Guggenmos 2020

The Effects of Creative Culture on Real Earnings Management*
RYAN D. GUGGENMOS, Cornell University†
ABSTRACT
Creativity and innovation have been identified by senior executives as some of the most desired
characteristics of corporate culture. Accordingly, managers strive to build these cultures within
their organizations. However, research in psychology suggests that these attempts may have
unintended negative consequences. In this study, I predict and find that managers in a more (versus
less) innovative company culture will engage in higher levels of real earnings management
(REM). I then test two construal level theory (CLT)-based interventions designed to reduce REM.
As I predict, I find that in more innovative corporate cultures an intervention that makes downside
risk more salient reduces REM, but an intervention that encourages managers to consider the “big-
picture” impact of their decision reduces REM to a greater extent. Unexpectedly, I also find that
the effect of the “big-picture” intervention reverses in a less innovative corporate culture leading to
an increase in REM. My findings contribute to the emerging accounting literature regarding REM.
I also extend the psychology literature investigating the link between opportunistic behavior and
creativity, and I also expand research into how interventions based on CLT can affect judgment
and decision making in an accounting context.
Keywords: creativity, real earnings management, financial reporting, construal level theory
Les répercussions de la culture de la créativité sur la gestion du

résultat réel
RÉSUMÉ
Pour certains hauts dirigeants, la créativité et l’innovation comptent parmi les caractéristiques les
plus recherchées de la culture d’entreprise. C’est pourquoi les gestionnaires s’efforcent d’introduire
ces cultures au sein de leurs organisations. Les recherches en psychologie paraissent toutefois
indiquer que ces efforts peuvent avoir des répercussions négatives imprévues. L’auteur pose
et démontre l’hypothèse selon laquelle les gestionnaires d’entreprises dont la culture est, par
comparaison, davantage axée sur l’innovation affichent des niveaux supérieurs de gestion du
résultat réel (GRR). Il vérifie ensuite deux interventions basées sur la théorie des niveaux
* Accepted by Jeffrey Hales. An earlier version of this paper was presented at the 2018 Contemporary Accounting
Research Conference, generously supported by the Chartered Professional Accountants of Canada. I am extremely grate-
ful for comments and guidance from my dissertation committee: Chris Agoglia (chair), Bradley Bennett, Linda Isbell,
Dave Piercey, and Elaine Wang. I am also appreciative of helpful comments from Scott Asay, Eddy Cardinaels, Shana
Clor-Proell, Mary Kate Dodgson, Mike Durney, Scott Emett, Jeffrey Hales (editor), Kathryn Kadous, Steve Kuselias,
Bob Libby, Nikki MacKenzie, Jessica Osgood, Jeff Pickerd, Bradley Pomeroy, Kristi Rennekamp, Aaron Saiewitz, Chez
Sealy, Wim Van der Stede (discussant), two anonymous reviewers, workshop participants at the 2015 ABO Midyear
Meeting and Research Conference, the 2018 Contemporary Accounting Research Conference, Cornell University, the
University of Kentucky, the University of Massachusetts Amherst, and Virginia Commonwealth University. Lastly, I
thank the University of Massachusetts Isenberg Scholar Program, Graduate School Dissertation Research Grant Program,
and accounting department for financial support. This manuscript is based on my dissertation at the University of Massa-
chusetts Amherst. Approval for this study was granted by the University of Massachusetts Amherst Institutional Review
Board, the institution where the study was conducted.
† Corresponding author.
Contemporary Accounting Research Vol. 00 No. 00 (Month 2020) pp. 1–38 © CAAA
doi:10.1111/1911-3846.12586
2 Contemporary Accounting Research
conceptuels, destinées à réduire la GRR. Conformément à ses prévisions, il constate que dans les
cultures d’entreprise davantage axées sur l’innovation, une intervention qui met en évidence le
risque de perte réduit la GRR, mais une intervention qui encourage les gestionnaires à appréhender
dans leur ensemble les conséquences de leurs décisions réduit la GRR dans de plus grandes pro-
portions. Fait inattendu, l’auteur constate également que l’incidence de l’intervention favorisant la
« vue d’ensemble » s’inverse dans une entreprise dont la culture est moins axée sur l’innovation,
ce qui induit une augmentation de la GRR. Ces observations contribuent à la littérature comptable
en émergence portant sur la GRR. L’étude vient également enrichir les travaux de psychologie
consacrés au lien entre le comportement opportuniste et la créativité, de même qu’elle invite les
chercheurs à se pencher sur la façon dont les interventions basées sur la théorie des niveaux
conceptuels peuvent influer sur le jugement et la prise de décisions dans le contexte de la
comptabilité.
Mots clés: créativité, gestion du résultat réel, information financiére, théorie des niveaux conceptuels
1. Introduction
According to a survey of over 1,500 CEOs, creativity is the most important leadership competency
for navigating today’s complex global business environment (IBM 2010). As a consequence, count-
less books, articles, and blog posts have been written to help managers build innovative cultures
(Chima 2013). Creativity and innovation can lead to exciting new products, greater agility in fast-
paced business environments, and the development of novel business processes. However, research
in psychology suggests that creativity and innovation may also be associated with less positive out-
comes, such as misreporting and greater acceptability of risk (Gino and Ariely 2012).
The purpose of this study is to explore how an innovation-focused corporate culture affects
employee willingness to engage in a specific risky behavior—real earnings management (REM).
Prior work in psychology has focused on how creativity can lead to dishonesty via misreporting
(Gino and Ariely 2012). However, it is an open question as to how creativity affects activities that
are reported accurately, but are simultaneously self-interested and myopic, and may increase risk to
the firm. Building off research in psychology and organizational theory, I predict that an innovative
corporate culture will lead to greater REM by increasing the acceptability of risky decision making.
Once this relationship is demonstrated, I test two consequence-focused interventions, based on con-
strual level theory (CLT), that are designed to reduce REM within the company’s existing culture.
One intervention, relying on lower-level construal, is intended to reduce REM by increasing down-
side risk salience. The other, based on higher-level construal, is expected to reduce REM by encour-
aging managers to consider the “big-picture” impact of their decisions on the organization as a
whole, decreasing preferences for instant gratification and increasing self-control.
My experimental results are largely consistent with my predictions. First, I find that managers in
a more innovative company culture engage in greater REM than managers in a less innovative com-
pany culture. Second, within the more innovative corporate culture, I find that the intervention that
highlights the downside risks of managers’ decisions diminishes REM, but the intervention that helps
managers see the “big-picture” impact of their decision reduces REM to a greater extent. However, I
find an unexpected effect of these interventions in a less innovative company culture. Contrary to
my predictions, I find that the same “big-picture” intervention that reduces REM to the greatest
extent in a more innovative company culture increases REM in a less innovative company culture.
Finally, I conduct additional analyses to investigate managers’ motivations and support
a post hoc explanation for the unexpected result. I demonstrate that, absent intervention,
employees in more innovative company cultures want to engage in more current year REM than
in less innovative company cultures. Additionally, I provide evidence consistent with a post
hoc explanation for why higher-level construal-based interventions affect managers’ motiva-
tions differently across cultures. The results of these analyses suggest that, in more innovative
CAR Vol. 00 No. 00 (Month 2020)

Effects of Creative Culture on REM 3
company cultures, higher-level construal interventions increase desire to engage in behavior

that is long-run beneficial to the firm even when it reduces the short-term benefit to the
decision-maker, as managers consider their decision’s “big-picture” impact on growth and
moving the firm forward. But in less innovative company cultures, focusing managers on the
“big-picture” impact of their decisions may not have the same effect, as managers may view the
“big-picture” to be preservation of the status quo.
The findings of this study offer several contributions. Although prior research has examined
how unscrupulous company leadership can foster environments where fraud and misreporting are
accepted (Mayer et al. 2012; Patelli and Pedrini 2013), accounting research has not investigated
unintended consequences of desirable corporate cultures. This is an important research area, as
noted by Dichev et al. (2013) and Graham et al. (2016, 2017), as they call for deeper analysis of
the effects of corporate culture on earnings quality and organizational outcomes. My findings also
add to the operational distortion literature examining the impact of using operational decisions to
achieve performance metrics (Bentley 2019; Bloomfield 2017).
Furthermore, as the psychology literature examining the downsides of creativity is still rela-
tively new, researchers have generally focused on laboratory settings in which creativity leads to
misreporting (Gino and Ariely 2012). However, investigating managers’ REM choices provides a
unique setting in which participants can engage in self-interested behavior and take on risk while
maintaining honest reporting. I contribute to the psychology literature by demonstrating that crea-
tivity can increase opportunistic behavior even within the bounds of accurate reporting. In addi-
tion, my unexpected finding that managers in more and less innovative corporate cultures respond
differently to higher-level construal-based interventions suggests that corporate cultures may
affect psychological processes in ways we do not yet fully understand. This provides ample
opportunity for future research.
This study’s results should be of interest to business professionals. First, company leadership
should be aware that cultivating innovative corporate culture may have downsides. Even though a
company’s culture may not appear harmful, elements of the culture that increase risky behavior
should be considered. Furthermore, as it is inadvisable to suggest that companies forgo innova-
tion, my results provide managers with interventions that can reduce REM behavior within more
innovative company cultures. In addition, my unexpected result provides initial evidence that
attempts to curb REM in less innovative corporate cultures by having managers focus on the
“big-picture” impacts of their behavior may not work as intended.
The remainder of the paper is organized as follows. The next section presents a review of rel-
evant literature and a development of my predictions. Section 3 describes the research design,
and section 4 reports data analysis and results. Section 5 concludes.
2. Literature review and hypothesis development

The pervasiveness of culture
Corporate cultures, systems of norms and values within organizations, have been a prominent research
area and topic in the business press for at least five decades. Generally, academics have taken a
descriptive approach to the topic, using both quantitative and qualitative research to identify cultural
indicators across companies and to map these indicators onto types of cultures (Amabile 1988; Miles
et al. 1978; Smircich 1983; Van den Steen 2010). Most recently in accounting, Graham
et al. (2016, 2017) offer survey and interview evidence that links corporate culture to firm value.
Prior research describes both the mechanism and strength of corporate culture’s effect on
day-to-day corporate life. Schein (2010) notes that culture determines who ascends through the
company’s ranks into leadership positions and simultaneously recognizes that those in leadership
roles create and manage culture. McShane and Von Glinow (2010) find that many companies for-
mally codify organizational culture into mission statements or value propositions but also explain
that codification is not the principal driver of company culture. Regardless, senior executives

interviewed by Graham et al. (2017) recognize corporate culture’s power to influence behavior
and describe culture as a “beliefs system,” “coordination mechanism,” “standard of behavior,”
and an “invisible hand.”
Creativity and innovation

Even with substantial cultural diversity across firms, there is a remarkable consensus with
respect to the value of innovation. In a synthesis of descriptive interview studies and experi-
mental laboratory findings, Amabile (1988, 126) distinguishes between creativity and organiza-
tional innovation, defining creativity as the “production of novel ideas and useful ideas by an
individual or small group,” and organizational innovation as “the successful implementation of
creative ideas within an organization.” In the same work, she proposes an organizational crea-
tivity model, which identifies crucial components of organizational innovation. These environ-
mental qualities, as originally identified in prior work by Amabile and Gryskiewicz (1987), can
either promote or inhibit creativity and, consequently, organizational innovation. For example,
a company culture that communicates the importance of innovation and expects some failure
along the way can serve as a powerful motivator of creative work. On the other hand, when
cues throughout the work environment suggest that management disregards innovation or sug-
gests that employees should not take risks, emphasize the status quo, or show reluctance to
change, culture may stifle innovation.1 This can even be the case when employees come to the
organization with high intrinsic motivation for creative work.
Even though executives report that innovative corporate culture is highly desirable (IBM 2010),
academic research suggests downsides as well. Prior research in psychology and organizational
behavior has shown that individuals who feel more at home in an innovative company culture also
tend to enjoy risk-taking, experimentation, and environments that are less rule-oriented or less cau-
tious (Amabile 1988; O’Reilly et al. 1991). Accordingly, individuals feeling more at home in innova-
tive organizations may have a preference for accepting greater organizational decision-making risk.
However, risk acceptance has different implications in different areas. Risk-taking may be welcomed
or desired in areas such as product development or research. However, stakeholders may prefer that
risk be minimized in other areas, such as consumer safety or financial reporting, where increased risk
could impact the organization negatively.
In addition to research suggesting that employees in more innovative corporate cultures may be
more accepting of risk, other research in psychology explores relationships between creativity and
misreporting. Gino and Ariely (2012) find that individuals primed to think creatively are more likely
to misreport performance on a laboratory task, even after controlling for individual differences in cre-
ative personalities. In addition, Gino and Wiltermuth (2014) find this association in reverse. That is,
misreporting leads to greater creativity. In both directions, the authors find that divergent thinking
processes (i.e., the feeling of being unconstrained by rules) serve to link creativity and misreporting.
This is consistent with Amabile and Gryskiewicz (1987) who find that perceptions of freedom over
one’s work can increase, and feelings of constraint can reduce, organizational innovation.
In a financial reporting context, if more innovative corporate cultures lead employees to
greater misreporting or risk-taking, this effect on employees in the company’s traditional financial
accounting and reporting function may be of first concern. However, while the finance and
accounting function is responsible for ensuring that the organization’s transactions are properly
recorded and reported, it is often employees outside of the financial reporting and accounting
function that make the decisions regarding when, and even whether, the real economic activities
that underlie the accounting records occur.
1. It is important to note that even though the study is motivated by the desirability of creative culture, this does not
imply that less innovative cultures are not viable or are less focused on achieving financial results. If this were the
case, then the importance of firm performance would differ between more and less innovative companies. Accord-
ingly, I hold the company’s results focus constant across conditions. This is discussed in more detail in section 3.

Prior research has shown that managers often choose to manage reported earnings through the
strategic timing of investing, operating, and financing activities, and may do so even when these deci-
sions are suboptimal to the long-term viability of the firm (Cohen et al. 2008; Roychowdhury 2006).2
This is known as REM. In the post–Sarbanes-Oxley era, managers are less likely to engage in
traditional accrual-based earnings management via the accounting and finance function and are
more likely to engage in REM (Cohen et al. 2008; Graham et al. 2005). While REM activities
vary, financial statement auditors have noted that managers often make operating decisions
that shift costs to subsequent periods in order to reduce current period expenses (Commerford
et al. 2016). For example, managers may halt advertising, R&D, or maintenance in the fourth
quarter to meet earnings targets (Graham et al. 2005; Roychowdhury 2006).
To the extent that REM increases firm risk, managers in companies that encourage risk-
taking may weight trade-offs between short-term incentives and long-term benefits differently
than managers in organizations that have cultures less tolerant of risk. As risk acceptability
increases, failure becomes more acceptable as well—but the incentives for myopic behavior stay
constant.3 In these cultures, managers may feel as if they have a “safety net” if they engage in
risky behavior and their decision leads to a negative outcome.
As discussed earlier, a company’s risk acceptance can be communicated implicitly through
corporate culture. Because more innovative corporate cultures have been associated with greater
risk acceptance (Graham et al. 2016, 2017; O’Reilly et al. 1991), I predict that managers will be
more likely to accept the risk that comes with engaging in REM in these cultures. This leads to
the following hypothesis:
HYPOTHESIS 1. Absent intervention, managers immersed in a more innovative company cul-

ture will engage in more real earnings management than managers immersed in a less
innovative company culture.
Construals and intervention

While I predict that more innovative culture will lead to greater REM, forgoing innovation is
impractical; the benefits are simply too great. An intervention that reduces REM within the firm’s
culture could be valuable to managers. To design this intervention, I turn to CLT.
CLT argues that individuals can only directly experience what is happening in the here and
now (Trope and Liberman 2010; Weisner 2015). Accordingly, events removed from this realm are
experienced abstractly, in that they are experienced indirectly from a perspective removed from
their own (Liberman and Trope 2008; Trope and Liberman 2010). Many managerial decisions are
2. Recent archival research has provided some evidence that engaging in REM is not universally negative. For exam-
ple, Vorst (2016) builds off work by Gunny (2010) to show that, on average, REM is associated with reductions in
future industry-adjusted ROA and cash flow from operations, but this relationship is less negative or disappears
when REM is used to hit a benchmark the firm would not otherwise achieve and the REM consists of a reversing
cut in investment.
3. Because REM is the intentional manipulation of firm activities to achieve an earnings benchmark, engaging in
REM often contains elements of risk, managerial myopia, and self-dealing. Firm risk can increase when engaging
in REM as uncertainty grows with respect to whether the operational activity will eventually occur, as the benefits
of the activity decrease over time, or as the costs of the activity increase over time (e.g., emergency repairs are often
more costly than planned maintenance) (Ewert and Wagenhofer 2005; Vorst 2016). Furthermore, managers have
incentives to engage in this behavior because they receive benefits from achieving benchmarks (e.g., bonuses tied
to accounting numbers, positive indications of performance for tenure decisions; Ewert and Wagenhofer 2005).
And finally, prioritization of a short-term earnings benchmark over long-term value is indicative of myopic behavior
(Bens et al. 2002). While all three of these elements often occur in tandem, risk is the most closely tied to the psy-
chological mechanism underlying the potential downside of more innovative company culture. With that said, it is
important to note that not all REM activities increase uncertainty-related risk. For example, engaging in REM by
providing sales discounts reduces the uncertainty of a sale occurring. I leave investigation of settings in which
REM does not result in increased uncertainty-related risk to future research.

experienced in this manner.4 For example, suppose a manager is considering outsourcing a main
component of a company’s flagship product. Many possible outcomes are important to consider,
but they cannot be experienced directly by the manager before the decision is made. Instead, the
manager thinks about potential actions, the situation, the events, how others will react, and how
likely the outcomes are to occur. This process is known as abstraction (Weisner 2015). Research
has consistently shown that the manner in which an individual engages in abstraction can have pre-
dictable decision-making effects (Liberman and Trope 2008; Trope and Liberman 2010).
Continuing the outsourcing example, managers might think about how outsourcing could impact
day-to-day plant operations. Alternatively, they may think about why outsourcing could impact the
company as a whole. When managers consider day-to-day impact, the picture in their mind is relatively
concrete and contextualized. For example, they may picture how to use new excess plant capacity. This
thought process is known as lower-level construal (Trope and Liberman 2003; Weisner 2015). In con-
trast, considering why outsourcing could impact the company as a whole is a more abstract and
decontextualized task. To consider this question, managers have to think more broadly about the entire
company. This is known as higher-level construal (Trope and Liberman 2003; Weisner 2015).
One way to reduce REM could be to use an intervention that prompts managers to consider their
decisions using lower-level construal.5 Research has shown that lower-level construal brings to mind
the concrete components of a choice and encourages pragmatism (Kivetz and Tyler 2007; Trope and
Liberman 2010). Especially applicable to concerns regarding REM, Chandran and Menon (2004)
show that lower-level construal can make the negative outcomes of risky behavior appear more con-
crete. In their study, the combination of lower-level construal and negative decision outcome framing
increased precaution, raised concerns and anxiety about hazards, and improved overall risk communi-
cation effectiveness. This suggests that a lower-level construal-based intervention may reduce REM.
Based on this work by Chandran and Menon (2004), managers thinking about a future deci-
sion using higher-level construal will weigh the immediate downside risks of engaging in REM
less than managers using lower-level construals. However, prior research suggests that higher-
level construal may be beneficial in reducing REM in other ways. Higher-level construal focuses
decision makers on the “big-picture” consequences of their decisions (Rogers and Bazerman 2008;
Trope and Liberman 2010). In addition, prior research has found that higher-level construal leads
to more self-control, decreased preferences for instant gratification, and less positive evaluations
of temptations (Fujita et al. 2006; Fujita and Carnevale 2012). These findings suggest that self-
interested behavior could be less appealing when managers use higher-level construal. To contrast
the two approaches, theory predicts that lower-level construal can reduce REM by increasing the
salience of downside risk, but higher-level construal can reduce REM in three different ways: by
decreasing preferences for instant gratification, by making “big-picture” consequences more
salient, and by increasing self-control. This provides multiple ways for the intervention to reduce
REM. This suggests that while lower-level and higher-level construal interventions will both
reduce REM, higher-level construal interventions will reduce REM to a greater extent. This
prediction, together with the prediction of greater baseline REM in more innovative company
cultures (Hypothesis 1), implies the following ordinal interaction hypothesis, stated formally:
HYPOTHESIS 2. The extent to which managers in a more innovative company culture engage
in greater real earnings management than managers in a less innovative company
4. While an extensive body of CLT-based research exists in psychology, this is not the case in accounting. In a litera-
ture review, Weisner (2015) argues that predictions and insights related to CLT are especially applicable to account-
ing contexts and could provide important insights with respect to the judgment and decision making of managers,
auditors, tax professionals, investors, and other stakeholders. Even so, Weisner (2015) only reports seven studies
informed by CLT-based hypotheses. Three of these studies relate to auditing, three to financial reporting, and one
to managerial accounting.
5. For ease of exposition, I refer to “interventions that prompt managers to think about their decision using lower
(higher)-level construal” as lower-(higher)-level construal interventions.

culture will be most pronounced absent any construal intervention, less pronounced
when a lower-level construal is emphasized, and least pronounced when a higher-level
construal is emphasized.
3. Method
Participants
I test my hypotheses using a 3×2 between-participants experiment. Study participants are 139 cur-
rent or past managers with an average of 5.8 years of management experience recruited from
Amazon Mechanical Turk (AMT).6 To ensure participants have adequate knowledge for the
experimental tasks, potential participants were prescreened to ensure that they have held or cur-
rently hold a management position and have experience making organizational budgeting deci-
sions.7 These characteristics are important to ensure that participants understand the implications
of missing an earnings target or other company goal and, therefore, are appropriate for the task
(Libby et al. 2002). Participants who did not meet the prescreening criteria were not included in
the study. On average, participants completed the study in 29.91 minutes and earned $4.20 ($3.00
in base pay and $1.20 in bonus) for their participation.8 AMT has become a popular source of
participants for studies in the social sciences, because data are inexpensive, easily obtainable, and
research has shown that online workers exert at least as much effort during experimental tasks as
traditional participants (Farrell et al. 2016; Krische 2019; Paolacci et al. 2010; Rennekamp 2012).
Experimental task
The study begins by providing participants with background information about “Bean’s Burgers,”
a publicly traded fast-casual restaurant chain. Participants are first exposed to the company culture
manipulations as they read through background information about the company. The primary
manipulation of corporate culture is achieved through the company’s slogan: “Think BEAN!”
The slogan is an acronym with the underlying meaning of the slogan manipulated between condi-
tions. The manipulation is discussed more fully below. Once participants finish reading back-
ground information and answer comprehension questions related to the company’s culture, they
proceed to the first task (Figure 1).
The first task is a product design initiative being held to celebrate the company’s 25th anni-
versary. This task requires employees to design a burger that they think, based on the company’s
culture, the company’s founder would most want to see on this special menu. The task’s purpose
6. The final participant count was 137, as one participant was excluded for having a duplicate IP address and one par-
ticipant was excluded for copying her experimental task submission from an Internet cooking website. All submis-
sions were run through an internet search engine as a rough assessment of whether participants generated their own
recipes—as this is necessary to infer that participants understood the company culture manipulation. Inferences are
unaffected by the inclusion or exclusion of these participants, as all p-values that were significant at the p < 0.05
level remain significant at that level.
7. Experimental participants were recruited for the study in two stages. First, a demographic survey was administered
to the general AMT population. This survey was comprised of eight questions; two of these questions contained the
questions used to determine study eligibility. This survey was administered to 3,076 individuals, of which
620 (20.4 percent) met the inclusion criteria. Second, a link to the experiment was sent to this pool of 620 individ-
uals, of which 139 responded. Because potential participants were blind to inclusion criteria, this procedure helps to
mitigate concerns that AMT workers would falsely indicate that they held the requisite experience in order to partic-
ipate in the study.
8. Paolacci et al. (2010) report that 41 percent of online workers complete tasks for entertainment purposes, and if this
the case, these participants may not be motivated by incentives—and especially on the order of magnitude that is
common for studies using online participant pools. Farrell et al. (2016) specifically test this concern and find that
online workers are sensitive to incentive contract design at pay levels that are low relative to traditional experiments
conducted in the laboratory. In addition, they do not find online workers to be sensitive to incentive contract design
at pay levels similar to traditional laboratory experiments. This suggests that compensation levels should be
matched to the norms of the platform, even if they may appear to be lower than one might expect.

Figure 1 Experimental design
Intervention Manipulation
No Intervention
Lower-level Construal
Higher-level Construal
Introduction Meeting with

Burger
and Outgoing Spending Experiment
Think BEAN! Creation PEQs
Background Regional Decision Ends
Task
Information Manager
Company Culture Manipulation
Less Innovative Culture

More Innovative Culture
Notes: This figure presents the study’s experimental design. Participants begin the experiment with the
Introduction and Background Information and proceed through until the experiment ends, after PEQs are
completed.
is to (i) reinforce the company culture manipulation and (ii) include a creative work task in both
conditions. This latter purpose is important, as without a creative work task in both conditions,
inferences may not be able to be isolated to the effect of more innovative corporate culture but
instead could be attributable to differential participation in a task that requires creativity (even if
that creativity is used to design a product that reflects a lack of creativity).9
The second task is a repair and maintenance spending decision. In this task, participants read
a transcript of a meeting with the outgoing regional manager to discuss the current state of the
region. In this vignette, participants are told that many of the region’s restaurants are part of the
company’s first wave of expansion. These restaurants, being older, have outdated kitchen equip-
ment budgeted for repairs and maintenance by year-end. Participants are told that the refrigeration
units in some of these restaurants are keeping food at the proper temperature intermittently and if
unsafe food is inadvertently served to customers and customers become ill, the impact of bad
publicity on the company could be severe. Finally, participants are told that even though the
repairs are budgeted, if the amount budgeted is spent, the company will miss its current year earn-
ings per share (EPS) target. Participants learn that even in a tough year for the industry, the comp-
any’s competitors have achieved their EPS targets. Study participants are then presented with the
intervention manipulation, reminded of the company’s “Think BEAN!” philosophy, and asked to
choose a current year kitchen update spending amount, which defers the remaining budgeted
spending into the next year. This design choice places the construal level intervention firmly
9. By holding the work task and compensation awarded for the task constant between levels of the company culture
conditions, I guard against perceptions that either culture is more or less focused on producing viable products. This
biases against finding increased risk acceptability in the more innovative culture conditions due to a more laissez-
faire attitude with respect to performance.

within the company’s culture.10 The primary dependent variable is the dollar amount of spending
on kitchen repairs deferred into the following year, a measure of REM.
To operationalize the incentive to meet or beat the company’s EPS target in both the current
and subsequent year, and the short- and long-term trade-offs that accompany this decision, partici-
pants are eligible for two bonuses. The first bonus is for meeting and/or beating the current year
EPS target. This bonus is displayed to participants as a point amount, changes as participants con-
sider different current year spending amounts, and is paid within 24 hours of study completion.
The second bonus is tied to achieving the EPS target in the subsequent year. This bonus is pres-
ented as a range of possible bonus amounts, with the actual bonus determined by a random draw
from a uniform distribution bounded by the range disclosed to participants, and is paid 10 business
days after the conclusion of the study. As participants increase expense deferral, the current year
bonus increases and the expected value of the subsequent year bonus decreases (i.e., expenses are
shifted from current year to the subsequent year, which increases EPS and bonus in the current
year, and reduces EPS and bonus in the subsequent year).11 Participants are not informed of the
actual amount earned until after the study’s conclusion.
Finally, to make the downside risk of engaging in REM more salient, participants are told that
if an outbreak of food-borne illness occurs they will not be paid a bonus for either year. Food-borne
illness risk increases as repairs and maintenance expense deferral increases. The level of risk, shown
to participants, ranges from 0.1 percent, if participants engage in zero REM, to 15 percent, if partic-
ipants defer the entire budgeted amount. A screenshot from the No Intervention condition is
included as Figure 2. The chance of food-borne illness is operationalized via a draw from a uniform
distribution, with parameters that match the amount of risk taken by study participants. Participants
are told after they complete post-experimental questions whether food-borne illness occurred. I
promptly pay the current year bonus as a known amount, delay an uncertain subsequent year bonus
payment, and include a penalty of zero bonus for a rare, but highly negative, outcome that increases
in likelihood as a consequence of expense deferral (REM). Taken together, this bonus structure
combines elements of risk, self-dealing, and myopia that coexist in REM settings.12
After the second task, participants answer post-experimental questions, are told the amount
and timing of their bonus payments, and are thanked for their participation.
Manipulations
I manipulate two independent variables, Company Culture and Intervention, between participants
using a fully-crossed 2×3 factorial design. Company Culture is manipulated at two levels: More Inno-
vative Culture and Less Innovative Culture. To encourage participants to internalize this
10. By presenting a concurrent manipulation of corporate culture and intervention, I am able to assess the effect of the
intervention within the overarching culture—the specific setting needed to test my research question. While partici-
pants were initially exposed to the company culture manipulation prior to the construal level intervention, by
reminding participants to “Think BEAN!” concurrent with the spending decision, I guard against concerns that the
interventions alter the company’s culture.
11. Repairs and maintenance expense available to defer ranged from $0 to $3.5 million, with a deferral of $1.035 mil-
lion or greater necessary to achieve the EPS target. Specifically, bonuses in Year 1 range from $0.00 (no expense
deferral) to $1.70 (maximum expense deferral). Bonus amounts in Year 2 are drawn from ranges of $0.95–$1.00
(no expense deferral) to $0.40–$0.45 (maximum expense deferral). The design choice to provide higher bonuses in
the current year was made to provide a more powerful test of theory, but as the bonus structure is held constant
across experimental conditions, does not affect the comparisons made between cells in the experimental design
(Libby et al. 2002).
12. The deferral of repairs and maintenance activity chosen to operationalize REM in this experiment makes risk espe-
cially salient when compared to other REM activities, such as providing end of year sales discounts or cutting
R&D. However, this design choice was made to strengthen the manipulation and would be expected to alter the
level of REM and, therefore, the amount of REM available for the interventions to reduce, but would not be
expected to have an effect on the overall pattern of results. As noted in footnote 3, I leave investigation of settings
in which REM and risk are not as closely intertwined to future research.

Figure 2 Spending decision screenshot
10

Contemporary Accounting Research
Notes: This figure presents the dependent variable screen from the No Intervention conditions as an example of the DV elicitation. Differences in wording between
levels of the Intervention conditions are given in Appendix 3.
manipulation, I use cultural artifacts and phrasing to signal the company’s pervasive
innovation-focused culture (Amabile 1988; McShane and Von Glinow 2010). The manipula-
tion captures the most important characteristics of innovative company culture, as described
by Amabile (1988), while holding the viability of the business strategy and importance of
good performance constant across cultures.
The primary manipulation of company culture is embedded in the company’s slogan, “Think
BEAN!” As participants read the company’s background information, they are told that “Think
BEAN!” is a prominent part of the culture at Bean’s Burgers and that this philosophy has been
the secret to the company’s success. The manipulation follows research in the management litera-
ture that shows persistent thinking patterns and shared meanings between employees are both
suggestive of a strong company culture (O’Reilly et al. 1991; Smircich and Morgan 1982; Van
den Steen 2010). In order to mirror the pervasive nature of strong corporate culture in firms,
participants are reminded throughout the experiment to “Think BEAN!” In addition, the slogan is
embossed into the background of the experimental instrument. The text of the “Think BEAN!”
slogan manipulation for each condition is provided in Appendix 1.
In the Less Innovative Culture condition, background information explains how the company
credits its success to a more traditional company culture. In this condition, the “Think BEAN!”
acronym stands for “Beautiful Burgers, Executing Elegance, Ageless Experiences, and Nostalgic
Interactions,” suggesting that more traditional or classic approaches are desired. The background
narrative, entitled “Our Story,” explains that Bean’s Burgers values “time-honored traditions,”
“proven, tried-and-true solutions,” and a “Why reinvent the wheel?” mindset. These phrases rep-
resent a viable approach to business strategy and successful operations, without demonstrating a
commitment to innovation.
In the More Innovative Culture condition, participants read how the company encourages
organizational innovation. The narrative explains that the company prefers “cutting-edge
solutions,” “innovative and creative solutions,” and a “think outside the box” mindset. In this
condition, “Think BEAN!” stands for “Beautiful Burgers, Exemplifying Eccentricity, Adventur-
ous Experiences, and Novel Interactions,” suggesting that the culture encourages a more creative
approach to work. In this condition, the background narrative is entitled, “From Sprout to Stalk:
How We Grew Up” and new employees are referred to as “sprouts.”
After reading background information and learning about “Think BEAN!,” participants
design a burger to submit to the company’s 25th anniversary product initiative. To reflect incen-
tive alignment with company culture, the 10 percent of submissions judged to be most aligned
with the manipulated culture receive a $1.00 monetary bonus. Bonus payments were decided
upon and delivered after the study closed. An example burger submission from each culture is
included in Appendix 2.
The second independent variable, Intervention, is manipulated at three levels: No Interven-
tion, Lower-level Construal Intervention, and Higher-level Construal Intervention. In all condi-
tions, participants are told that their decisions are important and are asked to think about their
plan. The two construal level intervention conditions also include language to prompt participants
to engage in lower- or higher-level construal. The manipulation follows from research in psychol-
ogy showing that “daily” framing prompts decision makers to think about outcomes using lower-
level construal, while “ongoing” framing activates higher-level construal (Chandran and
Menon 2004). Prior work has also shown that attending to the concrete details of a decision acti-
vates lower-level construal, while more abstract language activates higher-level construal
(Liberman and Trope 1998). To strengthen this manipulation, I manipulate whether participants
are told the decision affects each and every restaurant or kitchen (i.e., a concrete representation of
the sum of the underlying units that make up the company—activating lower-level construal), or
the region as a whole, (i.e., an abstract representation of the company taken in aggregate—

activating higher-level construal).13 A complete listing of textual differences between conditions

is included as Appendix 3.
4. Results
Manipulation checks
To assess whether participants deemed the company’s culture to be more or less innovative, all par-
ticipants responded to a question that asked how innovative they perceived the company to be. On
a 0–100 scale anchored at “Not at all” and “Extremely,” participants in the More Innovative Culture
conditions viewed the company to be significantly more innovative than participants in the Less
Innovative Culture conditions (means = 77.1 versus 59.5, p < 0.001, one-tailed).14
To determine whether participants internalized the company culture manipulation, six PhD
students evaluated burger submissions for creativity. Adapting the method used by Kachelmeier
et al. (2008) coders sorted submissions into 10 lots, putting the most creative submissions into lot
“10,” the least creative submissions into lot “1,” and assigning the remaining submissions to lots
“2” through “9” as coder assessment of the submission’s creativity increased.15 Once sorting was
complete, each submission was assigned the value of the lot that it was sorted into and the stack
of submissions was shuffled for the next rater. Cronbach’s alpha for these raters was 0.87, imply-
ing adequate rater agreement. Mean creativity ratings in the More Innovative Culture and Less
Innovative Culture conditions were 5.355 versus 3.792, respectively. This difference is significant
(t = 5.037, p < 0.001, one-tailed, untabulated). As submissions in the More Innovative Culture
conditions were judged to be significantly more creative than submissions in the Less Innovative
Culture conditions, this indicates a successful company culture manipulation.
To assess the intervention manipulation effectiveness, free responses to a question asking
participants to describe how they came to their decision were coded. Coding followed the proce-
dure used by Liberman and Trope (1998). This technique uses coders to examine the syntactical
structure of open-ended participant responses. By classifying the response structure, researchers
can gain insight into the construal level activated at the time of the response. Liberman and
Trope (1998) demonstrate that when higher-level construals are activated, participants are more
likely to describe activities using a description by activity form. For example, a description by
activity description of “reading a book” could be “broadening my horizons by reading.” On the
other hand, participants using lower-level construals are more likely to respond using an activity
by description syntactical form. Using the book-reading example, a lower-level construal consis-
tent response could be, “I read a book by flipping pages.” Following this coding scheme, decision
explanations were coded “1” when the response suggested higher-level construal (description by
activity) and coded as “0” when the response suggested lower-level construal (activity by descrip-
tion). Pearson’s chi-square test reveals that the difference between the lower- and higher-level
13. Prior accounting research informed by CLT (e.g., Backoff et al. 2018; Elliott et al. 2017) has employed more
psych-based construal level manipulations. While this may be a cleaner manipulation, manipulations like these
would not be able to be implemented jointly with the culture manipulation. For example, a common psych-based
construal level manipulation has participants consider how to complete a task, such as reading a book, versus why
to complete a task (Liberman and Trope 1998). If this method was employed in my experiment, it would have to
occur immediately prior to the spending decision and I would be unable to remind participants of the company’s
culture. However, by using a more practical intervention, and then assessing the effectiveness of the intervention
using a post-experimental manipulation check, I am able to manipulate construal level within the manipulated cul-
ture and still assess the success of the manipulation. Manipulation check results are discussed in the following
section of the paper.
14. All statistical analyses herein were performed using the R language and environment for statistical computing
(R Core Team 2019).
15. More specifically, the method used in this study mirrors the procedure that Kachelmeier et al. (2008) used as an ini-
tial creativity assessment in order to expedite payment to participants. Kachelmeier et al. (2008) used nine raters
who provided creativity assessments via radio frequency response devices for their main analysis.

TABLE 1
The effects of creative culture and intervention on repair expense deferral: Descriptive statistics
Mean (SD) Budgeted repairs spending deferred into the next year in dollars
No Lower-level Higher-level
intervention construal construal Row means
More innovative culture Cell 1 Cell 2 Cell 3

Mean 1,754,997 1,133,988 771,905 1,221,566
(SD) (1,281,729) (1,164,124) (825,239) (1,165,085)
[n] [23] [22] [23] [68]
Less innovative culture Cell 4 Cell 5 Cell 6
Mean 1,071,420 899,002 1,401,890 1,114,317
(SD) (937,066) (709,199) (1,144,098) (947,272)
[n] [22] [25] [22] [69]
Column means 1,420,804 1,008,996 1,079,898 1,167,550
(SD) (1,166,156) (946,117) (1,032,795) (1,058,438)
[n] [45] [47] [45] [137]
Notes: Dependent variable: Budgeted repairs spending deferred into the next year. Participants made
spending decisions for 2014 ranging from $0 to $3,500,000, with the remainder of the $3,500,000 being
deferred to 2015. Independent variables: Company Culture is manipulated as participants’ immersion in a
more or less innovative company culture. Intervention manipulated whether participants received (i) no
intervention, (ii) a lower-level construal intervention, or (iii) a higher-level construal intervention.
construal groups is significant in the predicted direction (χ 2 = 2.779, p = 0.047, one-tailed,

untabulated).16 This suggests a successful intervention manipulation.
Hypothesis testing
Hypothesis 1 predicts that, absent an intervention, managers in more innovative company cultures
will engage in more REM than do managers in less innovative company cultures. I use a planned
contrast to test this prediction, with weights of +1 for the More Innovative Culture/No Interven-
tion condition, −1 for the Less Innovative Culture/No Intervention condition, and 0 for all other
conditions.17 As shown in Table 1, mean deferred spending is $1,754,997 versus $1,071,420 in
the More Innovative Culture/No Intervention condition and the Less Innovative Culture/No
Intervention condition, respectively. Tested in Table 2, panel B, this difference is significant
(t131 = 2.24, p = 0.014, one-tailed), supporting Hypothesis 1.18 This provides support for my
16. Means of coded responses in the No Intervention, Lower-level Construal Intervention, and Higher-level Construal
Intervention conditions were 0.43, 0.30, and 0.47, respectively.
17. All pairwise comparisons in the paper are presented absent correction for increased Type I error that arises due to
multiple comparisons (Myers et al. 2013). Results throughout the paper are robust to conducting the analyses as t-
tests with an unpooled error term. The mean difference in p-values between the techniques is −0.02, and all results
found to be significant at an alpha level of 0.05 or 0.10 using a pooled error term remain significant at an alpha
level of 0.10 or lower. In addition, no results found to be statistically insignificant (p > 0.10) become statistically
significant with this change.
18. Levene’s test for homogeneity of variance indicates that the lack of heteroskedasticity assumed by the ANOVA
procedure is not met with these data (F = 2.544, p = 0.031, one-tailed). To assess the impact of heteroskedasticity
on inference, I reconducted all hypothesis testing and additional analysis using a linear model with a
heteroskedasticity-consistent covariance matrix (White 1980). In addition, I performed a Freeman and Tukey (1950)
pffiffiffiffiffiffiffiffiffi pffiffiffi
transformation, of the form x + 1 + x, designed to equate variances across cells, and reexamined both the overall
ANOVA and planned contrast for Hypothesis 1. In both cases, results were qualitatively similar and inferences
remain unchanged.

TABLE 2
The effects of creative culture and intervention on repair expense deferral: ANOVA, hypothesis testing, and
pairwise comparisons
Panel A: Analysis of variance (3×2)
Source df F-statistic p-value
Company Culture 1 0.30 0.584

Intervention 2 1.94 0.148
Company Culture×Intervention 2 4.78 0.009
Error 131
Panel B: Hypothesis testing
Hypothesis Test statistic p-value*
H1: Within the No Intervention condition: More innovative culture > less
innovative culture t131 = −2.24 0.014
H2: Custom contrast with weights of (+7, +3, −1, −1, −3, −5) F1,131 = 4.08 0.044
Residual between-cells variance F4,131 = 2.36 0.040
Contrast variance residual, q2 q2 = 0.72
Panel C: All pairwise comparisons (conducted with 131 error degrees of freedom)†
Comparison t-statistic p-value Comparison t-statistic p-value Comparison t-statistic p-value
Cell 1 vs. 2.03 0.044 Cell 2 vs. 1.19 0.238 Cell 3 vs. −0.43 0.668
Cell 2 Cell 3 Cell 5
Cell 1 vs. 3.25 0.002 Cell 2 vs. 0.20 0.840 Cell 3 vs. −2.06 0.040
Cell 1 vs. 2.24 0.028 Cell 2 vs. 0.78 0.434 Cell 4 vs. 0.58 0.566
Cell 1 vs. 2.89 0.006 Cell 2 vs. −0.87 0.388 Cell 4 vs. 1.07 0.288
Cell 1 vs. 1.16 0.250 Cell 3 vs. −0.98 0.330 Cell 5 vs. 1.68 0.048†
intervention, (ii) a lower-level construal intervention, or (iii) a higher-level construal intervention. *All
hypotheses are tested within the ANOVA. Hypothesis 1 is tested with a planned contrast using weights of
+1 for the No Intervention / More Innovative Culture condition, −1 for the No Intervention / Less
Innovative Culture condition, and 0 for all other cells. This test is presented as one-tailed because it is a
clear directional prediction. Hypothesis 2 is tested using a custom contrast following the procedure given in
Guggenmos et al. (2018). I use weights of +7 for the More Innovative Culture / No Intervention condition,
+3 for the More Innovative Culture / Lower-level Construal Intervention condition, −1 for the More
Innovative Culture / Higher-level Construal Intervention condition, −1 for the Less Innovative Culture / No
Intervention condition, −3 for the Less Innovative Culture / Lower-level Construal Intervention condition,
and −5 for the Less Innovative Culture / Higher-level Construal Intervention condition. †All pairwise
comparisons based on cells identified in Table 1 are presented as two-tailed tests. Note that the difference
between Cells 5 and 6 is significant, but in the direction opposite of my prediction.

hypothesis that, in the absence of any intervention, managers in a more innovative corporate cul-
ture engage in more REM than do managers in a less innovative corporate culture.
My second hypothesis tests the effectiveness of two interventions in reducing REM. My
theory implies an ordinal interaction with a specific pattern of cell means. As shown in Table 2,
panel A, the ANOVA interaction term is significant (F2,131 = 4.78, p = 0.009, two-tailed). I use
custom contrast coding to test for the specific form of interaction predicted and shown in
Figure 3, panel A. Following the procedure outlined by Guggenmos et al. (2018), I derive a con-
trast weighting based on a priori theory, evaluate the statistical significance of the contrast and
between-cells residual, assess visual fit, and evaluate the contrast residual variance (q2).
The predicted pattern of means in Figure 3, panel A, arises from the theoretical prediction that
the difference in REM behavior between more and less innovative corporate cultures is greatest
absent intervention, lesser with a lower-level construal intervention, and least when a higher-level
construal intervention is used. This suggests two downward-sloping lines with a steeper negative
slope in the More Innovative Culture conditions, as given Hypothesis 1 is supported, there is
more REM available to reduce in the More Innovative Culture conditions. This implies a custom
contrast coding of +7 for the More Innovative Culture / No Intervention condition, +3 for the More
Innovative Culture / Lower-level Construal Intervention condition, −1 for the More Innovative
Culture / Higher-level Construal Intervention condition, −1 for the Less Innovative Culture / No
Intervention condition, −3 for the Less Innovative Culture / Lower-level Construal Intervention con-
dition, and −5 for the Less Innovative Culture / Higher-level Construal Intervention condition.19 As
shown in Table 3, panel B, this contrast is significant (F1,131 = 4.08, p = 0.044, two-tailed). How-
ever, the between-cells residual variance is also significant (F4,131 = 2.36, p = 0.040, two-tailed) and
the contrast variance residual is 0.72, meaning that approximately 72 percent of the explainable vari-
ance is not explained by the predicted contrast. And finally, a comparison of panels A and B of
Figure 3 shows that, while the observed pattern of means for Cells 1, 2, 3, 4, and 5 fits the predicted
pattern, the result found in Cell 6 does not. Therefore, an overall visual fit between the predicted and
hypothesized pattern is not supported. Taken together, these results show that while there is support
for an interaction, it is not of the form predicted. Therefore, Hypothesis 2 is not supported.
While the predicted pattern as a whole is not supported, it does appear that the construal interven-
tions acted as predicted in the More Innovative Culture conditions (Cells 1, 2, and 3). That is, within
the More Innovative Culture conditions, the greatest REM occurred in the No Intervention condition,
less REM occurred in the Lower-level Construal Intervention condition, and the least REM occurred
in the Higher-level Construal Intervention condition. As reported in Table 1, the means for these con-
ditions were $1,754,997, $1,133,988, and $771,905, respectively. I conduct a Jonckheere–Terpstra
Test for ordering of cell means within the More Innovative Culture conditions, which confirms the
predicted pattern in the More Innovative Culture conditions (J = 609.5, p = 0.035, untabulated).20
I supplement my test of Hypothesis 2, and further explore the unexpected result in the Less
Innovative Culture conditions, by examining all pairwise cell comparisons. These results are
given in Table 2, panel C. As I predict, within the More Innovative Culture conditions, differ-
ences in REM are significant when comparing the No Intervention condition and the Lower-level
Construal Intervention condition (t131 = 2.032, p = 0.022, one-tailed) as well as the No
19. Guggenmos et al. (2018) note that “when unwieldy contrast weights are necessary,” this may provide a signal that
the assumptions being made may not be justified by theory. However, I note that while the [+7, +3, −1, −1, −3,
−5] contrast coding may appear arbitrary or unwieldy, this coding minimizes assumptions about relative magni-
tudes, given that the contrast coding must reflect the differential slopes predicted for the two lines. This suggests
that this coding is the least presumptuous available, given the specificity of the prediction provided by theory, the
complexity of the experimental design, and the requirement that the contrast coding sum to zero.
20. The exact Jonckheere–Terpstra test requires n < 100 and no ties to be present in the data (Jonckheere 1954). As ties
exist in these data, the Jonckheere–Terpstra statistic is calculated as a permutation test. This test result reported was
conducted with 10,000 permutations for the reference distribution, but inferences remain unchanged with permuta-
tions of 1,000 and 5,000 as well.

Figure 3 The effect of creative culture and intervention on repair expense deferral
Panel A : Predicted results

More Innovative Culture Less Innovative Culture
Budgeted repairs spending deferred into the next year ($)
No Intervention Lower−level Construal Higher−level Construal

Intervention
Panel B: Observed results

Budgeted repairs spending deferred into the next year ($)
$2,000,000
$1,500,000
$1,000,000
$500,000
No Intervention Lower−level Construal Higher −level Construal

Intervention
Notes: Panel A presents a graph of the predicted mean repairs expense deferred into the following year for
each of the six experimental conditions in my experiment. Panel B presents the observed results. Dependent
variable: Budgeted repairs spending deferred into the next year. Participants made spending decisions for
2014 ranging from $0 to $3,500,000, with the remainder of the $3,500,000 being deferred to 2015.
Independent variables: Company Culture is manipulated as participants’ immersion in a more or less
innovative company culture. Intervention manipulated whether participants received (i) no intervention, (ii) a
lower-level construal intervention, or (iii) a higher-level construal intervention.

TABLE 3
The effects of creative culture and intervention on repair expense deferral: Additional analysis
Panel A: ANOVA (2×2)—No Intervention conditions excluded
Source df F-statistic p-value
Company Culture 1 0.95 0.332

Intervention 1 0.12 0.729
Company Culture×Intervention 1 4.55 0.036
Error 88
Panel B: Tests of simple effects F1,90 p-value
SME of company culture within the lower-level construal intervention 0.69 0.409
SME of company culture within the higher-level construal intervention 4.77 0.032
SME of construal intervention within the less innovative company culture 8.97 0.089
SME of construal intervention within the more innovative company culture 1.57 0.214
intervention, (ii) a lower-level construal intervention, or (iii) a higher-level construal intervention. However,
the “no intervention” level of the Intervention factor is excluded from this analysis.
Intervention condition and the Higher-level Construal Intervention condition (t131 = 3.253,
p = 0.001, one-tailed). In addition, mean expense deferral in the Higher-level Construal Interven-
tion condition is lower when compared to the Lower-level Construal Intervention condition
(t131 = 1.185, p = 0.119, one-tailed). Even though this comparison is not statistically significant, this
difference is directionally consistent with my theory. However, comparisons between cells in the
Less Innovative Culture conditions are not significant in the direction predicted (all p values > 0.14,
one-tailed).21
Finally, I conduct a post hoc 2×2 between-participants ANOVA after removing observations
from the No Intervention conditions. This analysis, included as Table 3, panel A, shows a signifi-
cant interaction of Company Culture and Intervention (F1,88 = 4.55, p = 0.036, two-tailed). Sim-
ple main effects tests presented in Table 3, panel B, show that, within the Lower-level Construal
Intervention conditions, a significant effect of company culture is not supported (F1,90 = 0.69,
p = 0.409, two-tailed). However, within the Higher-level Construal Intervention conditions, the
21. In addition, I conduct two robustness checks. First, it is hard to imagine that managers would not attempt to defer
at least some expense to the following year in order to appear responsive to pressures to meet targets, yet a deferral
of the entire budget could signal to employees that the company does not care about repairs and maintenance or
might raise a red flag to auditors who consider REM to be a potential negative signal of management tone
(Commerford et al. 2016). Therefore, I reconduct the test of Hypothesis 1 excluding participants that either defer
the entire $3.5 million or engage in no expense deferral. Exclusion of these participants reduces the participant
count from 137 to 98 observations, limiting statistical power. However, results are consistent with the full sample.
Repeating the test of Hypothesis 1, the difference remains significant (t92 = 4.05, p = 0.023, one-tailed,
untabulated), supporting Hypothesis 1. These results demonstrate that the study’s inferences are not dependent on
including participants that defer all or no budgeted spending into the following year. I also conduct a binary logistic
regression with a “meet or beat” variable coded “one” if the expense deferral allows the company to meet or beat
the firms, and “zero” otherwise. Repeating the test of Hypothesis 1, this test remains significant (p = 0.05, one-
tailed, untabulated). This analysis provides evidence consistent with managers using REM to achieve the EPS
benchmark. As Hypothesis 2 is unsupported, I do not reconduct this test.

effect of company culture is significant (F1,90 = 4.77, p = 0.032, two-tailed), but in the direction
opposite of my initial prediction.
Unexpected effects of construal level interventions in less innovative company cultures

Even though the study’s primary purposes are to demonstrate that more innovative cultures can lead
to greater REM and to investigate interventions designed to curb this behavior within more innova-
tive cultures, the finding that the higher-level construal intervention led to greater REM in the Less
Innovative Corporate Culture condition is surprising and should not be ignored. While post hoc, a
further review of prior research does offer some suggestions as to why this may have occurred.
While prospect theory predicts that individuals are generally risk-averse (Kahneman and
Tversky 1979), more recent work has shown that motivations surrounding risk-seeking behavior
can be more complex. When facing loss, predictions of risk-seeking behaviors differ depending
on whether individuals are motivated more by status quo preservation or an interest in advance-
ment, growth, and the pursuit of ideals (Schneider and Lopes 1986; Scholer et al. 2010).
When considering these orientations with respect to a more or less innovative culture, it is
important to consider how motivational differences for preserving the status quo or pursuing
growth and advancement may differ. Less innovative (i.e., more traditional) company cultures
have been associated with reliability and predictability (O’Reilly et al. 1991), which is consistent
with a motivation to maintain the status quo. On the other hand, more innovative (i.e., less tradi-
tional) corporate cultures are associated with novelty, aspiration, and innovation, suggesting moti-
vation to move the firm forward by rejecting the status quo (O’Reilly et al. 1991). These
associations suggest that fundamental differences in values between corporate cultures may imply
differences in what motivates decision makers to engage in risk-seeking behavior.
With these differences in mind, I examine how this prior research alters my predictions of
construal level intervention effects in this setting. Analogous to my initial prediction for more
innovative corporate cultures, when individuals in less innovative corporate cultures are presented
with lower-level construal interventions, downside risks are made more salient (Chandran and
Menon 2004). This implies that lower-level construal interventions should not affect more and
less innovative corporate cultures differently.
However, one way that higher-level construal interventions are expected to reduce REM is by
making a decision’s “big-picture” impact more salient (Rogers and Bazerman 2008; Trope and
Liberman 2010). In more innovative cultures, considering the “big-picture” consequences, with
respect to the company’s goals and values, allows for the possibility that missing an earnings target
in the current period may still mean that the company has advanced, grown, and performed in line
with its ideals. This may be more important than preserving the status quo of consistent perfor-
mance. However, in less innovative cultures, where culture communicates that preserving the status
quo is highly desirable, individuals considering the “big picture” may be more likely to fixate on
preserving the status quo of consistent earnings performance. This is in line with prior research that
finds that individuals who wish to preserve the status quo display a preference toward taking risk
when it is a necessity for avoiding loss or failure (Molden and Higgins 2004; Scholer et al. 2010).
In my setting, this would suggest higher levels of risk acceptability, and therefore REM, when man-
agers in less innovative corporate cultures are prompted with higher-level construal interventions.
This description is consistent with the pattern of results I observe in my experiment, but it should
be interpreted with caution due to its post hoc nature. With that said, I provide evidence consistent
with this description in the additional analysis section below.
Additional analyses—Should versus want

While prior literature has chronicled how and when managers use REM strategies (Cohen
et al. 2008; Dichev et al. 2013), little has been done to understand managers’ motivations to
engage in REM. On one hand, managers may feel that they should forgo expense deferral to keep
from increasing the firm’s near-term risk, even if it will delay their own gain. However, they may

Figure 4 The effect of creative culture and intervention on should versus want to repair
Panel A: "Should" results

Extent to which participants feel that they should spend
100
on repairs and maintenance (0−100)
90
80
70
60
50

Intervention
Panel B: "Want" results

100
Extent to which participants want to spend
on repairs and maintenance (0−100)
90
80
70
60
50

Intervention
Notes: This figure presents a graph of the mean response to a question assessing the strength of participants’
desire to spend on repairs in the current period. Dependent variable—panel A: Response to question: “When
you made your kitchen update spending authorization decision, how strongly did you feel: I should spend
on kitchen updates soon (2014)? Dependent variable—panel B: Response to question: “When you made
your kitchen update spending authorization decision, how strongly did you feel: I want to spend on kitchen
updates soon (2014)? Independent variables: Company Culture is manipulated as participants’ immersion in
a more or less innovative company culture. Intervention manipulated whether participants received (i) no
intervention, (ii) a lower-level construal intervention, or (iii) a higher-level construal intervention.

also want the instant gratification that hitting a bonus target can provide, even if it subjects the
firm to greater overall risk. Tension that occurs as people consider what they want, usually in the
moment, and what they should do, usually for the long term, has been investigated as the “multi-
ple selves” phenomenon (Rogers and Bazerman 2008), and provides a framework relevant for
understanding the motivations that underlie managers’ REM decisions.
To investigate the multiple selves phenomenon in this setting, post-experimental questions
asked participants, “how strongly did you feel ‘I should spend on kitchen updates soon?’” and
“how strongly did you feel ‘I want to spend on kitchen updates soon?’” Participants recorded their
responses on a 101-point scale, where 0 = “Not at all” and 100 = “Very strongly.” A plot of mean
responses by condition is included as Figure 4, panels A and B. Participants in the No Intervention
condition report feeling a significantly greater obligation to upgrade the kitchens in the current year
than a desire to do so (means = 79.36 versus 68.73, p = 0.013, one-tailed, Table 4, panel C).
Because managers perceived updating the kitchens as more of an obligation than a desire, this sug-
gests a conflict between what they feel they should do and what they want to do. Even though par-
ticipants recognized more of an obligation than a want related to completing current period repairs,
they chose to forgo that obligation and defer repairs and maintenance expense anyway. This result
is consistent with conceptualizing REM as self-interested behavior.
Theory also predicts that one reason that managers in more innovative cultures may engage
in greater REM is because creative mindsets increase the appeal of self-interested behavior. If this
is the case, I expect that participants in more innovative cultures would be less likely to want to
make bonus-reducing repairs, as creative culture increases temptation. Consistent with theory,
absent an intervention, participants in the More Innovative Culture condition report a significantly
lower want to make current period repairs, as compared to participants in the Less Innovative
Culture condition (means = 60.04 versus 77.82, p = 0.016, one-tailed, Table 4, panel C).22 Con-
sistent with my theory, this implies that creative culture increases the desirability of self-interested
behavior.
In addition, research has shown that higher-level construals can guide people to make more
long-term focused decisions (Rogers and Bazerman 2008). If higher-level construal interventions
reduce REM in more innovative cultures by making self-interested behavior less appealing, then I
would expect these interventions to increase the extent to which participants feel that they want
to complete bonus-reducing repairs in the current period. Consistent with this expectation, partici-
pants in the More Innovative Culture/Higher-level Construal Intervention condition, as compared
to the More Innovative Culture / No Intervention condition, were more likely to report that they
want to make current period repairs (means = 83.48 versus 60.04, p = 0.002, one-tailed, Table 4,
panel C). To the extent that lower-level construals make the impacts of managers’ operational
choices and downside risk more salient, I would expect that these interventions also cause these
participants to be more likely to feel that they want to make bonus-reducing repairs in the current
year, in order to reduce downside risk. Consistent with this reasoning, participants in the More
Innovative Culture/Lower-level Construal Intervention condition, as compared with the More
Innovative Culture/No Intervention condition, are more likely to want to make current period
repairs (means = 79.27 versus 60.04, p = 0.020, one-tailed, Table 4, panel C).
In addition to providing insight into intervention efficacy in more innovative corporate cultures
where results were as predicted, further analysis may shed light on the unexpected finding that a
higher-level construal intervention increased REM in the Less Innovative Culture condition. Visual
examination of the pattern of results in Figure 4 suggests that, like the unexpected result found in
terms of the REM-dependent variable, higher-level construal interventions appear to affect man-
agers’ motivations to engage in REM differently in more versus less innovative cultures. A post hoc
t-test confirms this observation, as the difference between the Less Innovative Culture / Higher-level
22. This comparison, as well as the remaining comparisons in the paper, was made using Welch’s t-test as the homoge-
neity of variance assumption was not met with these data (Myers et al. 2013).

TABLE 4
The effects of creative culture and intervention on should versus want reporting
Panel A: Mean (SD) response to should measure
No intervention Lower-level construal Higher-level construal Row means
More innovative culture Cell 1A Cell 2A Cell 3A

Mean 75.22 76.59 94.30 82.12
(SD) (25.95) (32.16) (7.05) (25.28)
[n] [23] [22] [23] [68]
Less innovative culture Cell 4A Cell 5A Cell 6A
Mean 83.68 79.80 69.77 77.84
(SD) (20.08) (23.93) (28.83) (24.84)
[n] [22] [25] [22] [69]
Column means 79.36 78.30 82.31 79.96
(SD) (23.39) (27.82) (23.99) (25.06)
[n] [45] [47] [45] [137]
Panel B: Mean (SD) response to want measure
No intervention Lower-level construal Higher-level construal Row means
More innovative culture Cell 1B Cell 2B Cell 3B

Mean 60.04 79.27 83.48 74.19
(SD) (31.99) (29.21) (19.39) (28.88)
[n] [23] [22] [23] [68]
Less innovative culture Cell 4B Cell 5B Cell 6B
Mean 77.82 73.36 67.18 72.81
(SD) (20.67) (28.37) (28.19) (26.07)
[n] [22] [25] [22] [69]
Column means 68.73 76.13 75.51 73.50
(SD) (28.22) (28.61) (25.20) (27.41)
[n] [45] [47] [45] [137]
Panel C: Planned comparisons t-statistic p-value*
Within no intervention—Feelings of should versus want (Cell 1A and 4A versus Cell 1B

and 4B) 2.30 0.013
Within no intervention—Feelings of want: More versus less innovative culture (Cell 1B
versus 4B) −2.22 0.016
Within more innovative culture—Feelings of want: Higher-level construal versus no
intervention (Cell 3B versus 1B) 3.00 0.002
Within more innovative culture—Feelings of want: Lower-level construal versus no
intervention (Cell 2B versus 1B) 2.11 0.020
Notes: Dependent variables: (should / want) measure is participants’ response to how strongly did you feel “I
(should / want to) spend on kitchen updates soon?” on a 0–100 scale, where 0 = “Not at all” and 100 = “Very
strongly.” Independent variables: Company Culture is manipulated as participants’ immersion in a more or less
innovative company culture. Intervention is manipulated whether participants received (i) no intervention, (ii) a
lower-level construal intervention, or (iii) a higher-level construal intervention. *Given the directional
expectations suggested by my theory, these tests are one-tailed. The first comparison was conducted as a
paired within-subjects t-test. All other tests were conducted as a Welch’s two sample t-test as these tests are
between participants and the homogeneity of variance assumption is not supported throughout the data set.

Construal Intervention condition and the More Innovative Culture/Higher-level Construal Interven-
tion condition on both the should and want measures are statistically significant, but opposite of the
direction predicted (means = 94.30 versus 69.77, p = 0.001, two-tailed, untabulated and
means = 67.18 versus 83.46, p = 0.020, two-tailed, untabulated, respectively). To investigate this
further, I perform additional analysis within the Less Innovative Culture conditions.
As discussed earlier, one explanation for the increase in REM within the Less Innovative Culture
/ Higher-level Construal Intervention condition is because, after being prompted to consider the “big
picture,” managers view REM as a necessity for avoiding loss and for preserving the status quo
(Molden and Higgins 2004; Scholer et al. 2010). If this is the case, I expect that the extent that partic-
ipants feel they “should” spend on repairs and maintenance will be (i) lower in the Less Innovative
Culture/Higher-level Construal Intervention condition as compared to the Less Innovative Culture/No
Intervention condition and (ii) lower in the Less Innovative Culture/Higher-level Construal Interven-
tion condition as compared to the More Innovative Culture/Higher-level Construal Intervention condi-
tion. These comparisons are marginally significant (t = 1.66, p = 0.10, one-tailed, untabulated) and
significant (t = 8.90, p < 0.01, one-tailed, untabulated), respectively. In addition, I would expect no
differences between managers in More versus Less Innovative Culture condition within the lower-
level construal conditions on either the “should” or the “want” measure, as my post hoc expectation
also assumes that, within both cultures, the lower-level construal works via the same mechanism of
highlighting downside risk. These differences are not significant on the “should” (t = 0.48, p = 0.49,
two-tailed, untabulated) or “want” measure (t = 0.17, p = 0.69, two-tailed, untabulated), as expected.
Finally, my explanation suggests that higher-level construals work to make the “big-picture”
impact of manager’s decisions more salient in both more and less innovative cultures. However,
I propose that the big picture is focused more on “preserving the status quo” in less innovative cul-
tures and on ideals of “growth and moving forward” in more innovative cultures. To test this theory,
I coded free responses to a question asking participants to describe how they made their decision. I
coded responses as “one” if the response was justified based on concerns about the customer, health
and safety, or food-borne illness, as these represent customer focus and the long-term growth of the
company, and “zero” if the response was based on concerns about investors, EPS, or Wall Street,
representing a focus on short-term preservation of status quo performance. If both categories were
present, I coded responses based on the category listed first or on participant rank ordering, if avail-
able. If higher-level construal engenders status quo preservation in less innovative cultures, but
prompts status quo rejection in more innovative cultures, I expect an increase in customer focus from
the No Intervention condition to the Higher-level Construal Intervention condition within the more
innovative culture because participants are focused more on “big-picture” ideals. However, I expect
no change in customer focus from the No Intervention condition to the Higher-level Construal Inter-
vention conditions in the less innovative culture. Consistent with this expectation, this difference is
significant in the More Innovative Culture conditions (t = 1.37, p = 0.04, one-tailed, untabulated),
but not significant in the Less Innovative Culture conditions (t = 0.13, p = 0.55, two-tailed,
untabulated). This provides support consistent with my post hoc explanation that higher-level con-
struals may increase desire to preserve the status quo in less innovative corporate cultures, potentially
explaining increased REM in the Less Innovative/Higher-level Construal Intervention condition.
5. Conclusion
As executives pursue innovation in their firms, investigating unintended consequences of creative
culture on decision making is an important research area. Because innovative corporate culture may
increase risk acceptability and the desirability of self-interested behavior, some of the benefits of
these cultures could become eclipsed by costs. Recognizing this relationship has implications for
several different groups. For instance, understanding this relationship can help management under-
stand how culture affects organizational decision making. Furthermore, as REM is often detrimental
to long-term firm value, investors may benefit from knowing that innovation-focused organizations
may be more likely to engage in myopic behaviors.

However, this study’s main contribution is the demonstration of two interventions that reduce
REM in more innovative corporate cultures. My results demonstrate that prompting managers in
these cultures to consider the day-to-day impact of their decisions may reduce REM by highlight-
ing downside risk, and encouraging them to consider “big-picture” impacts of their decisions may
reduce REM even more. These interventions could be adapted to other settings where managers
are expected to act in the best interest of the company but simultaneously face incentives to
expose the company to greater risk or to make decisions contrary to the best interest of the firm.
This study also makes several contributions to the accounting and psychology literatures. To
my knowledge, this is the first study to investigate negative consequences of a desirable company
culture on financial reporting. This extends the tone-at-the-top literature to acknowledge that perva-
sive company initiatives do not need to be outwardly malevolent to have negative effects. Second,
this study adds to the emerging literature on the “dark side” of creativity, investigating these effects
in an applied setting. This study’s findings, combined with prior research, imply that even when
creativity incentivization is effective (Kachelmeier et al. 2008; Kachelmeier and Williamson 2010;
Chen et al. 2012; Grabner 2014), these incentives may have unintended consequences.
In addition, this study provides evidence that higher-level construal-based interventions may
not be universally effective in reducing REM, as I unexpectedly find that these interventions
increased REM in a less innovative company culture. Post hoc analysis suggests that this may be
due to an increased focus on status quo preservation and elevated risk preferences that occur
when accepting risk is necessary to avoid loss or failure in these cultures. Given its post hoc
nature, this explanation should be interpreted cautiously. Regardless, the effects of corporate cul-
ture appear to be more complicated than expected. This provides additional opportunity to investi-
gate how corporate cultures change decision making and information evaluation.
This research is subject to inherent limitations. As mentioned above, I examine a single facet
of company culture. While I consider this dimension to be important, company culture is multiface-
ted. In addition, managers in the More Innovative Culture conditions may have felt that the burger
creation task was more difficult than did the participants in the Less Innovative Culture condition.
This could lead to feelings of entitlement, which may provide justification for engaging in REM.
Although I do not believe this is the case, I cannot rule it out. However, this explanation would not
explain the efficacy of the intervention or the result in the Less Innovative Culture/Higher-level
Construal Intervention condition. Additionally, this study’s setting involves a decision where REM
leads to an increase in risk that is made salient to the decision maker. These findings may not gener-
alize to settings where the relationship between REM and risk is less salient. However, this limita-
tion should principally affect the magnitude, not existence, of the effect. Finally, my intervention
conditions emphasize decision consequences to a greater degree than the no intervention conditions.
This design choice was made to increase the effectiveness of the manipulation but makes me unable
to definitively state whether more generic construal level interventions that did not focus managers
on the consequences of their actions would have been as effective.
This study suggests several avenues for future research. Innovation is only one dimension of
corporate culture identified by O’Reilly et al. (1991). Future research could investigate additional
cultural impacts on financial reporting (Graham et al. 2016). In addition, researchers could inves-
tigate how creative culture affects external parties exposed to the company’s culture, such as audi-
tors or investors. Finally, psychologists could further investigate the interplay between creativity
and construal. Research in psychology finds that higher-level construals are associated with
enhanced creativity in laboratory tasks (Förster et al. 2004). However, in my setting, I find that
higher-level construal-based interventions can reduce opportunistic behavior. These findings,
taken together, imply that construal effects may interact with decision tasks. Finally, with the
unexpected result I observe in the less innovative corporate culture conditions, I provide initial
evidence that corporate culture may affect how psychological processes affect judgment and
information processing, but this evidence is preliminary. With these questions in mind, this area
should be a rich avenue for future research for years to come.

Appendix 1
Panel A: Think BEAN! Less Innovative Company Culture

Think BEAN!
Our founder, Aaron Bean, credits our success to the “Think BEAN!” philosophy and its effect on
the company’s culture. By living the “Think BEAN!” mindset at work, we always know what to
do. Think BEAN! is an acronym that stands for beautiful burgers, executing elegance, ageless
experiences, and nostalgic interactions.
Beautiful BURGERS
Like the Mona Lisa, a Bean’s Burger is a work of conventional beauty. Our customers love that
we stick to tradition. Our burgers stand for themselves. We don’t need fancy condiments or new-
fangled gimmicks to make a great burger! To be successful, we just need to do things the way
we’ve been doing them for the past 25 years. Over time, we’ve learned that traditional and con-
sistent approaches are the keys to success.
Executing ELEGANCE
When thinking about hamburgers, most people don’t think about elegance. However, we think
that adding a bit of simple elegance to a night out, whatever the cuisine, keeps customers coming
back. This way of thinking extends beyond our restaurants. As a Bean’s manager, you should
attack problems with elegantly simple, tried-and-true solutions because, even though an idea may
seem simple at first, it just might work.
AGELESS Experiences
The ritual of enjoying a meal out with family and friends has been an affordable luxury in the
best and worst of times. Accordingly, we strive to preserve the ageless nature of the family meal.
Our advertising often focuses on tradition and the timeless joy of getting together with loved
ones. This focus on tradition carries itself to the corporate office, as our managers prefer strategies
that have stood the test of time.
NOSTALGIC Interactions
Customer interactions are the most important part of Bean’s business. Because of this, we focus on
cultivating perfect interactions between diners and restaurant personnel. We believe that the perfect
interaction is one that inspires a patron to reminisce back to the best meal they have ever had and
to compare it to their meal at Bean’s. By providing interactions customers can reflect back on,
Bean’s can make dining nostalgic. Employees take this same approach when solving problems. As
management, we challenge employees to make comparisons between their proposed solution to an
issue and other well-established solutions that are proven to have worked in the past. By reflecting
back on our past successes, we believe that we can keep making great decisions far into the future.
Panel B: Think BEAN! More Innovative Company Culture
Think BEAN!
Our founder, Aaron Bean, credits our success to the “Think BEAN!” philosophy and its effect on
the company’s culture. By living the “Think BEAN!” mindset at work, we always know what to
do. Think BEAN! is an acronym that stands for beautiful burgers, exemplifying eccentricity,
adventurous experiences, and novel interactions.
Beautiful BURGERS
Like the creation of a modern sculptor, a Bean’s Burger is a work of unconventional beauty.
People come to our restaurants for the creative and unexpected—and we love to deliver. We
leave no stone unturned imagining ingenious new burgers for our customers to chow down
on. We are 25 years young and we do not see any reason to grow up now. We embrace

inventiveness, because we have learned that creative and innovative approaches are the keys to
success.
Exemplifying ECCENTRICITY
When thinking about hamburgers, most people do not think about eccentricity. However, we
think that adding a bit of unpredictability to a night out, whatever the cuisine, keeps customers
coming back. This way of thinking extends beyond our restaurants. As a Bean’s manager, you
should attack problems by thinking outside the box because, even though an idea may seem crazy
at first, it just might work.
ADVENTUROUS Experiences
The ritual of enjoying a meal out with family and friends has been an affordable luxury in the
best and worst of times. Accordingly, we strive to incorporate a little adventure into the family
meal. Our advertising often focuses on the fun that accompanies trying something new with loved
ones. This focus on adventure carries itself to the corporate office, as our managers prefer strate-
gies that capture the company’s sense of adventure.
NOVEL Interactions
Customer interactions are the most important part of Bean’s business. Because of this, we focus
on cultivating perfect interactions between diners and restaurant personnel. We believe that the
perfect interaction is one that is uniquely memorable. By providing new and innovative interac-
tions, Bean’s can make dining novel. We take this same approach to solving problems. As man-
agement, we challenge our employees to find new ways to attack issues. Just because a solution
has worked in the past, that does not mean there is not room for improvement. By focusing on
being novel, we believe that we can keep innovating far into the future.
Appendix 2
Burger submission examples
Panel A: Less Innovative Culture Burger Submission Example

Burger Name: The Standard Hamburger
Ingredients:
Hamburger
Tomato
Lettuce
Onion
Mayonnaise
Preparation/Serving Instructions:
Cook medium. Place lettuce, slice of ripe tomato, slice of onion and cover with mayonnaise.
Burger Name: Good Old American Burger
Ingredients:
Hamburger
Bun
Mustard
Ketchup
French Fries
Classically prepared burger, on a white bread bun, served with mustard and ketchup, with per-
fectly prepared French fries on the side.

Panel B: More Innovative Culture Burger Submission Example

Burger Name: Hang Ten
Ingredients:
Lean ground beef seasoned with Asian fish sauce
Grilled shrimp
Fried calamari
One oyster
Curly leaf lettuce
Grilled pineapple
Dollop of tartar sauce
Bun with dusting of stone ground flour
Avocado slice
[The] Hang Ten has 10 ingredients and is served with a surfboard-shaped avocado slice. Curly
lettuce represents waves and dusted bun represents sand.
Burger Name: Billy Beane Burger
Ingredients:
Ground goat meat
One packet onion soup mix
One teaspoon of salt, pepper
All spices
Onion powder
Quarter can of ground chick peas and ground black beans
Half a package of curry
One 1/3 pound burger
Ground 1/3 goat meat, black bean, and chickpeas and throw in bowl. Add all spices except curry
into a bowl of ground meat, peas, and beans. Form patty out of mixture. Cook the curry down in
the pan then add burger and flip liberally.

Text of intervention manipulations

Appendix 3

References
Amabile, T. M. 1988. A model of creativity and innovation in organizations. Research in Organizational
Behavior 10 (1): 123–67.
Amabile, T. M., and S. S. Gryskiewicz. 1987. Creativity in the R&D laboratory. Greensboro, NC: Center
for Creative Leadership.
Backoff, A. G., T. Carpenter, and J. M. Thayer. 2018. Auditing complex estimates: How do construal level
and evidence formatting impact auditors’ consideration of inconsistent evidence? Contemporary Account-
ing Research 35 (4): 1798–815.
Bens, D. A., V. Nagar, and H. F. Wong. 2002. Real investment implications of employee stock options exer-
cises. Journal of Accounting Research 40 (2): 359–93.
Bentley, J. 2019. Decreasing operational distortion and surrogation through narrative reporting. The Account-
ing Review 94 (3): 27–55.
Bloomfield, R. J. 2017. What counts and what gets counted, 2nd ed, https://doi.org/10.2139/ssrn.2899141,
retrieved November 15, 2018.
Chandran, S., and G. Menon. 2004. When a day means more than a year: Effects of temporal framing on
judgments of health risk. Journal of Consumer Research 31 (2): 375–89.
Chen, C. X., M. G. Williamson, and F. H. Zhou. 2012. Reward system design and group creativity: An
experimental investigation. The Accounting Review 87 (6): 1885–911.
Chima, C. 2013. Creative cultures: Mailchimp grants employees “permission to be creative,” http://www.
fastcompany.com/1767793/creative-cultures-mailchimp-grants-employees-permission-be-creative, retrieved
June 1, 2016.
Cohen, D. A., A. Dey, and T. Z. Lys. 2008. Real and accrual-based earnings management in the pre– and
post–Sarbanes-Oxley periods. The Accounting Review 83 (3): 757–87.
Commerford, B. P., D. R. Hermanson, R. W. Houston, and M. F. Peters. 2016. Real earnings management:
A threat to auditor comfort? Auditing: A Journal of Practice & Theory 35 (4): 39–56.
Dichev, I. D., J. Graham, C. R. Harvey, and S. Rajgopal. 2013. Earnings quality: Evidence from the field.
Journal of Accounting and Economics 56 (2): 1–33.
Elliott, W. B., S. M. Grant, and K. M. Rennekamp. 2017. How disclosure features of corporate social
responsibility reports interact with investor numeracy to influence investor judgments. Contemporary
Accounting Research 34 (3): 1596–621.
Ewert, R., and A. Wagenhofer. 2005. Economic effects of tightening accounting standards to restrict earn-
ings management. The Accounting Review 80 (4): 1101–24.
Farrell, A. M., J. H. Grenier, and J. Leiby. 2016. Scoundrels or stars? Theory and evidence on the quality of
workers in online labor markets. The Accounting Review 92 (1): 93–114.
Förster, J., R. S. Friedman, and N. Liberman. 2004. Temporal construal effects on abstract and concrete
thinking: Consequences for insight and creative cognition. Journal of Personality and Social Psychol-
ogy 87 (2): 177–89.
Freeman, M. F., and J. W. Tukey. 1950. Transformations related to the angular and the square root. Annals
of Mathematical Statistics 21 (4): 607–11.
Fujita, K., and J. J. Carnevale. 2012. Transcending temptation through abstraction the role of construal level
in self-control. Current Directions in Psychological Science 21 (4): 248–52.
Fujita, K., Y. Trope, N. Liberman, and M. Levin-Sagi. 2006. Construal levels and self-control. Journal of
Personality and Social Psychology 90 (3): 351–67.
Gino, F., and D. Ariely. 2012. The dark side of creativity: Original thinkers can be more dishonest. Journal
of Personality and Social Psychology 102 (3): 445–59.
Gino, F., and S. S. Wiltermuth. 2014. Evil genius? How dishonesty can lead to greater creativity. Psycholog-
ical Science 25 (4): 973–98.
Grabner, I. 2014. Incentive system design in creativity-dependent firms. The Accounting Review 89 (5):
1729–50.

Graham, J.R., C.R. Harvey, J. Popadak and S. Rajgopal. 2016. Corporate culture: Evidence from the field.
Working paper, Duke University.
Graham, J.R., C.R. Harvey, J. Popadak and S. Rajgopal. 2017. Corporate culture: The interview evidence.
Working paper, Duke University.
Graham, J. R., C. R. Harvey, and S. Rajgopal. 2005. The economic implications of corporate financial
reporting. Journal of Accounting and Economics 40 (1): 3–73.
Guggenmos, R. D., M. D. Piercey, and C. P. Agoglia. 2018. Custom contrast testing: Current trends and a
new approach. The Accounting Review 93 (5): 223–44.
Gunny, K. A. 2010. The relation between earnings management using real activities manipulation and future
performance: Evidence from meeting earnings benchmarks. Contemporary Accounting Research 27 (3):
855–88.
IBM. 2010. Capitalizing on complexity: Insights from the global chief executive officer study. Armonk, NY:
IBM Global Business Services.
Jonckheere, A. R. 1954. A distribution-free k-sample test against ordered alternatives. Biometrika 41 (5):
133–45.
Kachelmeier, S. J., B. E. Reichert, and M. G. Williamson. 2008. Measuring and motivating quantity, creativ-
ity, or both. Journal of Accounting Research 46 (2): 341–73.
Kachelmeier, S. J., and M. G. Williamson. 2010. Attracting creativity: The initial and aggregate effects of
contract selection on creativity-weighted productivity. The Accounting Review 85 (5): 1669–91.
Kahneman, D., and A. Tversky. 1979. Prospect theory: An analysis of decision under risk. Econometrica 47
(2): 263–91.
Kivetz, Y., and T. R. Tyler. 2007. Tomorrow I’ll be me: The effect of time perspective on the activation of ideal-
istic versus pragmatic selves. Organizational Behavior and Human Decision Processes 102 (2): 193–211.
Krische, S. 2019. Investment experience, financial literacy, and investment-related judgments. Contemporary
Libby, R., R. Bloomfield, and M. Nelson. 2002. Experimental research in financial accounting. Accounting,
Organizations and Society 27 (8): 775–810.
Liberman, N., and Y. Trope. 1998. The role of feasibililty and desirability considerations in near and distant
future decisions: A test of temporal construal theory. Journal of Personality and Social Psychology 75
(1): 5–18.
Liberman, N., and Y. Trope. 2008. The psychology of transcending the here and now. Science 322 (5905): 1201–5.
Mayer, D. M., K. Aquino, R. L. Greenbaum, and M. Kuenzi. 2012. Who displays ethical leadership, and
why does it matter? An examination of antecedents and consequences of ethical leadership. Academy of
Management Journal 55 (1): 151–71.
McShane, S. L., and M. A. Y. Von Glinow. 2010. Organizational behavior: Emerging knowledge and prac-
tice for the real world. Boston, MA: McGraw-Hill Irwin.
Miles, R. E., C. C. Snow, A. D. Meyer, and H. J. Coleman. 1978. Organizational strategy, structure, and
process. Academy of Management Review 3 (3): 546–62.
Molden, D. C., and E. T. Higgins. 2004. Categorization under uncertainty: Resolving vagueness and ambi-
guity with eager versus vigilant strategies. Social Cognition 22 (2): 248–77.
Myers, J. L., A. D. Well, and R. F. Lorch. 2013. Research design and statistical analysis. New York, NY:
Routledge.
O’Reilly, C. A., J. Chatman, and D. F. Caldwell. 1991. People and organizational culture: A profile compari-
son approach to assessing person-organization fit. Academy of Management Journal 34 (3): 487–516.
Paolacci, G., J. Chandler, and P. G. Ipeirotis. 2010. Running experiments on Amazon Mechanical Turk.
Judgment and Decision Making 5 (5): 411–19.
Patelli, L., and M. Pedrini. 2013. Is tone at the top associated with financial reporting aggressiveness? Jour-
nal of Business Ethics 126 (1): 3–19.
R Core Team. 2019. R: A language and environment for statistical computing. Vienna, Austria: R Founda-
tion for Statistical Computing, http://www.R-project.org, retrieved March 1, 2019.

Rennekamp, K. 2012. Processing fluency and investors’ reactions to disclosure readability. Journal of
Rogers, T., and M. H. Bazerman. 2008. Future lock-in: Future implementation increases selection of
“should” choices. Organizational Behavior and Human Decision Processes 106 (1): 1–20.
Roychowdhury, S. 2006. Earnings management through real activities manipulation. Journal of Accounting
and Economics 42 (3): 335–70.
Schein, E. H. 2010. Organizational culture and leadership, Vol. 2. San Francisco, CA: John Wiley & Sons.
Schneider, S. L., and L. L. Lopes. 1986. Reflection in preferences under risk: Who and when may suggest
why. Journal of Experimental Psychology: Human Perception and Performance 12 (4): 535–48.
Scholer, A. A., X. Zou, K. Fujita, S. J. Stroessner, and E. T. Higgins. 2010. When risk seeking becomes a
motivational necessity. Journal of Personality and Social Psychology 99 (2): 215–31.
Smircich, L. 1983. Concepts of culture and organizational analysis. Administrative Science Quarterly 28 (3):
339–58.
Smircich, L., and G. Morgan. 1982. Leadership: The management of meaning. Journal of Applied Behav-
ioral Science 18 (3): 257–73.
Trope, Y., and N. Liberman. 2003. Temporal construal. Psychological Review 110 (3): 403–21.
Trope, Y., and N. Liberman. 2010. Construal-level theory of psychological distance. Psychological Review
117 (2): 440–63.
Van den Steen, E. 2010. Culture clash: The costs and benefits of homogeneity. Management Science 56
(10): 1718–38.
Vorst, P. 2016. Real earnings management and long-term operating performance: The role of reversals in
discretionary investment cuts. The Accounting Review 91 (4): 1219–56.
Weisner, M. 2015. Using construal level theory to motivate accounting research: A literature review.
Auditing: A Journal of Practice & Theory 27 (1): 137–80.
White, H. A. 1980. Heteroskedasticity-consistent covariance matrix estimator and a direct test for hetero-
skedasticity. Econometrica: Journal of the Econometric Society 48 (4): 817–38.
Discussion of “The Effects of Creative Culture on Real Earnings

Management”*
WIM A. VAN DER STEDE, London School of Economics†
1. Introduction
This article is a discussion of, but mostly a further reflection on, the 2018 Contemporary Account-
ing Research conference paper by Ryan D. Guggenmos, “The Effects of Creative Culture on Real
Earnings Management” (this issue, henceforth cited as Guggenmos 2020). If I could have (nit)
picked a title for my discussion article, I would have called it “Alleged Unintended Consequences
of Apparently Desirable Company Cultures,” for reasons that will become obvious as I develop
my narrative. As there are two key terms in the title of Guggenmos (2020)—company culture
* Accepted by Jeffrey Hales. I thank Mike Welker for the opportunity to discuss the Guggenmos 2018 Contemporary
Accounting Research conference paper. I thank Ryan Guggenmos for the informal rejoinder following the presentation
of the paper and my comments there in October 2018. I also thank Jeffrey Hales for helpful suggestions and final edits.
† Corresponding author. Wim A. Van der Stede is also a visiting professor at Erasmus School of Economics and Peking
University’s Guanghua School of Management.

Discussion of “The Effects of Creative Culture on REM” 31
and earnings management—my discussion inevitably resolves around these two core themes, and
related, across the multiple sections that follow.
Earnings management comes in two forms: accounting-related earnings management, also
known as accruals-based earnings management, and accounting-triggered earnings management,
which is the type Guggenmos (2020) focuses on—the so-called “real earnings management”
(REM). My definition of REM is any operating action or decision management or employees
take to make accounting earnings look more desired than they otherwise would. Thus, intent is
key—that is, those taking the actions or decisions know that they may not be taking particularly
good actions or decisions, but they take them anyway because of the “desired” effect on earn-
ings.1 Accruals-based earnings management, in contrast, involves changing the actual accounting
numbers directly, either fraudulently or by way of stretching accounting conventions.
Next, I briefly summarize the key tenets and findings of Guggenmos (2020), after which I
will return to the paper’s two key concepts of company culture and earnings management as the
basis for my further reflections.
2. Setup and tension

Upfront in section 1, in the introduction, Guggenmos (2020) essentially argues and demonstrates
that “creativity can increase opportunistic behavior even within the bounds of accurate reporting.”
Two things about this quoted sentence: first, creativity is really the operationalization of what he
more broadly alludes to as an “innovative company culture”; second, “opportunistic behavior
even within the bounds of accurate reporting” clearly establishes his study’s focus on REM by
taking accruals-based earnings management out; that is, even when all the accounting is proper,
innovative company cultures may trigger “opportunistic behavior” or, in my terms, actions and
decisions intended to make earnings look more desirable than they otherwise would.
Pertinently stated in section 1, toward the end of the introduction when elaborating the
study’s contributions, the key tension in Guggenmos (2020) is that “Company leadership
should be aware that cultivating an innovative company culture may have downsides.” Put
bluntly, not everything that creativity produces is good. This is because “even though a comp-
any’s culture (i.e., creativity) may not appear harmful, elements of the culture that increase
risky behavior should be considered” (parentheses added). The key argument thus is that the
potentially riskier behaviors that creativity-based cultures encourage include a greater propen-
sity to engage in REM.
However, here is the rub, which is my favorite part of the setup of Guggenmos (2020) in that
same paragraph of his paper: “as it is inadvisable to suggest that companies forego innovation,
my results provide managers with interventions that can reduce REM behavior within more inno-
vative company cultures.” Thus, the point is about mitigating potentially undesirable conse-
quences within an otherwise desirable company culture. One should indeed not throw out the
baby with the bathwater. To be so clear about this is the refreshing part of the Guggenmos (2020)
paper. Less careful conceptualizing often inadvertently gets bogged down by wanting to eliminate
the bad while overlooking that it also often implies throwing out the good. Guggenmos (2020)
got it right with mitigation, not elimination.
During my Contemporary Accounting Research conference presentation, I illustrated this
subtle but crucial point with the generic graph I drew in Figure 1, implying that indeed there is a
“theoretical optimum” of creativity considering its benefits (upsides) and costs (downsides).
1. It is tempting to say “positive” effect on earnings, but it should really be “desired” because sometimes the benefit
to those taking the actions or decisions accrues from managing earnings up but sometimes also down. Moreover,
the notion of “desired” meshes well with the idea of “intent” being the rationale behind the actions or decisions.

Figure 1 Net effect of Ψ (“creative culture”) on π (“performance”)
Notes: This figure depicts the optimal amount of creativity (Ψ*) considering its benefits (upsides) and costs
(downsides). In this figure, Ψ is optimal where performance (π, equal to the benefits minus the costs of
creativity) is greatest or where the distance between the benefits and costs is the widest (dotted line).
3. Mitigating the downside

Because eliminating innovation is not an option, Guggenmos (2020) conceives of “an inter-
vention that reduces REM within (or given) the firm’s culture” (parenthesis added). For this,
he draws on construal level theory from (organizational) psychology. In my own words, the
core and theory-consistent idea behind construal level interventions for the purposes of
Guggenmos (2020) is to help curb the undesirable REM behaviors while maintaining the
firm’s innovative culture.
Without going into the details, construal level interventions do this (or attempt to achieve
this) by way of “abstraction”—that is, by prompting someone to weigh the consequences of their
actions or decisions before they make them, either in operational terms (lower-level construal) or
in a more strategic way (higher-level construal). Guggenmos (2020) gives examples of this. In
section 6 of this discussion, I will make some comments about what I believe may be issues of
(practical) feasibility and effectiveness of construal-type interventions.
In a nutshell, the idea is that lower-level construal can alleviate REM behaviors by increasing
the salience of downside risk, whereas higher-level construal is expected to do so by focusing
decision makers on the “big-picture” consequences of their decisions and, among other things,
muted preferences for instant gratification and self-interested behavior.
Because there is presumably “more” to higher-level construal (i.e., making big-picture conse-
quences more salient plus decreasing preferences for instant gratification plus some other features)
than there is to lower-level construal (which is characterized by prompting the salience of down-
side risk only), higher-level construal interventions are expected to reduce REM to a greater
extent or “more” than lower-level construal. Let me preamble now though I rather find this a
stretch, aside from the feasibility of using construal-level interventions in practice, which I will
discuss further in section 6.
Aside from this, what I find neat is the idea of trying to “mitigate the downside”—that is,
alleviate REM behaviors that may come with innovative company cultures without however

Figure 2 Post-intervention net effect of Ψ (“creative culture”) on π (“performance”)
Notes: This figure is an overlay of Figure 1. It depicts the mitigation of one of the costs of creativity (REM)
by way of an “intervention” (shown by the arrow). Such a construal level intervention that helps curb
undesirable REM thus pushes the cost curve to the right, and in so doing improves the level of creativity
(from Ψ* to Ψ*0 ) at which performance is maximized.
curtailing creativity itself. I illustrated that at the Contemporary Accounting Research conference
by way of drawing an overlay of Figure 1 as shown in Figure 2.
4. Guggenmos’s (2020) findings

Using an experimental research design, Guggenmos (2020) finds, first, that managers in a
more innovative company culture engage in greater REM than managers in a less innovative
company culture. Second, within the more innovative company culture, he finds that the
intervention that highlights the downside risks of managers’ decisions (i.e., the lower-level
construal intervention) diminishes REM, but the intervention that helps managers see the
big-picture impact of their decision (i.e., the higher-level construal intervention) reduces
REM to a greater extent. However, contrary to his prediction, he also finds that the same
big-picture (higher-level construal) intervention that reduces REM the most in a more inno-
vative company culture increases REM in a less innovative company culture.
Guggenmos (2020) then embarks on a post hoc analysis for this unexpected result. My
bare-bones, maximally reduced (maybe borderline simplistic) interpretation of this rather
complex post hoc analysis is that higher-level construal interventions appear to work—that is,
they manage to prompt the decision makers to consider the big picture; it is just that, as I infer
it, the “big picture” is different in more innovative than in less innovative company cultures.
In more innovative company cultures, the “big picture” is about the decision’s impact on
growth and moving the firm forward and engaging in behavior that is beneficial to the firm in
the long run. But in less innovative company cultures, managers instead appear to view the
“big picture” as preservation of the status quo. This may explain why, in a more innovative
culture, REM is mitigated, because the earnings management is inherently myopic. But, in the

less innovative culture, earnings management is not mitigated but instead exacerbated,
because REM “saves the day” (my phrase) even though, or just because, REM is immediate
in its short-term effect; that is, it helps to meet the target! Or, to quote from Guggenmos
(2020, in “Unexpected effects of construal level interventions in less innovative company cul-
tures” in section 4:
In more innovative cultures, considering the “big picture” consequences, with respect to the
company’s goals and values, allows for the possibility that missing an earnings target in the
current period may still mean that the company has advanced, grown, and performed in line
with its ideals. This may be more important than preserving the status quo of consistent perfor-
mance. However, in less innovative cultures, where culture communicates that preserving the
status quo is highly desirable, individuals considering the “big picture” may be more likely to
fixate on preserving the status quo of consistent earnings performance.
Hence, I think I am correct to suggest that if the “big picture” is different in more innovative
compared to less innovative company cultures, then evoking it should also be expected to have
different outcomes, the first more consistent with growth and the future and the latter more con-
sistent with steady earnings and the present.
Finally, despite my (apparently unfounded) initial reservations, the results do, however, indi-
cate “differential magnitudes” or “increased efficacy” commensurate with the “hierarchy” of
construal-type interventions. In other words, the findings bear out that, where the interventions
performed on the whole as predicted (i.e., in the more innovative culture conditions), the greatest
REM occurred where there was no intervention; less REM occurred where there was a lower-
level construal intervention, and the least REM occurred where the higher-level construal inter-
vention took place. Thus, I rest my case (or maybe not yet, see later, in section 6).
5. The bigger picture

With all the key features of Guggenmos (2020) now hopefully reasonably manageably summa-
rized, let me take a step back to make some bigger picture comments and offer some perspective
on company cultures and the underlying model of culture’s effects.
Company culture
There has been a recent uptick in culture research in accounting, finance, and economics. An easy
way to illustrate this is the 2016 and 2017 Graham et al. papers that Guggenmos (2020) cites. I
welcome this. As I said at the 2018 Contemporary Accounting Research conference, it helps to
“take culture out of the error term” in the (regression) equation that seeks to explain organizational
performance. Without doing so, culture has been both used and abused as it suits; that is, sometimes
praised as a boon for success and other times as a bane for failure. Worse, cultures that have been
praised can sometimes overnight turn into the alleged rot behind an unexpected failure. Cultures of
laser-focused, performance-driven meritocracy in banks praised during propitious times apparently
seamlessly became bastions of excessive greed when the financial crisis struck. Obviously, bank
culture did not change overnight; only bank performance did, dramatically, for the worse. And
hence, whether culture is “good” or “bad” appears to be inferred from performance—that is, it is
assumed that the culture must be good when a firm or a sector is doing well, with much being read
into what such a presumably good culture is. Enter failure, and the narrative about otherwise the
same culture changes. But culture is not an innocent bystander and infinitely acquiescent to perfor-
mance. Instead, culture deserves to be treated as a proper explanatory variable of organizational per-
formance instead of being retrospectively inferred from it.
One thing is sure: cultures are potent, for better or worse. To illustrate the “for worse” side,
and pertinent to Guggenmos (2020), I gave an example at the 2018 Contemporary Accounting
Research conference from the Toshiba 2015 accounting scandal (Financial Times 2015):

the panel found no evidence any of the three current and former chief executives had given spe-
cific instructions to division chiefs to inflate profit figures. It was a corporate culture—one of
exerting pressure on employees to meet aggressive, short-term profit targets spanning three gen-
erations of chief executives—in which employees were afraid to speak out against bosses when
they pushed for unrealistic earnings targets. [Italics added.]
In other words, bosses do not always have to give explicit instructions to have their
employees engage in earnings management, and the perpetrators do not always have to be looked
for in the accounting and finance functions of the organization, because the earnings management
can arise (or be aided and abetted, implicitly commissioned, or normalized) throughout the orga-
nization through implicit cultural norms. As Guggenmos (2020), in “Creativity and innovation”
in section 2, states:
while the finance and accounting function is responsible for ensuring that the organization’s
transactions are properly recorded and reported, it is often employees outside of financial
reporting that make the decisions regarding when, and even whether, the real economic activi-
ties that underlie the accounting records occur.
To the extent that these choices are subject to managerial discretion, they may have an effect
on reported firm performance. This is how Guggenmos (2020) came to focus on “real” earnings
management (REM). However, if culture is as potent as suggested, then indeed Guggenmos (2020)
is justified in maintaining that, even if finance and accounting subcultures discourage opportunistic
behavior, this may not carry over to the rest of the firm. We do not know whether in the Toshiba
case the conditional part of this argument (about the finance and accounting subcultures) applied.
However, the investigation into the case certainly was quite clear in suggesting that “it was a com-
pany culture—one of exerting pressure on employees to meet aggressive, short-term profit targets”
that had the effect on reported firm performance that Guggenmos (2020) is interested in (Financial
Times 2015). Hence, the Guggenmos study is highly relevant and a welcome addition to the
(accounting) literature for this reason.
Underlying model (or part-model) of culture’s effects

The potency of culture (in this and other contexts) also clearly warrants it to be taken out of the
error term to try to explain a variety of outcomes (including (real) earnings management, as
Guggenmos (2020) does). However, taking culture out of the error term requires that the
researcher is clear about
1. the construct—which culture is it? For example, company culture, the culture of a trade
or functional area (e.g., salespeople, engineers, lawyers) or a subgroup/division in the
organization (e.g., the accounting department);
2. the object—the effect of culture on what? For example, organizational performance,
fraudulent behavior, capacity to innovate;
3. the sign—that is, is culture a “good” or “bad” thing?; and
4. the channel—the type of effect through which culture is expected to have an effect; that
is, as a main, intervening, or moderating factor.
In the case of Guggenmos (2020), the answer to (1) is an “innovative company culture”
experimentally manipulated to distinguish “more” versus “less” of it, which is of course quite a
challenge in an experiment, but Guggenmos’s attempt is well conceived and well executed.
The answer to (2) is REM. I already discussed the very valid and relevant focus on this in
the paragraphs above. I deem REM to be a good choice, where the experimental analogue is quite
a classic one (why not?) and, again, properly operationalized as a decision to spend versus delay
presumably needed maintenance with an effect on the current year earnings per share target

(which would be missed if spent, but met if delayed). Thus, this gets at the usual intertemporal
trade-off often characterized under the rubric of “myopic” decision making.
Of course, one can never be entirely sure whether managers who delay maintenance expendi-
tures act truly “opportunistically” or actually “truly” believe that they are making a “good” deci-
sion not driven by self-interest or some bonus-maximizing, short-term rationale. Yet, the assumed
tenor or starting point of Guggenmos (2020), and typically every other study of this type, is that
the decisions are opportunistic. Whether the decisions are opportunistic, however, depends on
whether the maintenance (in this case) is “needed,” which is not an exact determination, espe-
cially in the real world. Such decisions (e.g., whether, how much, and when to “invest” or
“spend” on maintenance, training, R&D, customer loyalty, etc.) are inevitably gray, not black and
white.
When managers cut R&D, it is easy to take that as opportunistic. But what if the R&D project
is a dud? Would not cutting it then be akin to throwing just more good money after bad? Decisions
of this type should, therefore, be analyzed from a “real options” perspective (e.g., Gong et al. 2011),
which I do not mean to be taken in a technical finance meaning of the term, but rather in an organi-
zational or even behavioral way. For example, some managers may believe that spending just
enough on “upkeep” now is fine, whereas others may believe that it would be better to not spend
anything at all now thinking that a more significant “redevelopment of the estate” later is a better
course of action. Both “mere upkeep” and “redevelopment of the estate” are reasonable consider-
ations when pondering seemingly “simple” maintenance decisions. This is not to say that these con-
siderations are not colored by earnings management rationales, which may be particularly suspected
when the amounts spent happen to also help meet a budget or earnings target. Managers have
admitted to that being an important consideration indeed (see Graham et al. 2005). But how does
one really know? This harks back to the critical notion of “intent” in section 1 of this discussion
above. Intent is impossible to establish with certainty. However, experiments, when well designed,
can come closer than any other method. Altogether, Guggenmos (2020) manages to credibly cap-
ture the “opportunistic” or “earnings management” rationale of the maintenance decision in the
totality of his experimental setup, although I personally spent a lot of time reflecting and seeking
reasonable reassurance on this.
In this regard, I found the “should” versus “want” analysis in Guggenmos (2020, in “Addi-
tional analyses—Should versus want” in section 4) reassuring, as indeed the underlying idea was
to try to “understand managers’ motivations to engage in REM” [italics added]. To the extent that
“motivation” gets at intent, this type of additional analysis in Guggenmos (2020) perhaps will
trigger ideas for him and others to exploit this further to measure alleged earnings management
even more prominently in future work.
The answer to (3) about the sign of the effect is clear in Guggenmos (2020): it is about how
creativity (an innovative company culture) may have a “dark side” or “unintended consequences”
through increased REM, which addresses (4) about the “channel” through which culture is
expected to have an effect. Of course, if the ultimate variable were to be organizational perfor-
mance, then it is not entirely clear that even earnings management (if it can be established to be
earnings management indeed, see above) has a “universally negative” effect on performance
(as acknowledged in footnote 2 in Guggenmos 2020).
In my depiction of the effects of culture in Figure 1, organizational performance is the ulti-
mate dependent variable, where there are benefits and cost of culture (depicted separately as well
as showing their combined net effect). Guggenmos (2020) specifically studies the negative effect
of a culture that emphasizes creativity, but only (one of) its cost(s) as manifested by increased
REM. He then also examines how to mitigate this cost (which I depict in Figure 2); specifically,
how REM can be curtailed through two construal-type interventions. Thus, Guggenmos (2020)
focuses on a more limited model of culture’s effects. However, I deem his choice of this particu-
lar focus to be wise because he picked a relevant angle, both managerially (as I suggested in
“Company culture,” immediately above) as well as for accounting researchers to study.

Actually, if one were to seek to summarize the answers to (3) and (4) in one sentence, then I
would pick this one, straight from the key point leading into, and immediately preceding, Gug-
genmos’s statement of Hypothesis 1: “Because more innovative corporate cultures have been
associated with greater risk acceptance . . . I predict that managers will be more likely to accept
the risk that comes with engaging in REM in these cultures.” This exemplifies, in one short sen-
tence, the cost of creative cultures that Guggenmos (2020) focuses on, and how he expects it to
come about. Clear!
6. Final points and implications

Let me just mention two further points relating to the two key aspects of the study: culture and
REM. I will then also make a final point about the implications of the study. First, culture.
Whereas Guggenmos (2020) focuses on “innovative company cultures,” I found it sometimes dif-
ficult to not read this as “creativity” (and for the author to always stick with the chosen term in a
disciplined manner). If we were to substitute every instance of “innovative company culture” with
“creativity,” would the substantive reading, meaning, findings, and interpretations of the study be
different? Yet, is an innovative company culture reducible to just creativity? Can innovative com-
pany cultures entail more than just creativity? Alternatively, is it conceivable that, in some cases,
innovative company cultures could be quite different from creativity entirely? Maybe innovation
could be about experimenting with new operational aspects of the business, but does that require
creativity? Maybe it means spotting tried-and-tested approaches already in existence and
implementing these quickly (ahead of competitors). Hence, innovation then stems from speed
rather than creativity per se. I just mention this. It does not reduce the value and validity of the
study, but it may have implications for its application areas and generalizability, such as to con-
texts where creativity is less central to innovation, and hence, where the critical creativity-induced
risk-taking channel on which this study pivotally relies may be less pronounced or absent.
Second, REM. Again, the assumption is for the (real) earnings management to be opportunis-
tic. I have discussed above in “Underlying model (or part-model) of culture’s effects” in section 5
how difficult, if not impossible, it is to distinguish “real” REM from otherwise benign and good
faith decisions that involve inevitably difficult trade-offs, especially but not exclusively, with
respect to their timing (and hence, their likely determination as being “myopic” or not). The irony
is that this determination is likely to be especially difficult in innovative settings. Innovation
requires difficult trade-offs about whether, how much, and when to invest by its very nature, even
setting aside any possible earnings management or other incongruent or dysfunctional rationales.
This causes a difficult “real” endogeneity for studies of this kind. Indeed, the settings in which
the unintended behaviors are most likely to be relevant and exploitable—that is, more innovative
settings—are also the settings in which they are most difficult to distinguish from their benign
versions. Moreover, the types who self-select into creative environments perhaps are inherently
less myopically inclined, and if so may this introduce further error in our observation (measure-
ment) of what are deemed to be myopic decisions (yielding false positives)?2
REM indeed may be easier to study and/or less prone to measurement error in a sales or pro-
duction environment in a stable industry, say, than in a creative setting. There is less ambiguity
about the intent behind “shipping bricks and other tricks” at month-end than there is about the
amount that should be spent on training or R&D in fast-changing, uncertain environments. But
studying the former also stands to offer a less substantive incremental contribution compared to
the latter.
Finally, implications. Drawing on construal level theory, Guggenmos (2020) tests two “conse-
quence-focused” interventions that are designed to reduce REM within the company’s existing cul-
ture. I have two points about this. First, I still believe that Guggenmos (2020) makes too much out
2. Sorting effects that may occur in practice do not inflict Guggenmos (2020) due to the random assignment of experi-
mental subjects.

of the “hierarchy” of these two interventions to predict that the “higher-level” one will be “more”
effective than the “lower-level” one (even though that is what he finds, so maybe there is validity to
this, but I am still curious about any possible alternative reasons behind the two types of interven-
tions that may drive this result beyond one just being “more inclusive”—and thus, “higher-level”—
than the other). Be that as it may. But, second, I struggle to see how a firm would implement such
“consequence-focused” interventions in practice. It requires bringing decision makers into a mindset
where they are prompted to weigh the consequences of their decisions before they make them,
either in operational terms (lower-level construal) or in a more strategic way (higher-level con-
strual). But how to do that concretely? And who does the prompting? Given that REM often takes
place at quite high managerial levels, might there be not anyone to do this, or might the one who
should be doing it also be complicit in the REM (or stand to benefit from it, and hence, not wanting
to curb it)? Or, should decision processes have some built-in features that trigger the prompting or
pausing? What are these features and how do they work? Some more guidance about these ques-
tions could have brought the study perhaps even more to “real” life beyond it being very interesting
as it is.
References
Financial Times, 2015. Toshiba chief Hisao Tanaka resigns over $1.2bn accounting scandal. July 21. https://
on.ft.com/2Ux4Jg9, retrieved June 7, 2020.
Gong, J. J., W. A. Van der Stede, and S. M. Young. 2011. Real options in the motion picture industry evi-
dence from film marketing and sequels. Contemporary Accounting Research 28 (5): 1438–66.
Graham, J. R., C. R. Harvey, and S. Rajgopal. 2005. The economic implications of corporate financial
reporting. Journal of Accounting and Economics 40 (1): 3–73.
Guggenmos, R. D. 2020. The effects of creative culture on real earnings management. Contemporary
Accounting Research, this issue.

Guggenmos 2020

Uploaded by

Copyright:

Available Formats

Guggenmos 2020

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Guggenmos 2020

Uploaded by

Copyright:

Available Formats

The Effects of Creative Culture on Real Earnings Management*

RYAN D. GUGGENMOS, Cornell University†

Les répercussions de la culture de la créativité sur la gestion du

CAR Vol. 00 No. 00 (Month 2020)

company cultures, higher-level construal interventions increase desire to engage in behavior

2. Literature review and hypothesis development

CAR Vol. 00 No. 00 (Month 2020)

Creativity and innovation

CAR Vol. 00 No. 00 (Month 2020)

HYPOTHESIS 1. Absent intervention, managers immersed in a more innovative company cul-

Construals and intervention

CAR Vol. 00 No. 00 (Month 2020)

CAR Vol. 00 No. 00 (Month 2020)

CAR Vol. 00 No. 00 (Month 2020)

Figure 1 Experimental design

Introduction Meeting with

Company Culture Manipulation

Less Innovative Culture

CAR Vol. 00 No. 00 (Month 2020)

CAR Vol. 00 No. 00 (Month 2020)

CAR Vol. 00 No. 00 (Month 2020)

CAR Vol. 00 No. 00 (Month 2020)

activating higher-level construal).13 A complete listing of textual differences between conditions

CAR Vol. 00 No. 00 (Month 2020)

More innovative culture Cell 1 Cell 2 Cell 3

construal groups is signiﬁcant in the predicted direction (χ 2 = 2.779, p = 0.047, one-tailed,

CAR Vol. 00 No. 00 (Month 2020)

Panel A: Analysis of variance (3×2)

Source df F-statistic p-value

Company Culture 1 0.30 0.584

Panel B: Hypothesis testing

Hypothesis Test statistic p-value*

Comparison t-statistic p-value Comparison t-statistic p-value Comparison t-statistic p-value

CAR Vol. 00 No. 00 (Month 2020)

CAR Vol. 00 No. 00 (Month 2020)

Panel A : Predicted results

No Intervention Lower−level Construal Higher−level Construal

Panel B: Observed results

More Innovative Culture Less Innovative Culture

No Intervention Lower−level Construal Higher −level Construal

CAR Vol. 00 No. 00 (Month 2020)

Panel A: ANOVA (2×2)—No Intervention conditions excluded

Source df F-statistic p-value

Company Culture 1 0.95 0.332

Panel B: Tests of simple effects F1,90 p-value

CAR Vol. 00 No. 00 (Month 2020)

Unexpected effects of construal level interventions in less innovative company cultures

Additional analyses—Should versus want

CAR Vol. 00 No. 00 (Month 2020)

Panel A: "Should" results

No Intervention Lower−level Construal Higher−level Construal

Panel B: "Want" results

No Intervention Lower−level Construal Higher−level Construal

CAR Vol. 00 No. 00 (Month 2020)

CAR Vol. 00 No. 00 (Month 2020)

Panel A: Mean (SD) response to should measure

No intervention Lower-level construal Higher-level construal Row means

More innovative culture Cell 1A Cell 2A Cell 3A

Panel B: Mean (SD) response to want measure