possible agency
problems in IT investment decisions by integrating the behavioral perspective with the
agency perspective. By doing so, we not only examine key drivers of IT investment
34 DONG ET AL.

Agency Theory

IT Investment
tendency
(t – 1)

Corporate
governance
(t – 1)

Performance Innovation Long-term

IT investment
shortfalls outputs performance
(t)
(t – 1) (t + 1) (t + n)

Behavioral Theory of the Firm

Figure 1. An integrative model for behavioral agency theory.

decisions as informed by BTOF [e.g., 3, 101] but also address widespread agency problems
that plague IT investment decisions. Figure 1 presents an overview of our research model,
entailing a behavioral agency theory of IT investment decisions. Next, we develop asso­
ciated hypotheses with a focus on the moderating roles of IT investment tendency and
corporate governance, as well as the mediating role of innovation outputs, which reflect
our major contributions to filling the gaps in extant literature.

Hypotheses Development
Integrating Behavioral and Agency Drivers of IT Investment
BTOF argues that performance shortfalls, manifested performance decreases relative to
aspiration level, trigger resource investments [43, 48]. Salge, Kohli, and Barrett find that
when hospitals face decreases in performance relative to aspiration level, they increase IT
investment as a response [101]. Similarly, Anand, Sharma, and Kohli show that hospitals
increase their use of business intelligence and analytics when experiencing lower perfor­
mance relative to aspiration level [3]. For profit-seeking firms, we propose that when
performance decreases relative to aspiration level (i.e., performance shortfalls), IT invest­
ment is likely to increase.2 IT investment decisions, however, are also subject to agency
problems [24, 50, 58].
From a behavioral agency perspective, decision makers of a firm can make suboptimal
decisions leading to overinvestment and underinvestment in IT, because of their empire-
building or shirking incentives (in the absence of effective corporate governance). When
a firm tends to overinvest in IT (IT investment tendency = 1), decision makers are likely to
use performance shortfalls as an excuse to overinvest in IT and build a larger IT department,
such as hiring more IT personnel to increase their managerial power [58]. On the one hand,
this implies a stronger relationship between performance shortfalls and IT investment when
the firm has such an investment tendency. On the other hand, when a firm tends to
 JOURNAL OF MANAGEMENT INFORMATION SYSTEMS 35

underinvest in IT (IT investment tendency = 0), decision makers are likely to neglect
performance shortfalls and underinvest in IT. Since IT investment is very risky [34, 112],
managers of a firm tending to underinvest in IT may shirk and invest less than needed in
response to performance shortfalls by taking steps like cutting IT budgets and downsizing
the IT workforce. This indicates a weaker relationship between performance shortfalls and
IT investment when the firm tends to underinvest in IT. In sum, the negative relationship
between performance shortfalls and IT investment could be strengthened (weakened) by the
tendency to overinvest (underinvest) in IT. Therefore, we hypothesize:
Hypothesis 1 (H1): IT investment tendency moderates the negative relationship between
increases in performance relative to aspiration level and IT investment, such that the negative
relationship becomes stronger (weaker) if there is a tendency to overinvest (underinvest) in IT.

Agency theory suggests that corporate governance can mitigate IT overinvestment

problems by monitoring managers’ investment decisions. For example, board directors
can directly monitor the decisions for large-scale IT investment in regular board meetings
by reviewing and auditing financial statements [24]. It is more difficult for managers who
tend to overinvest to pursue empire-building motives when their IT investment decisions
are closely monitored. When corporate governance is strong, managers are more likely to
appropriately invest in IT in response to performance shortfalls. Similarly, corporate
governance can also address underinvestment problems by monitoring managers’ invest­
ment decisions. For example, a more independent board with external directors is likely to
question insufficient IT investment decisions made by managers. External directors are
equipped with knowledge outside the firm (e.g., investment practices and performance
trends of other firms in the industry), and can thereby identify new technological
opportunities that managers have not captured [58]. Thus, when managers are closely
monitored, it is more difficult for them to intentionally avoid necessary investment in IT.
In summary, in the presence of stronger corporate governance that monitors managers’
decisions, a tendency to overinvest in IT could be mitigated resulting in a more appro­
priate level of IT investment in response to performance shortfalls. Similarly, in the
presence of stronger corporate governance, a tendency to underinvest in IT could also
be mitigated resulting in a more appropriate level of IT investment in response to
performance shortfalls. Thus, we have the following hypothesis:
Hypothesis 2 (H2): The moderating effect of IT investment tendency on the negative relation­
ship between increases in performance relative to aspiration level and IT investment is less
outspoken when there is stronger corporate governance, such that corporate governance makes
the negative relationship stronger (weaker) to a smaller extent if there is a tendency to over­
invest (underinvest) in IT.

The Mediating Role of Innovation Outputs in the Performance Impact of IT

Investment
To address performance shortfalls, firms need to increase IT investment (and other
investments such as R&D) to enable problemistic search activities [67]. These activities
generate innovation that addresses performance gaps [48, 95]. IT investment can facilitate
problemistic search leading to innovation outputs for three reasons. First, IT investment
can improve knowledge management within a firm, and effective knowledge management
36 DONG ET AL.

can improve problem-solving activities pertinent to innovation. For instance, Alavi and
Leidner suggest that knowledge management systems facilitate knowledge creation, sto­
rage, and retrieval [2]. Forman and van Zeebroeck find that IT investment is associated
with increasing intraorganizational collaboration among knowledge inventors in firms
[42]. Joshi et al. document that IT use enables firms to develop knowledge management
capabilities, leading to new patents and products [64]. Firms’ IT investment has also been
found to be a key resource input to knowledge production with patent outputs [73, 118].
Recently, Dong and Yang showed that IT use supports firms’ knowledge recombination,
thereby expanding their patent outputs [37].
Second, IT investment can aid innovation through new product development by
supporting research and design of new products and services that either create new
markets or effectively respond to customer needs. For example, computer-aided design
is useful for conceiving and evaluating new product designs [73], and project management
systems facilitate task scheduling and resource allocation for complex new product devel­
opment projects [91]. In new product development processes, product quality can be
enhanced while cycle time and development costs can be reduced by IT use [9, 110]. IT
investment can mitigate the diminishing returns on R&D investment, generating more
patents with high commercial value [98]. IT investment can also benefit the commercia­
lization of new products and services [64, 66].
Third, IT investment can facilitate external knowledge search as well as knowledge
sharing between a firm and its business partners. Conducting a firm’s problemistic search
in collaboration with its suppliers and customers is vital for the focal firm’s innovation
outcomes [67]. Malhotra, Gosain and El Sawy find that standardized interfaces of inter-
organizational systems support a firm’s knowledge creation with its supply chain partners
[78, 79]. Relatedly, customer relationship management system use leads to process cou­
pling and knowledge sharing between a focal firm and its customers, leading to innovative
solutions [100, 102]. External knowledge search, enabled by IT investment for digital
platforms that facilitate new forms of transacting [70], is productive for a firm to generate
process innovation in supply chains [45, 113].
Following the aforementioned three reasons, IT investment is vital for problemistic
search and is likely to address performance shortfalls via developing innovation [38].
Innovation outputs resulting from problemistic search become the intermediate channel
through which IT investment is deployed to improve performance in the long run,
implying a mediating path from IT investment to innovation outputs to long-term
performance [60]. In other words, IT investment driven by performance shortfalls sup­
ports problemistic search leading to increased innovation outputs, which, in turn, improve
firm performance in the long run. Therefore, we hypothesize:
Hypothesis 3 (H3): Innovation outputs mediate the positive relationship between IT investment
and long-term performance.

Methodology
Data
We constructed a large-scale panel dataset from public U.S. firms based on five archival
sources. First, we obtained data about IT investment from the Harte Hanks’ Computer
 JOURNAL OF MANAGEMENT INFORMATION SYSTEMS 37

Intelligence (CI) database, which has been widely used in prior IS research [e.g., 25, 34,
73, 118]. Our sampling frame included Fortune 1000 firms starting from year 2001.
Second, we matched CI data with the Standard and Poor’s Compustat database. From
the Compustat database, we obtained financial data for firm performance and several
key control variables. Third, we obtained additional variables related to corporate
governance from the Investor Responsibility Research Center (IRRC) Directors data­
base and the Corporate Library database. Finally, we merged our data with the National
Bureau of Economic Research’s (NBER) Patent Citations database. Hall, Jaffe and
Trajtenberg developed this database based on all utility patents granted by the U.S.
Patent and Trademark Office (USPTO) between 1976 and 2006 [52]. This database
contains detailed information about more than 3 million patents, which are widely used
as a major data source of innovation [e.g., 1, 55, 73, 116, 118]. One advantage of using
this database is the availability of a dynamic match file linking patent assignee names to
firm identifiers in the Compustat database. We aggregated patent-level count data to
the assignee level and then to the firm level [52]. Since we need one-year time lag
between innovation outputs and other variables, we ended our panel in 2005. After
merging the above five sources of data, our final sample contained an unbalanced panel
structure from 1833 unique firms between 2001 and 2005 with 7237 firm-year observa­
tions. Depending on the variables included in regression analysis, some variables
contain missing values that lead to dropped observations.

Measures
Performance Relative to Aspiration Level
In line with the BTOF literature, we used performance relative to aspiration level to
capture performance shortfalls [18, 48, 101]. Increases in performance relative to aspira­
tion level indicate smaller performance shortfalls. We followed the BTOF literature to use
a profitability measure—return on assets (ROA), defined as operating income before
depreciation over total assets—as the measure of firm performance [e.g., 22, 23, 48,
61].3 ROA is the most widely used performance measure in the BTOF literature [18]. It
is particularly appropriate because (1) ROA is not sensitive to a firm’s capital structure
compared to other profitability measures like return on equity (ROE); and (2) rather than
relying on revenue or sales, BTOF is mostly concerned with profitability in terms of
utilization of invested assets [23, 61].
To measure aspiration level, we considered two different aspirations of performance as
proposed by BTOF: historical and social [28, 36]. We used (1) ROA in year t – 1 to
construct the proxy for historical aspiration in year t [23, 61]4, and (2) the median of ROA
in a Standard Industrial Classification (SIC) four-digit industry in year t – 1 as the proxy
for social aspiration in year t [23, 61]. Following a recent literature review on and
empirical comparison of different aspiration measures [18], we employed the switching
approach to construct a single measure for aspiration level (see the Appendix for
a detailed explanation). Specifically, aspiration level is equal to social aspiration if the
focal firm’s ROA is below social aspiration. When a focal firm’s ROA is above or equal to
social aspiration, aspiration level is slightly higher than the focal firm’s previous ROA (i.e.,
historical aspiration). The rationale is that firms need to catch up to industry peers if they
are lagging behind industry peers, but will continuously improve their own performance if
38 DONG ET AL.

Table 1. Summary of measures.

Construct Measure References
Performance ROA minus aspiration level Bromiley and Harris (2014) [18], Chen and
relative to Miller (2007) [23], Iyer and Miller (2008)
aspiration level [61]
IT investment A binary indicator equal to 1 if a firm’s IT investment Ho et al. (2017) [57]
tendency is greater than the industry average IT investment (a
tendency to overinvest in IT) and 0 otherwise (a
tendency to underinvest in IT)
Corporate Corporate governance index as the average Gillan et al. (2011) [44], Karuna (2009) [71]
governance percentile ranks of 1) board size, 2) board
independence, 3) inverse CEO-chairperson duality,
and 4) director attendance of meetings
IT investment IT labor expenditure scaled by total assets Tambe and Hitt (2012) [109], Tambe et al.
(2012) [110]
Innovation outputs The natural logarithm of the number of granted Gómez et al. (2017) [45], Joshi et al. (2010)
patents subtracted from the industry average [64], Saldanha et al. (2017) [100], Xue et al.
number of granted patents (2012) [118]
Long-term Tobin’s Q as firm market value over book value Bardhan et al. (2013) [10], Bharadwaj et al.
performance (1999) [14]
Deviation from A firm’s IT investment minus the industry average IT Ho et al. (2017) [57], Mithas et al. (2013)
industry average investment [82]
IT investment
R&D capital R&D expenditure or capital flow over total assets Brynjolfsson and Hitt (2003) [19], Hall et al.
(2005) [51]
Absorbed slack Selling, general, and administrative expenses over Greve (2003) [48], Iyer and Miller (2008)
total sales [61]
Unabsorbed slack Current assets divided by current liabilities Chen and Miller (2007) [23], Iyer and Miller
(2008) [61]
Potential slack Long-term debt over total assets Greve (2003) [48], Iyer and Miller (2008)
[61]
Diversification One minus the sum of squares of employment Hitt (1999) [56], Kobelsky et al. (2008) [74]
percentages in different industries
Firm growth Percentage of change in total sales from the Kobelsky et al. (2008) [74], Mitra (2005) [83]
previous year t – 1 to current year t
Firm size The natural logarithm of total assets Kobelsky et al. (2008) [74]
Notes: ROA = return on assets; IT = information technology.

they already outperform industry peers [18]. We followed prior studies [17, 18] to multi­
ply a focal firm’s previous ROA by 1.05 (i.e., an improvement rate of 5 percent) as its
historical aspiration.5 Finally, performance relative to aspiration level was computed as
ROA minus aspiration level [18].

IT Investment Tendency
We assessed a firm’s annual IT investment relative to the industry average IT investment
to reflect the firm’s tendency to overinvest or underinvest in IT [57]. We calculated the
industry average IT investment based on SIC three-digit industries (given the small
number of observations in some SIC four-digital industries) and create a binary indicator
with a value of 1 if a firm’s IT investment is above the industry average IT investment (i.e.,
a tendency of overinvestment in IT), and 0 otherwise (a tendency of underinvestment
in IT).

Corporate Governance
To capture corporate governance in a comprehensive manner, we followed the corporate
governance literature to create an index measure based on four variables pertinent to the
 JOURNAL OF MANAGEMENT INFORMATION SYSTEMS 39

monitoring functions of corporate board, including (1) board size, (2) board indepen­
dence, (3) CEO-chairperson duality, and (4) meeting attendance of directors [44, 71].
Board size is the total number of directors on board [119]. Board independence is the
percentage of directors on board from outside of the firm [11, 58]. CEO-chairperson
duality is measured by a binary variable indicating whether the CEO is the chairperson of
board, which is then reversely coded to be consistent with other variables [16, 41].
Meeting attendance of directors is measured by a binary variable indicating whether all
directors met the firm standards of attending board meetings [24, 62]. We constructed the
measure of corporate governance by computing the average of percentile ranks of these
four variables.

IT Investment
Recent IS studies suggest that a labor-based approach is better than a hardware-based
approach to improve the measurement of firms’ IT investment [109, 110]. Hardware
expenses compose only 10-15 percent of a firm’s total IT investment [20], and these
expenses have been reducing over recent years [103]. Instead, labor costs account for more
than 30 percent of a firm’s IT budgets, while other components of IT budgets, including
training and services, are more closely related to IT labor [109]. Thus, we used IT labor
expenditure to proxy for a firm’s annual IT investment.
This labor-based measure is particularly suitable for our study for two reasons. First, IT
labor expenditure reflects the resource flow invested by a firm in a particular year, rather
than IT hardware as resource stock that is consumed by a firm for several years [73], and
is thereby better aligned with the theoretical prediction of BTOF about adjustments of
resource flow in the short run based on performance feedback. Second, a measure related
to IT labor is also more consistent with our theoretical arguments based on agency theory,
as managers may overinvest or underinvest in IT labor because this is a more practical and
easier way to increase their managerial power or shirk compared to spending on or
adjusting IT hardware. Following the IS literature, we not only counted the number of
IT employees but also calculated IT labor expenditure by multiplying number of IT
employees with the average compensation per industry from the U.S. Bureau of Labor
Statistics (BLS) [76, 109]. We deflated the nominal value with reference to 2001 by using
the index for total compensation cost from BLS and, then, scaled the IT labor expenditure
by total assets to calculate the measure of IT investment.

Innovation Outputs
Following the digital innovation literature, we used the number of patents as a proxy of
innovation outputs [e.g., 37, 45, 64, 73, 98, 100, 118]. The natural logarithm was used to
reduce skewness of the number of patents. Given big differences in patenting traditions
across industries, we subtracted the industry average number of patents from each firm’s
number of patents. Thus, our patent measure has partialled out industry differences of
patenting. Since a patent by definition describes a technical problem and a solution to that
problem [49, 115], patents are essentially technical solutions as a result of problemistic
search that firms develop to solve their business problems causing unexpected perfor­
mance such as undesired product features [67, 72]. Therefore, patents indicate the amount
of successful solutions generated by a firm from problemistic search [47], and provide us
40 DONG ET AL.

a suitable measure for the consequences of problemistic search to evaluate the influence of
the triggers of problemistic search on the consequences.
Our approach is consistent with past work on problemistic search that has employed
research designs based on archival data sources where triggers and consequences of
problemistic search are likely to be observable, but the particular problemistic search
activities are typically not. A recent literature review on problemistic search indicates that
prior studies use such archival data to measure the triggers of problemistic search (e.g.,
performance shortfalls) and the behavioral consequences of problemistic search (e.g.,
patents or new products) [95], while employing problemistic search as an overarching
theoretical mechanism to explain the relationships.6

Long-Term Performance
BTOF suggest that firms adjust IT investment (and other investments such as R&D) in
support of problemistic search according to short-term performance feedback. The link
between innovative solutions to performance is however uncertain in the short run and,
therefore, it may take a long time to realize the value of IT investment [85]. This is
particularly true for IT investment that may be forward looking [e.g., 109]. Therefore, to
better capture the performance impact of IT investment, we followed prior IS studies [e.g.,
10, 14, 118] to use long-term performance measure, Tobin’s Q, to examine the perfor­
mance impact of IT investment. We calculated Tobin’s Q as the ratio of a firm’s market
value to its book value, where market value is the sum of market value of common equity,
liquidating value of preferred stock, current liabilities net of current assets and long-term
debt, and book value is total assets [122].

Control Variables
While we used a binary indicator to reflect firms’ tendency of IT overinvestment or
underinvestment and reduce measurement errors, we controlled the difference between
a firm’s IT investment and the average IT investment of other firms in a SIC three-digit
industry, which captures the actual deviation of a firm’s IT investment from the industry
average IT investment [57, 82]. R&D capital represents a firm’s absorptive capacity that is
critical for problemistic search and innovation [26]. We controlled R&D capital based on
annual R&D expenditure using the perpetual inventory method7 [51], scaled it by total
assets, and deflated R&D capital with reference to 2001 by using the gross domestic
product (GDP) deflator from the U.S. Bureau of Economic Analysis.
The BTOF literature suggests that it is crucial to include a fair set of controls for slack
(i.e., excess resources) as it triggers slack search, which is an alternative search mechanism
to problemistic search [28, 48].8 Bourgeois identifies three different types of slack
resources: absorbed, unabsorbed and potential slack [21], which have been widely used
in past research [e.g., 17, 48, 61]. Absorbed slack reflects organizational resources used for
general administrative purposes, which are not always necessary for the maintenance of
firm operations. We controlled absorbed slack by the ratio of selling, general, and
administrative expenses (SGAE) over total sales [48, 61]. Unabsorbed slack exists as
financial reserves, often measured by current ratio as current assets divided by current
liabilities [23, 61]. We controlled unabsorbed slack by current ratio. Higher debt and
interest payments can constrain a firm’s IT investment [74]. Potential slack exists if a firm
 JOURNAL OF MANAGEMENT INFORMATION SYSTEMS 41

borrows less from external financial sources than it potentially could borrow, manifested
by debt ratio as the long-term debt over total assets [48, 61]. We controlled potential slack
by debt ratio. Note that debt ratio is an inverse indicator of potential slack, as a lower debt
ratio reflects a higher level of potential slack.
Diversification influences a firm’s IT investment as it increases the need of coordination
across different lines of business [74]. We controlled diversification using a Herfindahl
index as one minus the sum of squares of a firm’s employment percentages in different
SIC four-digit industries [33, 56]. Sales growth rate refers to the annual percentage of
change in sales, which has been shown to influence a firm’s IT investment [74, 83]. We
controlled firm growth by sales growth rate. Firm size indicates the overall availability of
resources that can influence a firm’s IT investment; as a result, we controlled firm size by
the natural logarithm of total assets [74].
Table 1 describes our measures. Table 2 reports the descriptive statistics and correla­
tions. In addition to the control variables listed in Table 2, we also controlled firm fixed
effects and included SIC one-digit industry and year dummies in our analysis.

Results
Hypothesis Testing
To test H1 and H2, we used a fixed effects model to account for the panel structure of
our data and mitigate endogeneity of performance relative to aspiration level and
moderators. All other variables were one-year lagged to IT investment to avoid reverse
causality. To create the interaction terms, we mean-centered performance relative to
aspiration level and corporate governance, and used a binary measure to differentiate
between IT overinvestment tendency and IT underinvestment tendency as previously
described. We created three two-way interaction terms (performance relative to aspira­
tion level × IT investment tendency, performance relative to aspiration level × corpo­
rate governance, IT investment tendency × corporate governance) and one three-way
interaction term (performance relative to aspiration level × IT investment tendency ×
corporate governance). Table 3 reports the fixed effects regression results. After esti­
mating Model (1) with control variables only, we added performance relative to
aspiration level to Model (2). We found a statistically significant and negative effect
of performance relative to aspiration level on IT investment (β = -0.009, p < 0.01).
Consistent with prior findings from the non-profit hospitals context [e.g., 3, 101], this
result suggests that when performance relative to aspiration level decreases, firms
increase their IT investment. We then added IT investment tendency and its two-way
interaction term with performance relative to aspiration level to Model (3). We found
a statistically significant and negative interaction effect between performance relative to
aspiration level and IT investment tendency (β = -0.028, p < 0.001), suggesting that
firms make more (less) IT investment in response to performance relative to aspiration
level if there is IT overinvestment (underinvestment) tendency. Thus, H1 is supported.
Interestingly, after adding the two-way interaction term, performance relative to aspira­
tion level no longer demonstrated a significant main effect on IT investment. When there
is an underinvestment tendency, therefore, firms seem to not respond to performance
shortfalls by adjusting IT investment. This finding is in line with the BTOF literature,
 42
DONG ET AL.

Table 2. Descriptive statistics and correlations.

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14)
(1) Long-term performance
(2) Innovation outputs 0.03
(3) IT investment -0.00 -0.08
(4) Performance relative to aspiration level -0.99 0.00 0.00
(5) IT investment tendency 0.02 -0.14 0.34 0.01
(6) Corporate governance -0.06 0.13 -0.07 0.02 0.02
(7) Deviation from industry average IT investment -0.00 -0.07 0.89 -0.00 0.41 -0.03
(8) R&D capital 0.03 0.33 0.03 0.00 -0.04 -0.04 -0.00
(9) Absorbed slack 0.16 0.13 0.18 -0.13 0 -0.08 0.10 0.49
(10) Unabsorbed slack -0.00 0.03 0.12 0.01 0 -0.19 0.09 0.19 0.51
(11) Potential slack -0.02 -0.06 -0.07 0.01 -0.02 0.09 -0.02 -0.16 -0.11 -0.17
(12) Diversification -0.01 0.16 0.02 -0.00 0.04 0.16 0.02 -0.12 -0.12 -0.11 0.06
(13) Firm growth 0.11 -0.04 -0.05 0.23 -0.05 -0.05 -0.05 -0.06 -0.07 -0.03 -0.00 0.06
(14) Firm size -0.06 0.37 -0.26 0.08 -0.21 0.27 -0.17 -0.08 -0.17 -0.19 0.12 -0.00 -0.02
Obs. 6690 7237 7237 7237 7237 2308 7237 7210 7232 6762 7223 7237 7231 7237
Mean 1.25 0.34 0.01 -0.04 0.34 36.40 -0.00 0.02 0.20 2.15 0.21 0.35 0.07 7.08
SD 6.43 1.15 0.02 1.79 0.48 14.24 0.01 0.04 0.23 1.98 0.20 0.30 0.29 1.75
Notes: Correlations greater than 0.04 or smaller than -0.04 are significant at p < 0.05.
 JOURNAL OF MANAGEMENT INFORMATION SYSTEMS 43

Table 3. Fixed effects regression results.

(1) (2) (3) (4)
Performance relative to aspiration level -0.009** 0.003 0.000
(0.003) (0.004) (0.004)
IT investment tendency 0.000 0.000
(0.000) (0.000)
Performance relative to aspiration level × -0.028*** -0.020**
IT investment tendency (H1) (0.005) (0.007)
Corporate governance 0.000
(0.000)
Performance relative to aspiration level × -0.000
Corporate governance (0.000)
IT investment tendency × 0.000*
Corporate governance (0.000)
Performance relative to aspiration level × 0.003***
IT investment tendency × (0.001)
Corporate governance (H2)
Deviation from industry average IT investment 0.294*** 0.291*** 0.287*** 0.104***
(0.016) (0.016) (0.017) (0.022)
R&D capital -0.002 -0.005 -0.005 0.017
(0.015) (0.015) (0.015) (0.018)
Absorbed slack 0.007* 0.005 0.005 0.000
(0.004) (0.004) (0.004) (0.003)
Unabsorbed slack -0.001** -0.001** -0.001** -0.000
(0.000) (0.000) (0.000) (0.000)
Potential slack 0.001 0.001 0.001 0.005
(0.002) (0.002) (0.002) (0.003)
Diversification 0.005*** 0.005*** 0.005*** 0.001
(0.001) (0.001) (0.001) (0.001)
Firm growth -0.001 -0.000 -0.001 0.001
(0.001) (0.001) (0.001) (0.001)
Firm size -0.003*** -0.004*** -0.004*** -0.001
(0.001) (0.001) (0.001) (0.001)
Industry dummies Yes Yes Yes Yes
Year dummies Yes Yes Yes Yes
Constant 0.032*** 0.035*** 0.034*** 0.018
(0.009) (0.009) (0.009) (0.010)
F 22.820*** 22.170*** 21.480*** 4.700***
N 6117 6117 6117 2057
Notes: * p < 0.05; ** p < 0.01; *** p < 0.001. Standard errors are in parentheses. Dependent variable is IT investment in year
t + 1.

which documents that managers are more sensitive to performance shortfalls to increase
rather than decrease resource investments [48, 65].
We then added corporate governance, its two-way interaction term with performance
relative to aspiration level and IT investment tendency, as well as a three-way interaction
term to Model (4). We found that the interaction effect between performance relative to
aspiration level and IT investment tendency remains statistically significant and negative
(β = -0.020, p < 0.01), while the three-way interaction term is statistically significant and
positive (β = 0.003, p < 0.001). These results jointly suggest that although firms are more
likely to overinvest in IT when performance relative to aspiration level decreases, stronger
corporate governance makes this moderating role of IT investment tendency less out­
spoken, supporting H2. Consistent with our theory, our results indicate that stronger
corporate governance helps control overinvestment in IT.
We further plotted the three-way interaction effect in Figure 2, by defining high and
low levels of a variable as mean plus or minus one standard deviation. When performance
44 DONG ET AL.

IT investment in year t + 1
0.03

0.025

0.02

0.015

0.01
(1) IT overinvestment tendency, strong corporate governance
0.005 (2) IT overinvestment tendency, weak corporate governance
(3) IT underinvestment tendency, strong corporate governance
(4) IT underinvestment tendency, weak corporate governance
0
Low performance relative to aspiration level High performance relative to aspiration level
(i.e., large performance shortfalls) (i.e., small performance shortfalls)

Figure 2. Three-way interaction effect on IT investment.

Notes: Independent variable and moderators in year t.

relative to aspiration level decreases from a high level to a low level (i.e., small perfor­
mance shortfalls to large performance shortfalls), firms that have an IT overinvestment
tendency and strong corporate governance decrease IT investment. Thus, it seems that
effective corporate governance closely monitors IT investment decisions by controlling IT
budgets with a cut if performance gets worse and advising IT spending for competitive­
ness if performance is better. When performance relative to aspiration level decreases
from a high level to a low level (i.e., small performance shortfalls to large performance
shortfalls), firms with an IT overinvestment tendency and weak corporate governance
increase IT investment, which is likely to be an overinvestment when more earnings are
available. However, managers are not sensitive to performance shortfalls when making IT
investment decisions if there is an underinvestment tendency.
To test H3, we estimated simultaneous models with innovation outputs and long-
term performance as the consequences of problemistic search enabled by IT investment.
Since innovation outputs can impact long-term performance, the error terms of these
two equations are correlated. To correct for this correlation of error terms, we followed
the procedures used by Dewan and Ren when testing risk and return of IT investment
[34] and used seemingly unrelated regression [120], with innovation outputs as an
additional predictor of long-term performance. While seemingly unrelated regression
does not allow fixed effects, endogeneity is less a concern given that IT investment is the
endogenous variable in our research model.9 Again, IT investment and other variables
were one-year lagged to innovation outputs and long-term performance to avoid reverse
causality.
 JOURNAL OF MANAGEMENT INFORMATION SYSTEMS 45

Table 4. Seemingly unrelated regression results.

(1) (2)
DV: DV: DV: DV:
Innovation Long-Term Innovation Long-Term
Outputs Performance Outputs Performance
IT investment 10.491*** 4.392
(1.772) (3.567)
Innovation outputs 0.086*** 0.084***
(0.025) (0.025)
Deviation from industry average IT -0.174 7.668*** -10.282*** 3.436
investment (0.817) (1.635) (1.892) (3.806)
R&D capital 5.022*** 5.269*** 5.014*** 5.277***
(0.310) (0.634) (0.310) (0.634)
Absorbed slack 0.441*** 0.638*** 0.403*** 0.624***
(0.080) (0.160) (0.080) (0.160)
Unabsorbed slack 0.014 -0.016 0.014 -0.016
(0.007) (0.015) (0.007) (0.015)
Potential slack -0.344*** -0.089 -0.314*** -0.077
(0.057) (0.115) (0.057) (0.115)
Diversification 0.055 -0.341*** 0.022 -0.355***
(0.041) (0.082) (0.041) (0.083)
Firm growth -0.095** 0.234*** -0.088* 0.237***
(0.036) (0.073) (0.036) (0.073)
Firm size 0.265*** 0.043** 0.276*** 0.048**
(0.008) (0.017) (0.008) (0.017)
Industry dummies Yes Yes Yes Yes
Year dummies Yes Yes Yes Yes
Constant -1.883*** 0.922 -1.993*** 0.871
(0.264) (0.531) (0.264) (0.533)
Chi-square 2683.820*** 377.110*** 2733.610*** 378.720***
95 percent bias-corrected bootstrapping confidence interval (H3): [0.470, 1.322]
Notes: N = 6376. * p < 0.05; ** p < 0.01; *** p < 0.001. Standard errors are in parentheses.

Table 4 reports the seemingly unrelated regression results. After estimating Model (1)
with control variables only, we found that IT investment has a statistically significant and
positive effect on innovation outputs (β = 10.491, p < 0.001) in Model (2). Moreover, IT
investment does not directly influence long-term performance (β = 4.392, p > 0.05)10 and
innovation outputs have a statistically significant and positive effect on long-term perfor­
mance (β = 0.084, p < 0.001). We used a bootstrapping approach with 1000 times of
resampling—superior to the traditional Sobel test [96, 106]—to test the mediating effect of
innovation outputs and found that innovation outputs significantly mediate the relation­
ship between IT investment and long-term performance (95 percent bias-corrected boot­
strapping confidence interval: [0.470, 1.322]). Thus, H3 is also supported.

Robustness Checks
Sensitivity Analysis
First, we considered the possibility that IT investment decisions may be made in a forward-
looking manner. In other words, some investments in new technologies could be made for
developing sustainable competitive advantage and improving firm performance in the long
run. ROA as a short-term performance measure may not fully capture this motive. We
therefore replaced ROA by Tobin’s Q, which is a commonly used forward-looking perfor­
mance measure in the literature on business value of IT [e.g., 10, 14, 118], as an alternative
46 DONG ET AL.

performance measure to construct performance relative to aspiration level. Using Tobin’s

Q instead of ROA generated consistent results (see Table 5, Model (1)).
Second, we used the average of three-year ROA between year t – 3 and year t – 1 to
construct historical aspiration (see Table 5, Model (2)), which generated consistent results.
Similar to prior findings [88], our results are robust to the number of prior years’
performance used in formulating historical aspiration.
Third, we checked the sensitivity of our results to different improvement rates. We
followed the BTOF literature to change the improvement rate from 5 percent to 25 percent
and 50 percent [17, 18], which yielded consistent results [see Table 5, Models (3) and (4)].
Thus, our results are robust to the choice of improvement rate.

Ruling Out Alternative Explanation

Agency theory suggests two alternative mechanisms through which corporate governance
addresses agency problems: monitoring and incentive alignment [39, 55, 114]. Our
measure of corporate governance related to board of directors primarily addresses agency
problems based on monitoring. To rule out the alternative explanation of incentive
alignment, we controlled for incentive alignment. Incentive compensation plans (e.g.,
long-term incentive pay, bonuses, and options) are used by firm owners to align manage­
rial incentives with their own interests by making managers’ payments contingent on firm
long-term performance. Compensation plans with a larger proportion of long-term
incentive pay, bonuses, and options can align the interests of managers and firm owners,
thereby mitigating agency problems. CEO compensation plans have been documented to
influence IT outsourcing decisions [50].
We collected additional data for long-term incentive pay, bonuses, options, and total
compensation for all managers (e.g., CEO, CFO, COO, CIO, etc.) in each firm from the
Execucomp series of Compustat database. We then used the average proportion of long-
term incentive pay, bonuses and options in total compensation across all managers in each
firm to measure incentive alignment [8, 55, 80]. After controlling for incentive alignment,
we still found consistent results [see Table 5, Model (5)]. Thus, our measure of corporate
governance does not confound with the alternative explanation of incentive alignment.

Endogeneity Test
Our research model closes the loop of IT investment and firm performance, suggesting that
IT investment is endogenously determined by performance relative to aspiration level and
corporate governance given the tendency of IT overinvestment or underinvestment. To
address the endogeneity of IT investment, we followed the procedures used by Bharadwaj,
Bharadwaj, and Bendoly to estimate a Heckman two-stage model accounting for the endo­
genous propensity of the level of IT investment [15]. Specifically, we used the binary indicator
of IT investment tendency to reflect the level of IT investment: a value of 1 indicates high IT
investment greater than the industry average IT investment and a value of 0 otherwise.
In the first stage, we estimated a Probit model by regressing this binary indicator on
three exclusive restrictions from our research model (performance relative to aspiration
level, corporate governance, and their interaction term), as well as all control variables.
The Probit model demonstrated a good fit (Chi-square = 196.070, p < 0.001), allowing us
to calculate the inverse Mills ratio (IMR) representing the endogenous propensity of the
level of IT investment. In the second stage, we tested our model again with the IMR as an
 JOURNAL OF MANAGEMENT INFORMATION SYSTEMS 47

Table 5. Robustness checks.

(1) (2) (3) (4) (5) (6) (7)
Heckman Model
Tobin’s 25 Percent 50 Percent Control for Controlling for
Q Instead 3-Year Improvement Improvement Incentive Endogeneity of IT
of ROA ROA Rate Rate Alignment Investment
DV: DV: Long-
Innovation Term
DV: IT investment Outputs Performance
Performance relative -0.000 0.004 0.000 0.000 0.001
to aspiration level (0.000) (0.004) (0.004) (0.003) (0.004)
IT investment 0.001 0.000 0.000 0.000 0.000
tendency (0.001) (0.000) (0.000) (0.000) (0.000)
Performance relative -0.002*** -0.014* -0.016** -0.012* -0.020**
to aspiration level × (0.000) (0.006) (0.006) (0.005) (0.007)
IT investment
tendency
Corporate governance 0.000 0.000 0.000 0.000 0.000
(0.000) (0.000) (0.000) (0.000) (0.000)
Performance relative to 0.000 0.000 0.000 0.000 -0.000
aspiration level × (0.000) (0.000) (0.000) (0.000) (0.000)
Corporate
governance
IT investment 0.000 0.000* 0.000* 0.000* 0.000*
tendency × (0.000) (0.000) (0.000) (0.000) (0.000)
Corporate
governance
Performance relative 0.000+ 0.003*** 0.003*** 0.002*** 0.003***
to aspiration level × (0.000) (0.000) (0.000) (0.000) (0.001)
IT investment
tendency ×
Corporate
governance
Incentive alignment -0.001
(0.001)
IT investment 21.163*** -3.278
(5.198) (4.807)
Innovation outputs 0.067***
(0.020)
Inverse Mills ratio 1.431*** 1.146**
(0.439) (0.406)
Deviation from 0.093*** 0.118*** 0.100*** 0.096*** 0.103*** -20.728*** 0.006
industry average IT (0.023) (0.023) (0.022) (0.022) (0.022) (5.304) (4.903)
investment
R&D capital 0.035+ 0.021 0.019 0.022 0.021 6.576*** 2.008*
(0.019) (0.018) (0.018) (0.018) (0.019) (0.912) (0.850)
Absorbed slack 0.002 -0.000 0.000 0.000 0.000 0.488** 0.540***
(0.003) (0.003) (0.003) (0.003) (0.003) (0.167) (0.154)
Unabsorbed slack -0.000 -0.000 -0.000 -0.000 -0.000 -0.099** -0.066*
(0.000) (0.000) (0.000) (0.000) (0.000) (0.035) (0.032)
Potential slack 0.005+ 0.004 0.006* 0.006* 0.005+ -0.800 -1.847***
(0.003) (0.003) (0.003) (0.003) (0.003) (0.188) (0.174)
Diversification 0.000 -0.000 0.000 0.000 0.000 0.666** 0.124
(0.001) (0.001) (0.001) (0.001) (0.001) (0.217) (0.200)
Firm growth 0.000 0.000 0.001 0.001 0.001 -0.573*** 0.189
(0.001) (0.001) (0.001) (0.001) (0.001) (0.176) (0.162)
Firm size -0.001 -0.001 -0.001 -0.001 -0.001 0.144 -0.185*
(0.001) (0.001) (0.001) (0.001) (0.001) (0.097) (0.089)
Industry dummies Yes Yes Yes Yes Yes Yes Yes
Year dummies Yes Yes Yes Yes Yes Yes Yes
Constant 0.014 0.014 0.018+ 0.017+ 0.017+ -2.451*** 1.680***
(0.010) (0.010) (0.010) (0.010) (0.010) (0.540) (0.499)
F/Chi-square 3.100*** 4.180*** 4.680*** 4.320*** 4.510*** 1505.940*** 565.020***
N 2025 2056 2057 2057 2050 2111 2111
+
Notes: ROA = return on assets. p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001. Standard errors are in parentheses.
48 DONG ET AL.

additional control. The IMR is statistically significant, suggesting the necessity to address
the endogeneity of IT investment. After controlling for the IMR and addressing the
endogeneity, consistent results were found [see Table 5, Models (6) and (7)]. Thus,
endogeneity is not a serious concern for our results.

Discussion and Conclusion

Implications for Theory
Implication 1: While a behavioral perspective is useful to understand IT investment decisions,
complementing it with an agency perspective relaxes assumptions and enriches the theoretical
explanations for organizational decision making for IT investment.

Our findings suggest that (1) firms respond to performance shortfalls by increasing IT
investment, (2) IT investment made by firms in response to the performance shortfalls is
subject to agency problems—leading to overinvestment in IT, and (3) corporate govern­
ance reduces overinvestment in IT—leading to a smaller increase in IT investment in
response to the performance shortfalls. These findings jointly depict a behavioral agency
aspect of firms’ IT investment decisions.
Complementing past research on IT investment decisions [e.g., 68, 101], the wide
coverage of our sample across industries and over years should increase the confidence
of future research to employ a BTOF lens to understand IT investment decisions. More
importantly, by integrating BTOF and agency theory, we extend the behavioral perspective
into a behavioral agency perspective by simultaneously considering performance shortfalls
and agency problems, as well as regulation of these problems through governance, as key
drivers of IT investment decisions. An important implication is that future research needs
to address not only limitations stemming from decision makers’ bounded rationality but
also examine problems stemming from their incentives to pursue self-interest. Addressing
both issues provides a more complete explanation for firms’ decisions associated with the
allocation of IT investment.
Applying these findings to a broader theoretical perspective, we suggest that future
research use a BTOF lens to examine decision making for IT management across different
domains of IT decisions. BTOF maintains that managers at different levels are concerned
about different goals, and thus triggers for their decisions must be contextualized to the
specific management issues they are scrutinizing [28]. There is a long stream of research in
the IS literature on technology adoption at the organizational level. Different applications
support different processes; therefore, when managers are concerned about performance
shortfalls in specific value chain processes (e.g., marketing and sales), they may resort to
investing in specific technologies (e.g., customer relationship management and big data
analytics) for problemistic search addressing performance shortfalls [3].
Implication 2: IT investment supports problemistic search leading to innovation outputs,
through which firms increase long-term performance. Incorporating recent research on digital
innovation can deepen the understanding of business value of IT.

Our findings suggest that IT investment supports the search for innovation to solve
performance problems and improve firm performance in the long term. While prior
 JOURNAL OF MANAGEMENT INFORMATION SYSTEMS 49

studies have found support for a direct link between IT investment and innovation [e.g.,
45, 73, 98] and long-term performance measured by Tobin’s Q [e.g., 10, 14, 118], these
two research streams are largely disconnected. To the best of our knowledge, this study is
the first showing that innovation outputs serve as a mediating mechanism challenging the
performance impact of IT investment. We elaborate on the role of IT investment as a key
enabler of innovation and performance, as a result of problem-solving activities or
problemistic search addressing performance shortfalls. Overall, our findings answer the
call for research to unveil the mechanisms through which IT investment influences firm
performance [81]. It also has broader implications for future research on IT investment
and long-term performance.

Implications for Practice

Our research highlights the important roles of performance shortfalls and corporate
governance in driving IT investment decisions. Firms need to establish effective frames
of reference for aspirational performance to which they can compare their perfor­
mance and decide on future IT investment. Managers should consider IT as a key
resource to invest in, in order to address the firm’s performance shortfalls. In doing so,
they need to place salience on the role of IT to drive innovation and create value in
the long term. Such a way of goal setting can guide IT investment decisions based on
short-run feedback in a dynamic way that expand innovation outputs and improve
firm performance.
To safeguard IT investment decisions, firms should establish corporate governance
practices for board of directors to monitor managers’ key decisions for large IT invest­
ments. Corporate boards need to recognize that IT investment provides tremendous
opportunities to create value through innovation, but managers may have a tendency of
overinvestment in IT. This study shows that firms should mitigate managers’ empire
building via IT investment, and corporate governance can be an effective mechanism to
mitigate this tendency. Our results show that firms with IT investment above the industry
average are aggressive in making IT investment when there are performance shortfalls.
Effective board of directors can monitor IT investment decisions and mitigate overinvest­
ment in IT. Overall, firms need to pursue frames of reference for performance shortfalls
that are addressed by making IT investment to drive innovation and create value in the
long term while regulating IT overinvestment tendency through effective corporate
governance.

Limitations and Future Research

First, our analysis uses fixed effects model with time lag and a robustness check based on
Heckman model finds that endogeneity is not a serious concern. However, future studies
may explore possible instruments to better examine the causal effects in our theory.
Exogenous shocks such as the financial crisis in 2008 or the COVID-19 pandemic in
2020 may provide natural experiment settings to evaluate the causal influence of perfor­
mance shortfalls on IT investment.
Second, our sample consists of public U.S. firms only. Although public firms with
separation of administration and ownership are most likely to have agency problems and
50 DONG ET AL.

suitable for our study, we are not able to incorporate private firms, which could also
encounter agency problems. To further improve the generalizability of our findings, future
studies can also evaluate our findings based on firms in other emerging markets around
the world [e.g., 69].
Last but not least, our study provides evidence for IT investment at the organizational
level. Future research can examine the investment in specific technologies for solving
particular problems, especially emerging technologies such as artificial intelligence, big
data analytics, cloud computing, and social media, to address performance shortfalls in
different contexts [e.g., 3]. Such a fine-grained analysis may unveil new insights into
technological choices and performance impact of chosen technologies, which is likely to
be a fruitful avenue of research.

Conclusion
By integrating BTOF and agency theory, we advance a behavioral agency perspective to
theorize performance shortfalls, IT investment tendency and corporate governance as
key drivers that jointly determine IT investment in response to performance gaps. In
particular, we provide a nuanced explanation of IT investment decisions in response to
performance shortfalls across different organizational contexts, in which corporate
governance plays a dual role of controlling IT investment when there is a tendency
of overinvestment in IT and expanding IT investment when there is a tendency of
underinvestment in IT. Furthermore, innovation outputs serve as a mediating mechan­
ism linking resultant IT investment to firm performance. Overall, this study contri­
butes a behavioral agency theory to deepen our understanding about performance
drivers, governance, and outcomes of IT investment decisions.

Notes
1. In the empirical analysis, we also control for the deviation of a firm’s IT investment from the
industry average IT investment [57, 82].
2. One might view IT investment as costs and suggest a way to improve poor performance by
reducing these costs. However, BTOF contends that managers who seek to improve perfor­
mance are more likely to augment resource inputs to problemistic search and thus proactively
address performance shortfalls [43].
3. In a robustness check, we used Tobin’s Q as an alternative performance measure for forward-
looking IT investment decisions—some investments in new technologies may be made for
improving firm performance in the long run, which may not be fully driven by short-term
performance measure such as ROA. Tobin’s Q is a commonly used long-term performance
measure in the literature on business value of IT [e.g., 10, 14, 118]. Using Tobin’s Q as
performance measure generated consistent results [see Table 5, Model (1)]. We thank an
anonymous reviewer who suggested this point.
4. BTOF suggests that organizational learning is adaptive over time, indicating firm performance
in an earlier period is less important for decision making [18]. Given that firm performance is
highly correlated over time, using the average ROA in a three-year period from t – 3 to t – 1 and
using the ROA in the previous year t – 1 can empirically generate similar results [e.g., 88].
Having said that, we conducted a robustness check by using the average ROA from t – 3 to t – 1
as historical aspiration, which generated consistent results [see Table 5, Model (2)].
 JOURNAL OF MANAGEMENT INFORMATION SYSTEMS 51

5. We conducted a couple of robustness checks to test the sensitivity of results to this improve­
ment rate. We followed the BTOF literature to change the improvement rate from 5 percent
to 25 percent and 50 percent [17, 18], which yielded consistent results [see Table 5, Models
(3) and (4)] as what have been found in these prior studies [18, 88]. Thus, our results are
robust to the choice of improvement rate.
6. In contrast, some studies, primarily employing survey-based research designs, have focused
on advancing the conceptualization of problemistic search by developing novel constructs
which discern particular problem-solving activities in a specific context [e.g., 67]. While
survey methods are suitable to validate novel conceptualization of problemistic search, they
are likely to have limitations of self-reported data, cross-sectional design, and small sampling
scale.
7. R&D capital is calculated as CAPit = EXPit + (1 – δ)CAPit, where CAPit is the R&D capital of
firm i in year t, EXPit is the annual R&D expenditure of firm i in year t, and δ = 15 percent is
the depreciation rate of R&D capital [1, 19, 51]. Unlike IT labor expenditure that is totally
paid out every year, R&D expenditure depreciates over a four-year period [1, 19]. Initial R&D
capital is created as annual R&D expenditure multiplied by 4.3, because of perpetual
discounting of R&D capital assumed to be depreciated 15 percent per year and discounted
8 percent per year: 1/(0.15 + 0.08) = 4.3 [19]. We derived the difference of R&D capital in two
consecutive years as the measure for R&D capital to avoid double counting across years.
8. BTOF suggests that search can be induced by either problem-solving need in the form of
problemistic search, or slack resources in the form of slack search [28, 48]. Problemistic
search is goal-oriented and focused on the solutions to specific problems, while slack search is
more random and occasionally chaotic. Slack search is therefore less efficient in generating
innovative solutions than problemistic search, as it often tolerates loose discipline [86]. Slack
resources may or may not lead to resource investments for search, because slack resources
can also be used for other purposes such as alleviating intraorganizational collision among
subunits [93].
9. We addressed the endogeneity of IT investment by using a Heckman two-stage model in
a robustness check, and found that after controlling for the endogeneity our results still hold
[see Table 5, Models (6) and (7)].
10. Recent literature on mediation analysis suggests that it is not necessary to establish a direct
effect between the independent variable and the dependent variable in order to establish
mediation, and it only requires examination of whether the independent variable influences
the dependent variable through the mediator [121].

Acknowledgements
We thank the Editor-in-Chief, Vladimir Zwass, and three anonymous reviewers for their construc­
tive feedback. We also acknowledge the comments of seminar participants at the ESSEC Business
School and the University of Groningen. An early version of the paper received the Organizational
Communication and Information Systems (OCIS) Division Best Paper Award (First Runner-Up) at
Academy of Management Annual Meeting 2015 in Vancouver. The authors are listed alphabetically.

Funding
This research was partially funded by the Hong Kong Research Grants Council (HKU #17505018)
and the Tsinghua SEM Research Center for Artificial Intelligence & Management (AIM). The
Hong Kong University of Science and Technology, the University of Groningen and the
University of Hong Kong also provided partial financial support for conducting this research.
52 DONG ET AL.

ORCID
John Qi Dong http://orcid.org/0000-0002-3169-7790
Prasanna P. Karhade http://orcid.org/0000-0001-8078-2476
Arun Rai http://orcid.org/0000-0002-3655-7543
Sean Xin Xu http://orcid.org/0000-0002-9453-8404

References
1. Aghion, P.; Van Reenen, J.; and Zingales, L. Innovation and institutional ownership.
American Economic Review, 103, 1 (2013), 277–304.
2. Alavi, M.; and Leidner, D.E. Knowledge management and knowledge management systems:
Conceptual foundations and research issues. MIS Quarterly, 25, 1 (2001), 107–136.
3. Anand, A.; Sharma, R., and Kohli, R. The effects of operational and financial performance
failure on BI&A-enabled search behaviors: A theory of performance-driven search.
Information Systems Research, 31, 4 (2020), 1144–1163.
4. Anderson, M.C.; Banker, R.D.; and Ravindran, S. The new productivity paradox.
Communications of the ACM, 46, 3 (2003), 91–94.
5. Agarwal, R.; and Tiwana, A. Evolvable systems: Through the looking glass of IS. Information
Systems Research, 26, 3 (2015), 473–479.
6. Argote, L.; and Greve, H.R. A behavioral theory of the firm—40 years and counting:
Introduction and impact. Organization Science, 18, 3 (2007), 337–349.
7. Argyres, N.; Mahoney, J.T.; and Nickerson, J. Strategic responses to shocks: Comparative
adjustments costs, transaction costs, and opportunity costs. Strategic Management Journal, 40,
3 (2019), 357–376.
8. Balkin, D.B.; Markman, G.D.; and Gomez-Mejia, L.R. Is CEO pay in high-technology firms
related to innovation? Academy of Management Journal, 43, 6 (2000), 1118–1129.
9. Banker, R.D.; Bardhan, I.; and Asdemir, O. Understanding the impact of collaboration
software on product design and development. Information Systems Research, 17, 4 (2006),
352–373.
10. Bardhan, I.; Krishnan, V.; and Lin, S. Business value of information technology: Testing the
interaction effect of IT and R&D on Tobin’s Q. Information Systems Research, 24, 4 (2013),
1147–1161.
11. Baysinger, B.D.; Kosnik, R.D.; and Turk, T.A. Effects of board and ownership structure on
corporate R&D strategy. Academy of Management Journal, 34, 1 (1991), 205–214.
12. Benaroch, M.; and Chernobai, A. Operational IT failures, IT value destruction, and
board-level IT governance changes. MIS Quarterly, 41, 3 (2017), 729–762.
13. Berger, P.G.; and Hann, R. The impact of SFAS No. 131 on information and monitoring.
Journal of Accounting Research, 41, 2 (2003), 163–223.
14. Bharadwaj, A.S.; Bharadwaj, S.G.; and Konsynski, B.R. Information technology effects on firm
performance as measured by Tobin’s q. Management Science, 45, 6 (1999), 1008–1024.
15. Bharadwaj, S.; and Bharadwaj, A.; and Bendoly, E. The performance effects of complementa­
rities between information systems, marketing, manufacturing, and supply chain processes.
Information Systems Research, 18, 4 (2007), 437–453.
16. Boyd, B.K. CEO duality and firm performance: A contingency model. Strategic Management
Journal, 16, 4 (1995), 301–312.
17. Bromiley, P. Testing a causal model of corporate risk taking and performance. Academy of
Management Journal, 34, 1 (1991), 37–59.
18. Bromiley, P.; and Harris, J.D. A comparison of alternative measures of organizational
aspirations. Strategic Management Journal, 35, 3 (2014), 338–357.
19. Brynjolfsson, E.; and Hitt, L.M. Computing productivity: Firm-level evidence. Review of
Economics and Statistics, 85, 4 (2003), 793–808.
 JOURNAL OF MANAGEMENT INFORMATION SYSTEMS 53

20. Brynjolfsson, E.; Fitoussi, D.; and Hitt, L.M. The Information Technology Iceberg. Cambridge,
MA: Massachusetts Institute of Technology, 2006.
21. Bourgeois, III, L.J. On the measurement of organizational slack. Academy of Management
Review, 6, 1 (1981), 29–39.
22. Chen, W.-R. Determinants of firms’ backward- and forward-looking R&D search behavior.
Organization Science, 19, 4 (2008), 609–622.
23. Chen, W.-R.; and Miller, K.D. Situational and institutional determinants of firms’ R&D
search intensity. Strategic Management Journal, 28, 4 (2007), 369–381.
24. Cheng, Z.J.; Rai, A.; Tian, F.; and Xu; S.X. Social learning in information technology
investment: The role of board interlocks. Management Science, 67, 1 (2021), 547–576.
25. Chwelos, P.; Ramirez, R.; Kraemer, K.L.; and Melville, N.P. Does technological progress alter
the nature of information technology as a production input? New evidence and new results.
Information Systems Research, 21, 2 (2010), 392–408.
26. Cohen, W.M.; and Levinthal, D.A. Absorptive capacity: A new perspective on learning and
innovation. Administrative Science Quarterly, 35, 1 (1990), 128–152.
27. Corporate Board Member. Board and Information Technology Strategies Report: Maximizing
Performance through IT Strategy. Stamford, CT: Corporate Board Member, 2007.
28. Cyert, R.M.; and March, J.G. A Behavioral Theory of the Firm. Englewood Cliffs, NJ: Prentice
Hall, 1963.
29. Deephouse, D.L.; and Wiseman, R.M. Comparing alternative explanations for accounting
risk-return relations. Journal of Economic Behavior and Organization, 42, 4 (2000),
463–482.
30. Deloitte Insights. Technology and the Boardroom: A CIO’s Guide to Engaging the Board,
https://www2.deloitte.com/us/en/insights/focus/cio-insider-business-insights/boards-
technology-fluency-cio-guide.html, 2019.
31. Demirhan, D.; Jacob, V.S.; and Raghunathan, S. Information technology investment strategies
under declining technology cost. Journal of Management Information Systems, 22, 3 (2005),
321–350.
32. Denis, J.D.; and Sibilkov, V. Financial constraints, investment, and the value of cash holding.
Review of Financial Studies, 23, 1 (2010), 247–269.
33. Dewan, S.; Michael, S.C.; and Min, C. Firm characteristics and investments in information
technology: Scale and scope effects. Information Systems Research, 9, 3 (1998), 219–232.
34. Dewan, S.; and Ren, F. Information technology and firm boundaries: Impact on firm risk and
return performance. Information Systems Research, 22, 2 (2011), 369–388.
35. Doms, M. IT investment: Will the glory days ever return? FRBSF Economic Letter, 4, 1 (2005),
25–28.
36. Dong, J.Q. Technological choices under uncertainty: Does organizational aspiration matter?
Strategic Management Journal, In Press (2021).
37. Dong, J.Q.; and Yang, C.-H. Information technology and innovation outcomes: Is knowledge
recombination the missing link? European Journal of Information Systems, 28, 6 (2019),
612–626.
38. Dong, J.Q.; Karhade, P.P.; Rai, A.; and Xu, S.X. Information Technology and Innovation
Outputs: The Missing Link of Search Evolution. Vancouver: Academy of Management, 2015.
39. Eisenhardt, K.M. Agency theory: An assessment and review. Academy of Management Review,
14, 1 (1989), 57–74.
40. Fama, E.F. Agency problems and the theory of the firm. Journal of Political Economy, 88, 2
(1980), 288–307.
41. Finkelstein, S.; and D’Aveni, R.A. CEO duality as a double-edged sword: How boards of
directors balance entrenchment avoidance and unity of command. Academy of Management
Journal, 37, 5 (1994), 1037–1108.
42. Forman, C.; and van Zeebroeck, N. From wires to partners: How the Internet has fostered
R&D collaborations with firms. Management Science, 58, 8 (2012), 1549–1568.
43. Gavetti, G.; Greve, H.R.; Levinthal, D.A.; and Ocasio, W. The behavioral theory of the firm:
Assessment and prospects. Academy of Management Annals, 6, 1 (2012), 1–40.
54 DONG ET AL.

44. Gillan, S.L.; Hartzell, J.C.; and Starks, L.T. Tradeoffs in corporate governance: Evidence from
board structures and charter provisions. Quarterly Journal of Finance, 1, 4 (2011), 667–705.
45. Gómez, J.; Salazar, I.; and Vargas, P. Does information technology improve open innovation
performance? Information Systems Research, 28, 3 (2017), 661–675.
46. Gomez-Mejia, L.R.; Neacsu, I.; and Martin, G. CEO risk-taking and socioemotional wealth:
The behavioral agency model, family control and CEO option wealth. Journal of
Management, 45, 4 (2019), 1713–1738.
47. Guo, B.; and Ding, P. A behavioral theory of patent application rhythm: The moderating roles
of CEO overconfidence and discretion. Management Decision, 58, 4 (2019), 743–758.
48. Greve, H.R. A behavioral theory of R&D expenditures and innovations: Evidence form
shipbuilding. Academy of Management Journal, 46, 6 (2003), 685–702.
49. Griliches, Z. Patent Statistics as Economic Indicators: A Survey. Cambridge, MA: National
Bureau of Economic Research, 1990.
50. Hall, J.A. Financial performance, CEO compensation, and large-scale information technology
outsourcing decisions. Journal of Management Information Systems, 22, 1 (2005), 193–221.
51. Hall, B.H.; Jaffe, A.B.; and Trajtenberg, M. Market value and patent citations. RAND Journal
of Economics, 36, 1 (2005), 16–38.
52. Hall, B.H.; Jaffe, A.B.; and Trajtenberg, M. The NBER Patent Citations Data File: Lessons,
Insights and Methodological Tools. Cambridge, MA: National Bureau of Economic Research,
2001.
53. Han, K.; Chang, Y.B.; and Hahn, J. Information technology spillover and productivity: The
role of information technology intensity and competition. Journal of Management
Information Systems, 28, 1 (2011), 115–146.
54. Harris, J.; and Bromiley, P. Incentives to cheat: The influence of executive compensation and
firm performance on financial misrepresentation. Organization Science, 18, 3 (2007),
350–367.
55. He, J.; and Wang, H. Innovative knowledge assets and economic performance: The asym­
metric roles of incentives and monitoring. Academy of Management Journal, 52, 5 (2009),
919–938.
56. Hitt, L. M. Information technology and firm boundaries: Evidence from panel data.
Information Systems Research, 10, 2 (1999), 134–149.
57. Ho, J.L.Y.; Tian, F.; Wu, A.; and Xu, S.X. Seeking value through deviation? Economic impacts
of IT overinvestment and underinvestment. Information Systems Research, 28, 4 (2017),
850–862.
58. Ho, J.L.Y.; Wu, A.; and Xu, S.X. Corporate governance and returns on information technol­
ogy investment: Evidence from an emerging market. Strategic Management Journal, 32, 6
(2011), 595–623.
59. Holst, A. IT Budgets and Investments. Hamburg: Statista, 2020.
60. Iacobucci, D. Mediation Analysis. Thousand Oaks, CA: Sage Publications, 2008.
61. Iyer, D.N.; and Miller, K.D. Performance feedback; slack, and the timing of acquisitions.
Academy of Management Journal, 51, 4 (2008), 808–822.
62. Jensen, M.C. The modern industrial revolution, exit and the failure of internal control
systems. Journal of Finance, 48, 3 (1993), 831–880.
63. Jensen, M.C.; and Meckling, W.H. Theory of the firm: Managerial behavior, agency costs and
ownership structure. Journal of Financial Economics, 3, 4 (1976), 305–360.
64. Joshi, K.D.; Chi, L.; Datta, A.; and Han, S. Changing the competitive landscape: Continuous
innovation through IT-enabled knowledge capabilities. Information Systems Research, 21, 3
(2010), 472–495.
65. Kahneman, D.; and Tversky, A. Prospect theory: An analysis of decision under risk.
Econometrica, 47, 2 (1979), 263–291.
66. Karhade, P.P.; and Dong, J.Q. Information technology investment and commercialized
innovation performance: Dynamic adjustment costs and curvilinear impacts. MIS
Quarterly, In Press (2021).
 JOURNAL OF MANAGEMENT INFORMATION SYSTEMS 55

67. Karhade, P.P.; and Dong, J.Q. Innovation outcomes of digitally enabled collaborative pro­
blemistic search capability. MIS Quarterly, In Press (2021).
68. Karhade, P.P.; Shaw, M.J.; and Subramanyam, R. Patterns in information systems portfolio
prioritization: Evidence from decision tree induction. MIS Quarterly, 39, 2 (2015), 413–433.
69. Karhade, P.P.; and Kathuria A. Missing impact of online ratings on platform participation in
India: A call for research in GREAT Contexts. Communications of the Association for
Information Systems, 47, 1 (2020), 364–381.
70. Kathuria, A.; Karhade, P.P.; and Konsynski, B. In the realm of hungry ghosts: Multi-level
theory for supplier participation on digital platforms. Journal of Management Information
Systems, 37, 2 (2020), 396–430.
71. Karuna, C. CEO Reputation and Internal Corporate Governance. Irvine, CA: University of
California at Irvine, 2009.
72. Katila, R.; and Ahuja, G. Something old, something new: A longitudinal study of search
behavioral and new product introduction. Academy of Management Journal, 45, 6 (2002),
1183–1194.s
73. Kleis, L.; Chwelos, P.; Ramirez, R.V.; and Cockburn, I. Information technology and intangible
output: The impact of IT investment on innovation productivity. Information Systems
Research, 23, 1 (2012), 42–59.
74. Kobelsky, K.W.; Richardson, J.V.; Smith, R.E.; and Zmud, R.W. Determinants and conse­
quences of firm information technology budgets. Accounting Review, 83, 4 (2008), 957–995.
75. Kohli, R.; and Grover, V. Business value of IT: An essay on expanding research directions to
keep up with the times. Journal of the Association for Information Systems, 9, 1 (2008), 23–39.
76. Levina, N.; and Xin, M. Comparing IT workers’ compensation across country contexts:
Demographic, human capital, and institutional factors. Information Systems Research, 18, 2
(2007), 193–210.
77. Li, X. Managerial entrenchment with strategic information technology: A dynamic
perspective. Journal of Management Information Systems, 25, 4 (2009), 183–204.
78. Malhotra, A.; Gosain, S.; and El Sawy, O.A. Absorptive capacity configurations in supply
chains: Gearing for partner-enabled market knowledge creation. MIS Quarterly, 29, 1 (2005),
145–187.
79. Malhotra, A.; Gosain, S.; and El Sawy, O.A. Leveraging standard electronic business interfaces
to enable adaptive supply chain partnerships. Information Systems Research, 18, 3 (2007),
260–279.
80. Miller, J.S.; Wiseman, R.M.; and Gomez-Majia, L.R. The fit between CEO compensation
design and firm risk. Academy of Management Journal, 45, 4 (2002), 745–756.
81. Mithas, S.; Tafti, A.R.; Bardhan, I.R.; and Goh, J.M. Information technology and firm profit­
ability: Mechanisms and empirical evidence. MIS Quarterly, 36, 1 (2012), 205–224.
82. Mithas, S.; Tafti, A.R.; and Mitchell, W. How a firm’s competitive environment and digital
strategic posture influence digital business strategy. MIS Quarterly, 37, 2 (2013), 511–536.
83. Mitra, S. Information technology as an enabler of growth in firms. An empirical assessment.
Journal of Management Information Systems, 22, 2 (2005), 279–300.
84. Morgado, A.; and Pindado, J. The underinvestment and overinvestment hypotheses: An
analysis using panel data. European Financial Management, 9, 2 (2003), 163–177.
85. Nelson, R.R.; and Winter, S.G. An Evolutionary Theory of Economic Change. Cambridge, MA:
Harvard University Press, 1982.
86. Nohria, N.; and Gulati, R. Is slack good or bad for innovation? Academy of Management
Journal, 39, 5 (1996), 1245–1264.
87. Nolan, R.; and McFarlan, F.W. Information technology and the board of directors. Harvard
Business Review, 83, 10 (2005), 96–106.
88. O’Brien, J.P.; and David, P. Reciprocity and R&D search: Applying the behavioral theory of
the firm to a communitarian context. Strategic Management Journal, 35, 4 (2014), 550–565.
89. Officer, M.S. Overinvestment, corporate governance, and dividend initiations. Journal of
Corporate Finance, 17, 3 (2011), 710–724.
56 DONG ET AL.

90. Paravisini, D. Local bank financial constraints and firm access to external finance. Journal of
Finance, 63, 5 (2008), 2161–2193.
91. Pavlou, P.A.; and El Sawy, O.A. From IT leveraging competence to competitive advantage in
turbulent environments: The case of new product development. Information Systems
Research, 17, 3 (2006), 198–227.
92. Pepper, A.; and Gore, J. Behavioral agency theory: New foundations for theorizing about
executive compensation. Journal of Management, 41, 4 (2015), 1045–1068.
93. Pitelis, C.N. A behavioral resource-based view of the firm: The synergy of Cyert and March
(1963) and Penrose (1959). Organization Science, 18, 3 (2007), 478–490.
94. Platt Retail Institute. Retailers’ Investment in Technology: An Industry Perspective. Hinsdale,
IL: Platt Retail Institute, 2012.
95. Posen, H.E.; Keil, T.; Kim, S.; and Meissner, F.D. Renewing research on problemistic search—
A review and research agenda. Academy of Management Annals, 12, 1 (2018), 208–251.
96. Preacher, K.J.; and Hayes, A.F. Asymptotic and resampling strategies for assessing and
comparing indirect effects in multiple mediator models. Behavior Research Methods, 40, 3
(2008), 879–891.
97. PricewaterhouseCoopers. Boards Confront an Evolving Landscape: PwC’s Annual Corporate
Directors Survey. London: PricewaterhouseCoopers Center for Board Governance, 2013.
98. Ravichandran, T.; Han, S.; and Mithas, S. Mitigating diminishing returns to R&D: The role of
information technology in innovation. Information Systems Research, 28, 4 (2017), 812–827.
99. Ryan, S.D.; and Harrison, D.A. Considering social subsystem costs and benefits in informa­
tion technology investment decisions: A view from the field on anticipated payoffs. Journal of
Management Information Systems, 16, 4 (2000), 11–40.
100. Saldanha, T.J.V.; Mithas, S.; and Krishnan, M.S. Leveraging customer involvement for fueling
innovation: The role of relational and analytical information processing capabilities. MIS
Quarterly, 41, 1 (2017), 367–396.
101. Salge, T.O.; Kohli, R.; and Barrett, M. Investing in information systems: On the behavioral
and institutional search mechanisms underpinning hospitals’ IS investment decisions. MIS
Quarterly, 39, 1 (2015), 61–89.
102. Saraf, N.; Langdon, C.S.; and Gosain, S. IS application capabilities and relational value in
interfirm partnerships. Information Systems Research, 18, 3 (2007), 320–339.
103. Saunders, A. Has Information Technology Leveled the Competitive Playing Field? Philadelphia,
PA: University of Pennsylvania, 2010.
104. Shleifer, A.; and Vishny, R.W. A survey of corporate governance. Journal of Finance, 52, 2
(1997), 737–783.
105. Simon, H. A behavioral model of rational choice. Quarterly Journal of Economics, 69, 1
(1955), 99–118.
106. Sobel, M.E. Asymptotic intervals for indirect effects in structural equation models. In
S. Leinhart (ed.), Sociological Methodology. San Francisco, CA: Jossey-Bass, 1982, pp. 290-312.
107. Stulz, R.M. Managerial discretion and optimal financing policies. Journal of Financial
Economics, 26, 1 (1990), 3–27.
108. Susarla, A.; Subramanyam, R.; and Karhade, P.P. Contractual provisions to mitigate holdup:
Evidence from information technology outsourcing. Information Systems Research, 21, 1
(2010), 37–55.
109. Tambe, P.; and Hitt, L.M. The productivity of information technology investments: New
evidence from IT labor data. Information Systems Research, 23, 3 (2012), 599–617.
110. Tambe, P.; Hitt, L.M.; and Brynjolfsson, E. The extroverted firm: How external information
practices affect innovation and productivity. Management Science, 58, 5 (2012), 843–859.
111. Technology CEO Council. E-Health Readiness Guide. Washington, DC: Technology CEO
Council, 2014.
112. Tian, F.; and Xu, S.X. How do enterprise resource planning systems affect firm risk?
Post-implementation impact. MIS Quarterly, 39, 1 (2015), 39–60.
 JOURNAL OF MANAGEMENT INFORMATION SYSTEMS 57

113. Trantopoulos, K.; von Krogh, G.; Wallin, M.W.; and Woerter, M. External knowledge and
information technology: Implications for process innovation performance. MIS Quarterly, 41,
1 (2017), 287–300.
114. Tosi, H.L.; Katz, J.P.; and Gomez-Mejia, L.R. Disaggregating the agency contract: The effects
of monitoring, incentive alignment, and term in office on agent decision-making. Academy of
Management Journal, 40, 3 (1997), 584–602.
115. Walker, R. Patents as Scientific and Technical Literature. Metuchen, NJ: Scarecrow Press,
1995.
116. Wang, H.; Choi, J.; Wan, G.; and Dong, J.Q. Slack resources and the rent-generating potential
of firm-specific knowledge. Journal of Management, 42, 2 (2016), 500–523.
117. Wiseman, R.M.; and Bromiley, P. Toward a model of risk in declining organizations: An empirical
examination of risk, performance and decline. Organization Science, 7, 5 (1996), 524–543.
118. Xue, L.; Ray, G.; and Sambamurthy, V. Efficiency or innovation: How do industry environ­
ments moderate the effects of firms’ IT asset portfolios? MIS Quarterly, 36, 2 (2012), 509–528.
119. Yermack, D. Higher market valuation of companies with a small board of directors. Journal of
Financial Economics, 40, 2 (1996), 185–211.
120. Zellner, A. An efficient method of estimating seemingly unrelated regression equations and tests
for aggregation bias. Journal of the American Statistical Association, 57, 298 (1962), 348–368.
121. Zhao, X.; Lynch Jr., J.G.; and Chen, Q. Reconsidering Baron and Kenny: Myths and truths
about mediation analysis. Journal of Consumer Research, 37, 2 (2010), 197-206.
122. Zhao, Y.; and Chen, K.H. Staggered boards and earnings management. Accounting Review,
83, 5 (2008), 1347–1381.

About the Authors

John Qi Dong ([email protected]) is Chair and Professor of Business Analytics at the Trinity
Business School at Trinity College Dublin, University of Dublin. His research interests include
business analytics, digital innovation and a variety of topics related to behavioral strategy. His work
has been published or is forthcoming in MIS Quarterly, Strategic Management Journal, Journal of
Management, Journal of Management Information Systems, Journal of the Association for Information
Systems, Journal of Product Innovation Management, European Journal of Information Systems,
Information Systems Journal, and Journal of Strategic Information Systems, among others. Dr. Dong
serves as associate editor for MIS Quarterly and Journal of the Association for Information Systems.
Prasanna P. Karhade ([email protected]) is the Shidler College Faculty Fellow and a Faculty
Fellow at the Pacific Asian Center for Entrepreneurship (PACE) at the Shidler College of Business,
University of Hawaiʻi at Mānoa. He holds a Ph.D. in Business Administration from the University
of Illinois at Urbana-Champaign. He previously worked as a software engineer. Dr. Karhade’s
research interests, which include IT governance, and the impact of IT on firm innovation, are at the
intersection of Management Information Systems and Strategic Management. His work has been
published in Information Systems Research, Journal of Management Information Systems and MIS
Quarterly.
Arun Rai ([email protected]; corresponding author) is Regents’ Professor at the Robinson College of
Business at Georgia State University, holds the Robinson Chair, and is the director of the Center for
Digital Innovation. His research has focused on the development and deployment of information
systems to drive innovation and create value. His work has contributed to understanding the digital
transformation of organizations and supply chains, governance of IT investments and platform
ecosystems, and deployment of digital innovations at scale to empower individuals and address
societal problems including poverty, health disparities, infant mortality, and digital inequality. Dr.
Rai served as editor-in-chief for the MIS Quarterly. He is a Fellow of the Association for
Information Systems (AIS) and a Distinguished Fellow of the INFORMS Information Systems
Society, and received the AIS LEO Award for Lifetime Exceptional Achievement in the Information
Systems discipline.
58 DONG ET AL.

Sean Xin Xu ([email protected]) is a professor and an associate dean at the School of

Economics and Management, Tsinghua University, China. His work has been published in MIS
Quarterly, Information Systems Research, Journal of Management Information Systems, Management
Science, Strategic Management Journal, and Contemporary Accounting Research, among others. Dr.
Xu is a senior editor for MIS Quarterly and was an associate editor for Information Systems
Research. He received the MIS Quarterly Best Paper Award and was named the Best Associate
Editor for Information Systems Research.

APPENDIX. Measuring aspiration level

Measuring aspiration level has long been challenging in the behavioral theory of the firm (BTOF)
literature, because aspiration level is unobservable to researchers [36]. Cyert and March suggest that
decision makers of a firm formulate a single aspiration at the organizational level based on historical
and social information about firm performance [28]. Specifically, aspiration level is developed based
on historical aspiration (i.e., a firm’s historical performance) and social aspiration (i.e., the average
performance of other firms in the industry), but how historical aspiration and social aspiration
constitute the ultimate aspiration level is not very clear.
In prior literature, three empirical approaches have been used to solve this problem. The first
approach simply does not aggregate historical and social aspirations and uses them separately [e.g.,
23, 54, 61]. A potential issue of this approach is that employing two measures for a single construct
is theoretically inconsistent with BTOF and these two measures may not always generate consistent
results.
The second approach relies on a weighted average method to combine historical and social
aspirations by assigning weights to the historical and social components that give the highest model
fit to empirical data [e.g., 22, 48, 88]. This approach can derive a single aspiration level but has other
issues. First, it is essentially data-driven and the weights are theoretically unjustifiable. It takes the
measurement model that generates the best model fit but ignores other possible ways of aggregating
historical and social aspirations. Second, the estimates of weights for aspiration level depend on
other constructs in the research model. When the research model changes, the weights are likely to
change as well. Third, this approach imposes a zero-sum structure for the historical and social
components. Such a structure may even be implausible, because it may require that a firm with
a performance above the average performance in the industry aspires to lower its performance.
The third approach derives a theory-based measure for a single aspiration level by assuming that
the limited attention of boundedly rational decision makers of a firm systematically switches from
one aspiration to the other [e.g., 17, 29, 117]. Aspiration level is equal to social aspiration if the focal
firm’s performance is below social aspiration and a slightly higher historical performance otherwise.
The rationale is that firms need to catch up to industry peers if they are lagging behind, but will
continuously improve their own performance if they are in the lead. This approach has several
advantages over the other two approaches, especially its parsimony in capturing a priori theoretical
logic for the switching of limited attention about what decision makers attend to and why based on
the assumption of bounded rationality in BTOF [18].
Bromiley and Harris empirically compare these three approaches and find strong evidence in
favor of the switching approach and the separate measures approach over the weighted average
approach [18]. By using Akaike information criterion (AIC) and Bayesian information criterion
(BIC) to assess these three approaches in estimating 18 different models, they find that the switching
approach and the separate measures approach are better than the weighted average approach in 17
out of 18 estimates. Given that separate measures approach approach is inconsistent with BTOF
that suggests a single aspiration level [18], we chose the equally valid switching approach to
calculate a single measure of aspiration level.