Behavioral field evidence on psychological and social factors in dishonesty and misconduct

Accepted Manuscript Title: Behavioral Field Evidence on Psychological and Social Factors in Dishonesty and Misconduct Author: Lamar Pierce Parasuram Balasubramanian PII: DOI: Reference:

S2352-250X(15)00139-6 http://dx.doi.org/doi:10.1016/j.copsyc.2015.04.002 COPSYC 117

To appear in: Received date: Accepted date:

31-3-2015 13-4-2015

Please cite this article as: http://dx.doi.org/10.1016/j.copsyc.2015.04.002 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Behavioral Field Evidence on Psychological and Social Factors in Dishonesty and Misconduct Lamar Pierce

Olin Business School Washington University in St. Louis

cr

Parasuram Balasubramanian

ip t

Olin Business School Washington University in St. Louis

us

ABSTRACT: We review recent behavioral field evidence on dishonesty and other unethical behaviors from psychology and related fields. We specifically focus on individual-level studies that use explicitly behavioral data in natural settings, covering research topics

an

relevant to psychology from across disciplines. Our review shows both the paucity and potential of behavioral field evidence on the psychology of dishonesty--although such

M

research can provide actionable and realistic conclusions, it presents a host of practical and identification-related challenges that have limited its use. We explain the major methodological approaches, and discuss the multiple identification challenges for

d

researchers using archival and other non-experimental data.

te

Laboratory research on dishonesty and other unethical and illicit behaviors has proven invaluable in helping to understand the behavioral underpinnings of misconduct. Similarly, survey-based studies

Ac ce p

have provided a wealth of data and insights on self-reported dishonesty as well as its motivations, mechanisms, and prevalence. Yet an emerging stream of research is using behavior data from field experiments, direct observation, and archival sources to address concerns about the generalizability of often low-stakes laboratory studies and potentially biased self-reported data. This review details the current state of this emerging literature in psychology and related fields, and provides guidelines for future research. We focus specifically on studies that use individual-level behavioral data from “natural” settings—those where people engage in their typical work or personal activities. Related reviews on organizational-level misconduct [1] and broader literatures in business and behavioral ethics [2, 3] are also valuable reading. Existing Behavioral Field Research on Dishonesty We first review the existing behavioral field research on topics of interest to psychologists and behavioral scientists.

Page 1 of 22

Social processes: One of the most promising and important topics on dishonesty is how social processes influence behavior, with a growing body of work using behavioral field evidence to explore it. Bucciol et al. [4] used direct observation and interviews to identify how bus passengers travelling with family members were more likely to have a valid ticket, but not those travelling with friends. Similarly, a field experiment on customers keeping excessive change in Israeli restaurants found

ip t

almost no improved honesty from groups, with the higher average honesty of women exerting little pressure on their male dining companions [5]. These results suggest that social pressure may

cr

selectively increase honesty, but that the specific social dynamics are critical. Two recent studies of performance enhancing drugs in baseball [6] and cycling [7] show that social and professional

us

interactions are critical in disseminating both knowledge and acceptance of illicit drug usage. These follow an important early study of social processes in sports cheating, where Duggan and Levitt [8] showed that sumo wrestlers reciprocally throw matches to aid one another in achieving a minimum

an

win count. It is also consistent with recent work using communication data to examine information transmission among networks of dishonest parties [9, 10]. This is consistent with a field experiment by Wenzel [11] that found information on others’ behavior improved tax compliance, as well as

M

results showing employees become more dishonest when joining dishonest firms [12].

d

Fairness, equity, and social comparison: Social comparison and related fairness and equity concerns are also a focus of recent work. Early work by Greenberg [13] was one of the first to address this topic

te

using behavioral field data, showing increased theft following a pay decrease at two out of three factories. A related study [14] also showed higher theft when the employer, not coworkers, was the

Ac ce p

likely victim. A notable later study by Edelman and Larkin [15] found social comparison as a motivation among faculty fraudulently downloading their own papers on SSRN. Related to social comparison is a small set of field studies on socioeconomic class and dishonesty. Although Gino and Pierce [16] found evidence of dishonest helping within socioeconomic class in mechanics, Balafoutas et al. [17] find no differences in fraud by taxi drivers across customer income levels. Related work [18] examines the socioeconomic class of aggressive drivers, although mechanisms linking dishonesty with personal wealth are difficult to separate. Moral reminders and preferences: Multiple large-scale field experiments have focused on testing the efficacy of moral reminders previously established in laboratory studies. Studies of individual taxpayers [19] and newspaper buyers [20] found that the inclusion of a moral reminder increased honesty in disclosures and payments. In contrast, a field experiment by Fellner et al. [21] found that Austrians only improved their honesty in paying TV licensing fees when mailed threats of enforcement, not when sent moral appeals. These build on an earlier important study of bagel

Page 2 of 22

customers by Levitt [22], who found that payments under the honor system were largely a function of internal moral preferences. Furthermore, he found that the September 11 terrorist attack significantly increased honesty in payments, suggesting the power of moral reminders. Related to this, Shu et al. [23] used a field experiment to show that insurance customers who signed at the top of forms reported higher annual mileage than those who signed at the bottom, presumably because signing

ip t

provided a moral reminder.

cr

Culture: Several recent studies have also found the influence of ethnic or national culture and identity on dishonest behavior. A foundational study in economics correlated national corruption measures

us

with the unpaid parking tickets of diplomats [24]. Other papers focused on how interactions within and across ethnic and national groups can change levels of dishonesty, including favoritism in Olympic judging [25], ethnic diversity and corruption in Indonesia [26], and stock market fraud in

an

Kenya [27]. One approach by Bianchi and Mohliver [28] links economic conditions during executives’ formative periods to stock option backdating.

M

Professionalism: One growing area of interest is how the professional identity and pro-social motivation of an expert can clash with her career and financial incentives. Although dishonesty in certain

d

professions might be expected (e.g., auto mechanics) [29], for others the public’s trust in expert honesty is critical. Medicine provides several examples, such as how liver transplant surgeons’

te

financial and prosocial motivations can lead to dishonest patient reporting [30]. Similarly, teachers who are expected to instill ethical values in children have been shown to cheat when pressured with

Ac ce p

strong financial and career incentives [31].

Incentives and control: One of the largest bodies of behavioral field studies centers on extrinsic motivation from incentives and control—how financial payoffs, monitoring, and penalties can alter dishonest behavior. Although the majority of this work is in economics [32], the work on monitoring has particular implications for psychological theories of dishonesty. Monitoring, for example, has been shown to reduce theft [33, 34], unexcused absenteeism [35], and dishonest reporting [36] in organizational settings such as call centers, restaurants, schools, and banks. Similarly, recent field experiments have targeted tax fraud [37, 38] and corruption [39, 40] through the explicit manipulation of increased monitoring through audits and transparency. Although economic theory implies the efficacy of monitoring, evidence from a field experiment on factory productivity monitoring [41] suggests that psychological mechanisms may make monitoring counterproductive in reducing dishonesty. Behavioral field research that can test the multiple psychological and economic mechanisms invoked by monitoring is clearly needed.

Page 3 of 22

Methodological Approaches: Field Experiments and Archival Three principal methodological approaches dominate behavioral field research on dishonesty: direct observation, randomized field experiments, and archival data analysis. Direct observation involves actively observing and recording behavior under multiple conditions to infer relationships between

ip t

dishonesty and environmental conditions or individual differences. This method lacks the

randomized manipulation of a field experiment, and observation is typically covert to avoid

cr

Hawthorne effects. The broad and random sampling of both honest and dishonest behavior across

conditions and differences is important to avoid selection bias. Examples include researchers actively

us

watching and recording problematic behavior such as illegal parking [42], bus fare evasion [4], aggressive driving [18], or bribes paid by truckers in Indonesia [43].

an

Field experiments typically involve direct observation, but also include random assignment of manipulations to treatment and control groups, as in laboratory experiments. Although highly stylized experiments outside of a laboratory are often considered to be “field experiments,” those

M

conducted in natural behavioral settings are most valuable for understanding behavior in the field. Such “natural field experiments” study individuals in plausibly normal daily behavior, rather than in

d

contrived tasks or jobs they would not normally do. Random assignment can occur either as individuals [23] or groups [20]; field settings often make individual randomization impossible for

te

practicality reasons or because of the inseparability of organizational or social settings. The strengths of natural field experiments on dishonesty are threefold: they are immediately generalizable to

Ac ce p

specific social or organizational settings, they provide strong causal inference, and they frequently provide immediate and measurable policy implications for managers or government. The third methodological approach is the analysis of archival behavioral field data. Archival data on dishonesty can be easier to acquire than opportunities for true field experiments or even observational data, particularly in organizational settings where management can be reluctant to allow researchers to manipulate or interact with the work environment. Archival data present several important identification challenges, however, because of the lack of random assignment. Causality is therefore difficult to establish because of selection bias and often cross-sectional data structure. Just as commonly, omitted variables might explain the relationship between the variable of interest and dishonesty in a way inconsistent with the proposed theoretical explanation. Researchers therefore often rely on “natural experiments”—exogenous shocks to some individuals or groups that can create quasi-treatment and control groups, such as exogenous restaurant monitoring technology implementation [33].

Page 4 of 22

Measuring Dishonest Behavior One of the great challenges of studying dishonesty in the field is measurement precision and accuracy. Dishonesty and other forms of misconduct are intentionally committed under a veil of secrecy, due to explicit costs from detection and punishment. This yields two substantial problems in

ip t

empirical work. First, measures of dishonesty will be inherently imprecise, with low observability rates across subjects and conditions. So long as this observability is randomly generated, it serves

cr

only as noise in any empirical model, and does not bias any results. The second and larger problem is when the observability of dishonesty is correlated with some other factor, such as individual

us

competence, and is thus inaccurate and biased. This most commonly occurs because dishonest acts are rarely randomly detected and recorded. Instead, the actual detection of dishonest behavior is usually endogenous to any model of behavior. When we observe data on detected dishonesty, it

an

reflects many observations of false negatives, and these false negatives almost certainly reflect their own psychological and economic processes.

M

Given these measurement challenges, there are several ways in which researchers typically measure dishonesty using behavioral field data. The simplest is to directly observe the behavior, either by the

d

researcher or recorded in an archival data set. Jin and Kato [44] provide an excellent example through a field experiment purchasing baseball cards on eBay—the comparison of reported versus true

te

condition directly measures dishonesty. Pierce et al. [45] provide an example from archival data,

Ac ce p

observing revenue theft by workers in point-of-sales data from a large set of restaurants. The second common method is to identify suspicious patterns of behavior that are inconsistent with the known counterfactual of honest behavior. Although this measurement strategy does not allow researchers to precisely identify individual dishonest acts, it provides probabilistic measures that still allow for hypothesis testing. Jacob and Levitt [31] provide a classic example of this through suspicious answer patterns that identify teacher cheating. Pierce and Snyder’s [12, 45] use of improbably high pass rates from vehicle emissions testing experts provide a good example of fraud in a firm setting. Within academia, Simonsohn [46] and colleagues [47] provide clear examples of how counterfactual probability distributions of hypothesis tests can help identify both intentional and subconscious dishonesty in data reporting. This measurement method can be particularly useful in the context of a field experiment, where a manipulation that should only affect dishonest people (such as the risk of a tax audit) can impact reported income and deductions [19, 33, 38]. Other Identification Challenges in Archival Work

Page 5 of 22

Addressing alternative explanations Although field data may show patterns consistent with the theoretical explanation advanced by researchers, they carry an additional challenge endemic to non-experimental approaches—alternative explanations for observed correlations. In a tightly controlled laboratory setting, one can precisely manipulate the variable of interest, and thereby hold all other factors constant. But in a field setting,

ip t

such precision is rarely possible, and identification can be haunted by multiple plausible or even

probable alternative explanations for the results. Convincing archival and observational studies of

cr

dishonesty must not only provide proof for a hypothesis, but must also cast substantial doubt on

us

alternative hypotheses.

Most significantly, causality is often difficult to establish, and researchers must be careful not to infer a causal relationship from correlational results. Even panel data, where multiple individuals are

an

observed across time, often is unable to reveal causal evidence because of independent variables that are endogenously determined. One promising approaches is to exploit a natural experiment, where the impact of plausibly exogenous shock on individual behavior can be reasonably called causal [22,

M

20, 33]. Another is to exploit discontinuous policy or rules through the equivalent of a quasi-

d

experimental regression discontinuity design [8,30, 48]

In many field settings, alternative hypotheses cannot be fully dispelled, and the scope of identification

te

problems must be weighed against the novelty of the research question and field setting. Yenkey’s [27] study of ethnicity and stock market fraud in Kenya, for example, is exceptionally novel on both

Ac ce p

dimensions. In all cases, researchers should be, and should be encouraged by editors to be, honest about identification problems in their research. Identifying psychological (and economic) mechanisms Perhaps the biggest challenge with behavioral field data is the identification of specific psychological and economic mechanisms driving dishonest behavior. Theory may propose multiple mechanisms that could explain observed behavior, yet without the benefit of a controlled experimental setting, separating these mechanisms can be tricky. Behavioral field data are rarely accompanied by the selfresponse data frequently used by psychologists to measure psychological processes. Consequently, researchers must instead attempt to identify mechanisms either by using variation in the main effect through moderators. The Promise of Mixed Methods

Page 6 of 22

One promising approach to addressing limitations in behavioral field data on dishonesty is to combine them with laboratory studies or surveys, particularly for the purpose of identifying specific psychological or economic mechanisms [15, 16, 18, 20]. Researchers can test competing hypothesized mechanisms in controlled laboratory settings, but must be careful that both the experimental design and subject pool can be reasonably generalized to the field data setting. This is particularly important

ip t

when the field setting involves experts with long-run reputational or incentive concerns, which may

match poorly to undergraduate students playing highly stylized, low-stakes cheating games. Although

cr

scenario studies can help target the specific setting, subjects may not have the expertise and settingspecific knowledge to provide insights on the mechanisms behind observed behavior in the field.

us

Similarly, surveys can help explicate the decision-making process and emotions, particularly when directly conducted with the population from which the behavioral field data were drawn.

an

Conclusion

Behavioral field research presents great opportunities for better understanding the psychological and economic foundations of dishonesty, yet they present their own set of unique challenges. The lack of

M

randomized assignment in archival and observational data presents arguably the great challenge because they make causality difficult to establish and alternative explanations hard to dispel.

d

Consequently, researchers must approach behavioral field data with an appropriate level of skepticism, and with the fundamental question of “what else might explain these results?” Results

te

that are consistent with a theory are not sufficient, because they produce false positive conclusions that are particularly dangerous in settings involving dishonesty. Instead, results should ideally be

Ac ce p

consistent with a theory and inconsistent with alternative theories for observed relationships. The behavioral field studies discussed here are somewhat arbitrarily bounded by necessity, and are part of a much broader set of field studies on dishonesty. For example, survey-based can be extremely valuable to researchers, particularly when embedded in a randomized experimental design [49].

Given the rapidly expanding literature of laboratory studies on cognitive biases in ethical decisionmaking, there is a tremendous need for further behavioral field evidence on bias in dishonesty. A few recent studies address this shortage using data from banks [50] and illegal parking [42], but also demonstrate the difficulty of separating cognitive bias mechanisms from others in field data. Perhaps most promising is recent work by Rees-Jones [51] that uses a quasi-experimental regression discontinuity approach to identify how loss aversion influences tax fraud. Further field studies that

Page 7 of 22

can directly identify cognitive biases would provide a substantial contribution to the field of behavioral ethics. Finally, behavioral field data on dishonesty bring additional risks and concerns regarding human subjects protection. Identifiable and private data on dishonesty have the potential to produce

ip t

substantial harm, such as job loss, social shame, and even incarceration. Furthermore, field

experiments must be typically limited to interventions believed to decrease dishonesty because of the

us

cr

potential cost to subjects and other parties.

References *of special interest

an

**of outstanding interest

M

[1] Greve HR, Palmer D, Pozner J: Organizations gone wild: The causes, processes, and consequences of organizational misconduct. The Academy of Management Annals 2010, 4:53–107.

d

[2] Tenbrunsel AE, Smith‐Crowe K: 13 Ethical decision making: Where we’ve been and where

te

we’re going. The Academy of Management Annals 2008, 2:545–607.

Ac ce p

[3] Moore C, Gino F: Approach, ability, aftermath: A psychological process framework of unethical behavior at work. The Academy of Management Annals 2015, 9:235–289. [4] Bucciol A, Landini F, Piovesan M: Unethical behavior in the field: Demographic characteristics and beliefs of the cheater. Journal of Economic Behavior & Organization 2013, 93:248– 257. *[5] Azar OH, Yosef S, Bar-Eli M: Do customers return excessive change in a restaurant? Journal of Economic Behavior & Organization 2013, 93:219–226. [6] Gould ED, Kaplan TR: Learning unethical practices from a co-worker: The peer effect of Jose Canseco. Labour Economics 2011, 18:338-348. [7] Palmer D, Yenkey C: Drugs, sweat and gears: An organizational analysis of performance enhancing drug use in the 2010 Tour de France. Social Forces (forthcoming). *[8] Duggan M, Levitt SD: Winning isn’t everything: Corruption in sumo wrestling. The American Economic Review 2002, 92:1594-1605. [9] Baker WE, Faulkner RR: Social networks and loss of capital. Social Networks 2004, 26:91–111. [10] Aven B: The paradox of corrupt networks Organization Science (forthcoming).

Page 8 of 22

*[11] Wenzel M: Misperceptions of social norms about tax compliance: From theory to intervention. Journal of Economic Psychology 2005, 26:862-883. [12] Pierce L, Snyder J: Ethical spillovers in firms: Evidence from vehicle emissions testing. Management Science 2008, 54:1891–1903.

ip t

**[13] Greenberg J: Employee theft as a reaction to underpayment inequity: The hidden cost of pay cuts. Journal of Applied Psychology 1990, 75:561-568.

cr

[14] Greenberg J: Who stole the money, and when? Individual and situational determinants of employee theft. Organizational Behavior and Human Decision Processes 2002, 89:985–1003.

us

**[15] Edelman B, Larkin I: Social comparisons and deception across workplace hierarchies: Field and experimental evidence. Organization Science 2015, 26:78-98.

an

**[16] Gino F, Pierce L: Robin Hood under the hood: Wealth-based discrimination in illicit customer help. Organization Science 2010, 21:1176–1194.

M

*[17] Balafoutas L, Beck A, Kerschbamer R, Sutter M: What drives taxi drivers? A field experiment on fraud in a market for credence goods. The Review of Economic Studies 2013, 80:876– 891.

d

[18] Piff PK, Stancato DM, Côté S, Mendoza-Denton R, Keltner D: Higher social class predicts increased unethical behavior. Proceedings of the National Academy of Sciences of the United States of America 2012, 109:4086-4091.

te

[19] Bott K, Cappelen AW, Sørensen EØ, Tungodden B: You’ve got mail: A randomised field experiment on tax evasion. NHH Norwegian School of Economics 2014.

Ac ce p

*[20] Pruckner GJ, Sausgruber R: Honesty on the streets: A field study on newspaper purchasing. Journal of the European Economic Association 2013, 11:661-679. *[21] Fellner G, Sausgruber R, Traxler C. Testing enforcement strategies in the field: Threat, moral appeal and social information. Journal of the European Economic Association 2013, 11:634-660. [22] Levitt SD: White-collar crime writ small: A case study of bagels, donuts and the honor system. The American Economic Review 2006, 96:290-294 **[23] Shu LL, Mazar N, Gino F, Ariely D, Bazerman MH: Signing at the beginning makes ethics salient and decreases dishonest self-reports in comparison to signing at the end. Proceedings of the National Academy of Sciences 2012, 109:15197–15200. **[24] Fisman R, Miguel E: Corruption, norms, and legal enforcement: Evidence from diplomatic parking tickets. Journal of Political Economy 2007, 115:1020–1048. [25] Zitzewitz E: Does transparency reduce favoritism and corruption? Evidence from the reform of figure skating judging. Journal of Sports Economics 2014, 15:3–30. [26] Olken BA: Corruption and the costs of redistribution: Micro evidence from Indonesia. Journal of Public Economics 2006, 90:853–870.

Page 9 of 22

**[27] Yenkey CB: Tribes and (dis)trust. University of Chicago Working Paper. Available at SSRN: http://ssrn.com/abstract=2562539. [28] Bianchi E, Mohliver AC: Do good times breed cheats? Entering the workforce in an economic boom predicts later unethical behavior. London Business School Working Paper.

ip t

[29] Schneider HS: Agency problems and reputation in expert services: Evidence from auto repair. The Journal of Industrial Economics 2012, 60:406–433.

cr

**[30] Snyder J: Gaming the liver transplant market. Journal of Law, Economics, and Organization 2010, 26:546–568.

us

*[31] Jacob BA, Levitt SD: Rotten apples: An investigation of the prevalence and predictors of teacher cheating. The Quarterly Journal of Economics 2003, 118:843-877. [32] Zitzewitz E: Forensic economics. Journal of Economic Literature 2012, 50: 731-769.

an

*[33] Pierce L, Snow DC, McAfee A: Cleaning house: The impact of information technology monitoring on employee theft and productivity. Management Science (forthcoming)

M

*[34] Nagin DS, Rebitzer JB, Sanders S, Taylor LJ: Monitoring, motivation, and management: The determinants of opportunistic behavior in a field experiment. The American Economic Review 2002, 92:850-873.

d

*[35] Duflo E, Hanna R, Ryan S: Incentives work: Getting teachers to come to school. The American Economic Review 2012, 102: 1241-1278.

te

[36] Hertzberg A, Liberti J, Paravisini D: Information and incentives inside the firm: Evidence from loan officer rotation. The Journal of Finance 2010, 65:795–828.

Ac ce p

[37] Slemrod J, Blumenthal M, Christian C: Taxpayer response to an increased probability of audit: Evidence from a controlled experiment in Minnesota. Journal of Public Economics 2001, 79:455–483. [38] Kleven HJ, Knudsen MB, Kreiner CT, Pedersen S, Saez E: Unwilling or unable to cheat? Evidence from a tax audit experiment in Denmark. Econometrica 2011, 79:651-692. *[39] Olken BA: Monitoring corruption: Evidence from a field experiment in Indonesia. Journal of Political Economy 2007, 115: 200-249. [40] Peisakhin L: Transparency and corruption: Evidence from India. Journal of Law and Economics 2012, 55:129–149. [41] Bernstein ES: The transparency paradox a role for privacy in organizational learning and operational control. Administrative Science Quarterly 2012, 57:181–216. [42] Yap AJ, Wazlawek AS, Lucas BJ, Cuddy AJC, Carney DR: The ergonomics of dishonesty: The effect of incidental posture on stealing, cheating, and traffic violations. Psychological Science 2013, 24:2281–2289. [43] Olken BA, Barron P: The simple economics of extortion: Evidence from trucking in Aceh. Journal of Political Economy 2009, 117:417–452.

Page 10 of 22

[44] Jin GZ, Kato A: Price, quality, and reputation: Evidence from an online field experiment. The RAND Journal of Economics 2006, 37:983–1005. *[45] Pierce L, Snyder JA: Unethical demand and employee turnover. Journal of Business Ethics 2015, doi: 10.1007/s10551-013-2018-2.

ip t

**[46] Simonsohn U: Just post it: The lesson from two cases of fabricated data detected by statistics alone. Psychological Science 2013, 24:1875–1888.

cr

[47] Simmons JP, Nelson LD, Simonsohn U: False-positive psychology: Undisclosed flexibility in data collection and analysis allows presenting anything as significant. Psychological Science 2011, 22:1359–1366.

us

[48] Bennett VM, Pierce L, Snyder JA, Toffel MW: Customer-driven misconduct: How competition corrupts business practices. Management Science 2013, 59: 1725-1742.

an

[49] Mayer DM, Nurmohamed S, Treviño LK, Shapiro DL, Schminke M: Encouraging employees to report unethical conduct internally: It takes a village. Organizational Behavior and Human Decision Processes 2013, 121:89–103.

M

[50] Derfler-Rozin R, Moore C, Staats BR: Enhancing ethical behavior through task variety. London Business School Working Paper.

Ac ce p

te

d

**[51] Rees-Jones A: Loss aversion motivates tax sheltering: Evidence from U.S. tax returns. University of Pennsylvania Working Paper.

Page 11 of 22

Annotated Bibliography for Pierce and Subramanian [1] Greve HR, Palmer D, Pozner J: Organizations gone wild: The causes, processes, and consequences of organizational misconduct. The Academy of Management Annals 2010, 4:53–107.

cr

ip t

In this review article, Greve et al. critique the literature on misconduct with an emphasis on organizational causes. They argue for more rigor from researchers in defining and labeling misconduct; since the definition of misconduct is often implicit in the literature and the role of social-control agents has been neglected in identifying misconduct. They also examine the consequences of misconduct for individuals, networks and organizations and propose possible areas for future research.

an

going. The Academy of Management Annals 2008, 2:545–607.

us

[2] Tenbrunsel AE, Smith‐Crowe K: 13 Ethical decision making: Where we’ve been and where we’re

M

Tenbrunsel and Smith-Crowe review and summarize research on ethical decision making in organizations with specific focus on behavioral or descriptive ethics. Using an inductive approach, the authors focus on theoretical frameworks that arise from data. To summarize the field and provide future direction, they develop a model of ethical decision making and a typology of key dependent variables.

d

[3] Moore C, Gino F: Approach, ability, aftermath: A psychological process framework of unethical behavior at work. The Academy of Management Annals 2015, 9:235–289.

Ac ce p

te

Moore and Gino in this review important contributions in moral psychology from over a couple of decades, and suggest ways for this literature to enrich and inform research on workplace unethical behavior. They argue that dominant models of workplace unethical behavior of the past two decades have failed to explain how and why people make the moral decisions they do – from moral psychology and cognitive neuroscience in particular. The authors explain the role of intuition, affect, physiology and identity in supporting and informing a more deliberative reasoning process in the construction and enactment of moral behavior. [4] Bucciol A, Landini F, Piovesan M: Unethical behavior in the field: Demographic characteristics and beliefs of the cheater. Journal of Economic Behavior & Organization 2013, 93:248–257. In this field study, Bucciol et al. randomly interviewed 541 bus passengers in the city of Reggio Emilia in Italy. They exploit a high level of fare evasion among passengers to find young individuals, male and non-European immigrants are more likely to travel without a ticket. They also find a social component with cheating: traveling with relatives increases probability of holding a valid ticket and friends have no bearing on unethical behavior while traveling, except around noon when it was mostly male students. Field interviews were conducted during September to December 2011 in collaboration with a local NGO. [5] Azar OH, Yosef S, Bar-Eli M: Do customers return excessive change in a restaurant? Journal of Economic Behavior & Organization 2013, 93:219–226. Azar et al. conduct a field experiment in a large restaurant in the center of Israel to study dishonest behavior in a natural setting. When customers who paid with cash at the restaurant were returned excessive cash in change, authors found that a majority of these customers did not return the excessive change. They also found that repeat customers returned the excessive change back to the restaurant as compared to one-time customers; and among repeat customers, more women returned excess change as compared to men. The

Page 12 of 22

experiment was conducted between March and September 2011 and a total of 192 observations of excessive change and customer behavior were recorded. [6] Gould ED, Kaplan TR: Learning unethical practices from a co-worker: The peer effect of Jose Canseco. Labour Economics 2011, 18:338-348.

cr

ip t

Gould and Kaplan examine if workers learn productive skills from their co-workers even when those skills are unethical. The authors use data from the “Baseball Archive”, freely available on the internet and restrict the sample to seasons between 1970 and 2009. They find performance of players increased significantly after playing with John Canseco, consistent with his claims that he started a contagion of steroid use in the 1980’s by influencing his peers and trainers. The authors do not find any evidence for 30 other power-hitting stars affecting their teammates in the way Canseco did. Their findings thus suggest that workers not only learn productive skills from their co-workers, but sometimes those skills may have come from unethical practices.

us

[7] Palmer D, Yenkey C: Drugs, sweat and gears: An organizational analysis of performance enhancing drug use in the 2010 Tour de France. Social Forces (forthcoming).

M

an

Palmer and Yenkey analyze data on 198 cyclists and 22 teams that competed in the 2010 Tour de France against the backdrop of increasing, but incomplete social control of doping in cycling. They theorize two under-explored causes of wrongdoing related to informal peer relationships and to an organizational participant’s embeddedness in formal organizational structures. Their results suggest that actors who perform critical organizational roles are more likely to engage in wrongdoing. They also find that social ties to unpunished offenders increased likelihood of wrongdoing while ties to severely punished offenders reduced it.

d

[8] Duggan M, Levitt SD: Winning isn’t everything: Corruption in sumo wrestling. The American Economic Review 2002, 92:1594-1605.

Ac ce p

te

Duggat and Levitt present statistical analysis documenting match rigging in sumo wrestling. They use data from more than a decade of Japan’s sumo elite to find overwhelming evidence that match rigging occurs in the final days of sumo tournaments. The data set contains almost every official match between January 1989 and January 2000, a total of 32,000 bouts of top rank Japanese sumo wrestling. With focus on economic analysis of corruption, they find wrestlers win a disproportionate share of the matches when on the margin, not explained by increased effort. Reciprocity agreements between stables of wrestlers appear to exist suggesting collusive behavior is not restricted to individual actors. [9] Baker WE, Faulkner RR: Social networks and loss of capital. Social Networks 2004, 26:91–111. Baker and Faulkner use a strategic case in an oil and gas company (Fountain Oil & Gas Company) to determine the effect of social networks on the probability of loss of capital after investments have been made. Legal actions against Fountain provided publicly available records at the District Attorney’s Office of Ventura County, California; in addition they obtain data from the receiver of the bankruptcy court and Fountain’s financial statements. They find investors with highest risk of loss of capital with probability 79%, did not use social ties or conduct due diligence; while investors with lowest risk of loss of capital with probability 14%, effectively used within-network exchange and conducted due diligence. [10] Aven B: The paradox of corrupt networks Organization Science (forthcoming). In this paper, Aven looks at Enron Corporation prior to its bankruptcy using longitudinal data of organizational crimes, and combines qualitative coding techniques with social network analysis to study effects of corruption on organization behavior. She identifies three corrupt projects for analysis using data from government documents and employee testimonies, and also analyzes an email dataset of

Page 13 of 22

communications from Enron for examination of communication of organizational crime. An important finding of the paper is that corrupt information is a strong determinant of communication behaviors that coincide with information sharing, but this effect weakens over time. Also it was found that individuals within corrupt projects would communicate less and share fewer reciprocal ties than those involved in non-corrupt projects.

ip t

[11] Wenzel M: Misperceptions of social norms about tax compliance: From theory to intervention. Journal of Economic Psychology 2005, 26:862-883.

us

cr

Wenzel conducts a field study with 1,500 Australian taxpayers in addition to a questionnaire-based experiment with university students. With the support of the Australian Taxation Office (ATO), random taxpayers were sent a survey regarding their own views as well as of other taxpayers, specifically concerning honesty in deduction claims. He argues that taxpayers are likely to estimate others’ acceptance of tax evasion as being greater than their own; and they find that feedback about other self-other discrepancy could improve compliance and correct misperception of social norms.

an

[12] Pierce L, Snyder J: Ethical spillovers in firms: Evidence from vehicle emissions testing. Management Science 2008, 54:1891–1903.

d

M

Pierce and Snyder use a unique data set of over three million vehicle emission tests from the department of motor vehicles of a large northern state to show how organizations influence unethical behavior of their employees. The study contains data of all vehicle inspections conducted in 2003 and 2004 for gasolinepowered vehicles under 8,500 lbs. Their results suggest individual ethical behavior is influenced by the ethics of the employer and when individuals move from facility to another, their leniency seems to shift to that of the employer. In particular, their findings show ethical spillovers from firm to individual are strongest at large facilities, corporate chains and weakest for large-volume inspectors.

te

[13] Greenberg J: Employee theft as a reaction to underpayment inequity: The hidden cost of pay cuts. Journal of Applied Psychology 1990, 75:561-568.

Ac ce p

In this field experiment, Greenberg studies employee theft rates in a manufacturing plant in a period when pay was temporarily reduced by 15%. The study includes nonunion employees working for 30 consecutive weeks in one company spread across three manufacturing plants in mid-western US. He uses two categories of dependent variables to measure theft behavior: actuarial data on employee theft, and self-report measures tapping some of the processes. Greenberg finds employees when workers experienced a 15% pay reduction, reported feeling underpaid and stole twice as much as against when they felt equitably paid. Theft rate reduced and feelings of inequity lessened, when the basis for pay cuts was fully explained to employees. [14] Greenberg J: Who stole the money, and when? Individual and situational determinants of employee theft. Organizational Behavior and Human Decision Processes 2002, 89:985–1003. Greenberg studies the personal and situational variables affecting employee theft in a real work environment. The final sample in the experiment consisted 270 customer service representatives employed by a large financial services company in the US; 142 of the representatives were in an office with an ethics program while the rest were in a distant city with no ethics program in place. He found theft was higher with employees having lower moral development, in the office that had no ethics program and when the company stood to be victimized, instead of coworkers. [15] Edelman B, Larkin I: Social comparisons and deception across workplace hierarchies: Field and experimental evidence. Organization Science 2015, 26:78-98.

Page 14 of 22

ip t

Edelman and Larkin use field data and scenario-based experiments (M-Turk) to study how unfavorable social comparisons varyingly engage employees of different hierarchical levels to engage in deception. For the field study, the authors use data of downloads from SSRN between 2001 and late 2007 to examine deceptive downloads of academic working papers. Their results provide the first evidence that the impact of negative social comparison differs across work hierarchies while confirming the importance of social comparison in the workplace. They find that successful and long-tenured employees face a greater loss of self-esteem from negative social comparisons and are more likely to engage in deceptive behavior in response to reported performance that is lower than that of peers.

cr

[16] Gino F, Pierce L: Robin Hood under the hood: Wealth-based discrimination in illicit customer help. Organization Science 2010, 21:1176–1194.

an

us

Gino and Pierce combine data from the field and a laboratory experiment to propose that envy and empathy lead employees to discriminate in illicitly helping customers based on customer wealth. In the field study, they use a database of over six million emission tests from a metropolitan area (of a large northern state) from 2001 to 2004 and identify relative levels of leniency for individual inspectors and customer wealth based on their ownership of luxury versus standard vehicles. The results show that individuals appear to be influenced by social comparisons when choosing to engage in illicit behavior. They find a significant number of inspectors treated standard vehicles differently than luxury cars, and majority of such inspectors appeared to be illegally helping customers who exhibited less wealth.

M

[17] Balafoutas L, Beck A, Kerschbamer R, Sutter M: What drives taxi drivers? A field experiment on fraud in a market for credence goods. The Review of Economic Studies 2013, 80:876–891.

Ac ce p

te

d

Balafoutas et al. conduct a field experiment on taxi rides in Athens, Greece to measure fraud and examine influence of passengers’ presumed income and information on the extent of fraud. Primarily they find passengers who do not have information about optimal routes are taken on longer detours while lack of information about the local tariff system increases the likelihood of cheating of inflated bills by about fifteen percentage points. They use five experimenters switching between two different income roles and three different information roles, collecting observations during two weeks in July 2010 and one week in March 2012. [18] Piff PK, Stancato DM, Côté S, Mendoza-Denton R, Keltner D: Higher social class predicts increased unethical behavior. Proceedings of the National Academy of Sciences of the United States of America 2012, 109:40864091. Using two field experiments and five laboratory experiments, Piff et al. argue that upper-class individuals behave more unethically than lower-class individuals. In the first field study, the authors investigate a busy four-way intersection with four stop signs to see if upper-class drivers were more likely to cut off other vehicles. In the second field study, they tested if pedestrians at a crosswalk are more likely to be cut off by upper-class drivers. Authors collected data for 274 vehicles in the first study and 152 vehicles in the second, both in the San Francisco Bay Area. [19] Bott K, Cappelen AW, Sørensen EØ, Tungodden B: You’ve got mail: A randomised field experiment on tax evasion. NHH Norwegian School of Economics 2014. Bott et al. conduct a randomized field experiment in Norway with taxpayers who according to the tax administration may have misreported their foreign income. The authors limited the sample to 18,000 taxpayers for the income year 2011 since these taxpayers had a substantial negative deviation between selfreported foreign income and third party reports. They find that including a moral appeal in a letter from the tax administration nearly doubled the average foreign income reported, as compared to a base letter without

Page 15 of 22

such an appeal. The moral appeal helps increasing the amount of reported foreign income while the probability of detection increases the number of taxpayers who report foreign income.

ip t

[20] Pruckner GJ, Sausgruber R: Honesty on the streets: A field study on newspaper purchasing. Journal of the European Economic Association 2013, 11:661-679.

us

cr

Pruckner and Sausgruber conducted a field experiment and two complementary studies to examine honesty in the honor system of newspapers selling on the street in Austria. They find that a moral reminder increases the level of honesty in payments but the same message has no effect on an individual’s honesty. The authors collect data in three different ways – field experiment at 40 booth locations over 6 days; a quasi-field experiment at 250 locations over 7 weeks and lastly using a survey of customers who had taken a newspaper from a booth. [21] Fellner G, Sausgruber R, Traxler C. Testing enforcement strategies in the field: Threat, moral appeal and social information. Journal of the European Economic Association 2013, 11:634-660.

M

an

In a natural field experiment, Fellner et al. evaluate alternative strategies to enforce compliance with the law. They use the setting of the radio license fees that funds public broadcasters in Austria, to manipulate mailings sent out by GIS (Gebuhren Info Service) to potential license fee evaders. The mailing contained information about a citizen’s payment duty, size of the fee and maximum fine imposable in case of detection. The results show that the mailings had a positive impact on fee payments in the treatment group as compared to an untreated control group. The authors suggest that mailing may signal surveillance and thus an increased perceived risk of sanction.

te

d

[22] Levitt SD: White-collar crime writ small: A case study of bagels, donuts and the honor system. The American Economic Review 2006, 96:290-294.

Ac ce p

Levitt in this decade long field study recorded results of nearly 75,000 deliveries of bagels and donuts to corporate offices. In particular, he analyzed the discrepancy between prices of goods consumed and actual payments deposited in a lock box. The data reflects deliveries to offices over a 12-year period from 1993 to 2004 along with a list of prices and a lockbox where customers make payments on the honor system. He finds that payment rates fall in response to increases in the posted price, suggesting some marginal consumers to be willing to sustain the moral cost associated paying less than full price or not stealing goods. He also finds payment rates are higher when bagels and donuts are uneaten, suggesting that the marginal consumer pays a lower share of the posted price than the inframarginal consumer. [23] Shu LL, Mazar N, Gino F, Ariely D, Bazerman MH: Signing at the beginning makes ethics salient and decreases dishonest self-reports in comparison to signing at the end. Proceedings of the National Academy of Sciences 2012, 109:15197–15200. Shu et al. use field and laboratory experiments to find that signing before an opportunity instead of after, makes ethics salient and significantly reduces dishonesty. The authors conduct the field experiment with an insurance company in southeastern United States, with customers required to sign an honesty statement at the beginning of an insurance form instead at the end. They collected data from 13,488 insurance policies for a total of 20,741 cars. They find that asking customers to sign at beginning of the form led to a 10.25% increase in calculated miles in comparison to signing at the end; suggesting that customers were actually checking the odometer when they signed at the beginning. [24] Fisman R, Miguel E: Corruption, norms, and legal enforcement: Evidence from diplomatic parking tickets. Journal of Political Economy 2007, 115:1020–1048.

Page 16 of 22

ip t

Fisman and Miguel analyze parking behavior of United Nations officials in Manhattan to study cultural norms and legal enforcement in controlling corruption. The authors use field data of parking violations covered from November 1997 to November 2005 from New York City Department of Finance, Kaufmann et al. data on country corruption levels and also employ country-level data for economic, political and social characteristics. The sample was restricted to all countries that had 1998 populations greater than 500,000. They find a strong effect of corruption norms that diplomats from high-corruption countries accumulated significantly more unpaid parking violations. In 2002 when enforcement authorities acquired the right to confiscate diplomatic license plates of violators, following which unpaid violations by diplomats dropped sharply in response.

cr

[25] Zitzewitz E: Does transparency reduce favoritism and corruption? Evidence from the reform of figure skating judging. Journal of Sports Economics 2014, 15:3–30.

an

us

In this field study, Zitzewitz examines data from the International Skating Union (ISU) following vote trade scandals in the 1998 and 2002 Olympics. The ISU introduced a number of changes to its judging system in order to disrupt collusion by groups of judges. For the study, the author considers three samples: one from prior to 2002 scandal, one for events in 2002-04 which retained the 6.0 system and having judge anonymity, and the third sample with events from both the code of points system and judge anonymity. As a result of aggregate favoritism and corruption (vote trading), the results suggest that the compatriot-judge effect increased slightly after the reforms.

M

[26] Olken BA: Corruption and the costs of redistribution: Micro evidence from Indonesia. Journal of Public Economics 2006, 90:853–870.

Ac ce p

te

d

Olken examines corruption in a large Indonesian anti-poverty program that distributed subsidized rice to poor households. In particular, he compares survey data on the amount of rice actually received by households with the administrative data on amount of rice distributed. The survey analysis is done using two datasets, one a detailed survey of 100 villages and the other a nationally representative survey of 200 Indonesian districts. The author finds ethnically heterogeneous and sparsely populated areas have higher likelihood of missing rice. He also finds at least 18% of rice in the program disappeared due to corruption, substantial enough to make a welfare enhancing program become welfare reducing on net. [27] Yenkey CB: Tribes and (dis)trust. University of Chicago Working Paper. Available at SSRN: http://ssrn.com/abstract=2562539. In this paper, Yenkey examines the socially diverse and contentious population of investors in Kenya’s emerging stock market. The author uses data from Nairobi Securities Exchange (NSE) back office databases maintained by Central Depository and Settlement Corporation (CDSC) and archival records from Kenya Investor’s Compensation Fund. He analyzes investor-level data to suggest that investors choose a same-tribe stockbroker and avoid rival-tribe brokers, particularly in areas of inter-tribal violence and where one’s tribe is a minority. On the other hand, a corrupt stockholder is more likely to steal from clients of his own tribe instead of rival tribes; and the stockbrokers are twice as likely to steal from their own tribe members at the top of the wealth distribution rather than the bottom. [28] Bianchi E, Mohliver AC: Do good times breed cheats? Entering the workforce in an economic boom predicts later unethical behavior. London Business School Working Paper. Bianchi in this study argues that during an economic boom, entering the workforce will increase the likelihood that people will engage in ethically questionable practices later in their careers. The authors use data from 14,872 executives and directors of publicly traded companies in the US between 1996 and 2005; executives who received at least one stock option grant during this period. They find that people who began

Page 17 of 22

their careers in economically prosperous times are more likely to backdate stock options later in their careers. They also suggest that at a formative period in life, environmental conditions can affect ethical impulses for a long time. [29] Schneider HS: Agency problems and reputation in expert services: Evidence from auto repair. The Journal of Industrial Economics 2012, 60:406–433.

us

cr

ip t

With undercover visits to auto repair garages using a test vehicle, Schneider conducts a field experiment to examine how asymmetric information between motorists and mechanics over auto repair service quality affects outcomes. With the help of mechanics working for a Canadian public interest group, data is derived from a study conducted at forty independent auto repair shops in four towns in two Connecticut counties. The author also uses data from repair receipts in 51 undercover garage visits conducted by the Canadian group. Results are largely consistent with large agency problems that the mechanic is not likely to present efficient service whether a customer represents one-time or repeat business. [30] Snyder J: Gaming the liver transplant market. Journal of Law, Economics, and Organization 2010, 26:546– 568.

M

an

In a field study of the liver transplant market, Snyder uses a policy change to examine changes in ICU (intensive care unit) admission behavior. The policy change in effect after March 1, 2002 mandated that liver transplant waiting lists were not to be influenced by ICUs. The author uses data provided by the United Network for Organ Sharing (UNOS), a comprehensive database on all liver transplants performed in the US from mid-1987 till the end of 2008. The results suggest that post-reform, areas with multiple transplant centers saw a reduction in ICU usage. In competitive areas, he finds that in order for the sickest patients to receive a liver, the waiting list seems to be manipulated.

te

d

[31] Jacob BA, Levitt SD: Rotten apples: An investigation of the prevalence and predictors of teacher cheating. The Quarterly Journal of Economics 2003, 118:843-877.

Ac ce p

Jacob and Levitt use data from Chicago public schools to find serious cases of teacher or administrator cheating on standardized tests each year in at least 4-5 percent of elementary school classrooms. To detect teacher cheating, they develop an algorithm that combines information on expected score fluctuations and suspicious patterns of student answers in a classroom. Data for this study comes from Chicago Public Schools administration which includes detailed question-answers for every student in grades 3 to 8 taking the Iowa Test of Basic Skills (ITBS) from 1993 to 2000. Their results primarily suggest that teacher cheating appears responsive to relatively minor changes in incentives. [32] Zitzewitz E: Forensic economics. Journal of Economic Literature 2012, 50: 731-769. Zitzewitz reviews a new emerging field of forensic economics that uncovers evidence on hidden behavior in a variety of domains. This article focuses on uncovering evidence of wrongdoing across different research questions, drawing out common threads and opening possibilities of further research across fields. The author chooses to deal with economically important hidden behavior which may usually be considered least unethical; examples include road builders skimping on materials, violations of UN sanctions, unnecessary heart surgeries and racial biases in employment decisions. [33] Pierce L, Snow DC, McAfee A: Cleaning house: The impact of information technology monitoring on employee theft and productivity. Management Science (forthcoming). Pierce et al. look at how firm investments in technology-based employee monitoring impacts both misconduct and productivity. The data for the field study obtained from a restaurant system IT vendor includes five restaurant chains (firms) with a total of 392 restaurant locations in 38 American states from

Page 18 of 22

March 2010 to February 2012. The data includes weakly theft data in addition to all transactions at each restaurant for the two year period. They find significant treatment effects in reduced theft and improved productivity driven by changed worker behavior rather than worker turnover. The results in the study suggest that employee misconduct is not solely a function of individual differences in ethics or morality, but can be influenced by managerial policies that can benefit both firm and individual.

ip t

[34] Nagin DS, Rebitzer JB, Sanders S, Taylor LJ: Monitoring, motivation, and management: The determinants of opportunistic behavior in a field experiment. The American Economic Review 2002, 92:850873.

an

us

cr

In a field experiment, Nagin et al. examine the rational cheater model by observing how experimentally induced variation in monitoring employees of a telephone call center influences opportunism. The experiment was run at a telephone marketing firm which was geographically dispersed at 16 sites; the job of the employees at these sites was to call potential donors and request contributions. Cheating is detected by measuring suspicious bad calls (SBC) during a week. The authors find a considerable fraction of employees behave according to predictions of the rational cheater model – employees engage in malfeasance at a higher rate in response to reduced monitoring. Evidence from a survey of employees who responded to reduced monitoring shows that these employees perceived their employer as being uncaring and unfair. [35] Duflo E, Hanna R, Ryan S: Incentives work: Getting teachers to come to school. The American Economic Review 2012, 102: 1241-1278.

te

d

M

Duflo et al. use a randomized field experiment and a structural model to examine if monitoring and financial incentives can increase learning and reduce teacher absenteeism. Out of 113 schools in India, the authors randomly choose 57 to verify a teacher’s daily attendance through photographs with the help of time and date stamps; and salary was a non-linear function of attendance. In the schools with treatment effect, they found teacher absence rate fell from 42 percent to 21 percent compared to non-treatment schools. They suggest that teacher response was almost entirely due to financial incentives while the program improved child learning.

Ac ce p

[36] Hertzberg A, Liberti J, Paravisini D: Information and incentives inside the firm: Evidence from loan officer rotation. The Journal of Finance 2010, 65:795–828. Hertzberg et al. in this field study of a commercial bank argue that reassigning tasks among agents can alleviate moral hazard in communication. The authors study loan officers of a commercial bank to find the effects of officer rotation policy on their reporting behavior. From the internal records of a large multinational US bank operating in Argentina, the authors analyze the reporting behavior of loan officers in the small and medium business division. The sample covers 7-year period from December 1997 to 2004, 1248 firms and 100 loan officers. They find that reports are more accurate when an officer anticipates rotation, contains more bad news about borrower’s repayment prospects and affects decisions of bank lending. [37] Slemrod J, Blumenthal M, Christian C: Taxpayer response to an increased probability of audit: Evidence from a controlled experiment in Minnesota. Journal of Public Economics 2001, 79:455–483. Slemrod et al. conduct a controlled experiment regarding income tax compliance of 1,724 randomly selected Minnesota taxpayers in 1995. This control group of 1,724 taxpayers were informed by a letter that they their file would be closely examined. The Minnesota Department of Revenue provided data of taxpayers for income years 1993 and 1994. They find that low and middle-income taxpayers on average increased their tax payments compared to the previous year, indicating the presence of noncompliance. They also find reason to suspect that high-income taxpayers may react by reporting lesser income compared to the previous year, probably perceiving that final outcomes are based on initially reported incomes.

Page 19 of 22

[38] Kleven HJ, Knudsen MB, Kreiner CT, Pedersen S, Saez E: Unwilling or unable to cheat? Evidence from a tax audit experiment in Denmark. Econometrica 2011, 79:651-692.

cr

ip t

Kleven et al. analyze a tax enforcement field experiment of over 40,000 individual incomes in Denmark. In the experiment, one group (half) of tax filers were randomly selected for an audit in the second year and in the following year, tax filers in two groups were randomly sent threat of audit letters. The authors use data from the Danish tax administration’s (SKAT) Business Object Database for incomes reported during 2006 and 2007, which includes all income items from third-party reports and are pre-populated, filed and audited tax returns for each year and taxpayer. The authors find substantial tax evasion rate for self-reported income as compared to third party reporting. They also find that random prior audits and threat-of-audit letters have significant effect on self-reported income but no effect on third-party reported income.

us

[39] Olken BA: Monitoring corruption: Evidence from a field experiment in Indonesia. Journal of Political Economy 2007, 115: 200-249.

M

an

Olken conducts a randomized field experiment in Indonesia on monitoring corruption in over 600 Indonesian village road projects. Data for this study comes from World Bank funded Kecamatan Development Projects (KDP) in two of Indonesia’s most populous provinces of East Java and Central Java, and collected between September 2003 and August 2004. He finds that even in a highly corrupt environment, increasing government audits can play an important role in reducing corruption. In particular, he finds an eight percentage points reduction in missing expenditures, as measured by discrepancies between official project costs and an independent engineers’ estimate of costs.

d

[40] Peisakhin L: Transparency and corruption: Evidence from India. Journal of Law and Economics 2012, 55:129–149.

Ac ce p

te

Peisakhin examines the relative effectiveness of using the freedom-of-information law, compared to bribery, for middle and lower class individuals applying for public services. The author uses data from field experiments with New Delhi’s middle and low income residents and slum dwellers, and finds that applicants who use the freedom-of-information law attain the same rate of success and those who employ bribery, suggesting that the law erases class differences in procuring public services. [41] Bernstein ES: The transparency paradox a role for privacy in organizational learning and operational control. Administrative Science Quarterly 2012, 57:181–216. Bernstein examines the impact of transparent organizational design on workers’ productivity and performance using a field experiment and data from embedded participant-observers at the second largest mobile phone factory in China. Two of the sixteen production lines were used for experimental condition where curtains were installed to instill privacy, while the author tracked hourly production and quality data using the company’s management information systems. The findings from the experiment argues for the transparency paradox, that zones of privacy under certain conditions increases performance; under observability, employees conceal their activities through codes and other costly means inhibiting performance. [42] Yap AJ, Wazlawek AS, Lucas BJ, Cuddy AJC, Carney DR: The ergonomics of dishonesty: The effect of incidental posture on stealing, cheating, and traffic violations. Psychological Science 2013, 24:2281–2289. Yap et al. in this article study the impact of physical environments on dishonest behavior using a field experiment, two laboratory studies and one observational field study. In the first field experiment, eightyeight members were recruited from South Station in Boston and outside Columbia University to study the

Page 20 of 22

relationship between stretching and impression formation. In the observational field study to test the realworld generalizability of the incidental-posture effect, the authors use information of legally parked vehicles blocked by double-blocked cars. Overall, the four studies together provide evidence that expansive postures inadvertently lead to feeling more powerful and these feelings can lead to dishonest behavior.

ip t

[43] Olken BA, Barron P: The simple economics of extortion: Evidence from trucking in Aceh. Journal of Political Economy 2009, 117:417–452.

us

cr

In this field study, Indonesian truck drivers are accompanied by surveyors on 304 trips during which they observe over 6,000 illegal payments to soldiers, police and weigh station attendants. During the trips, drivers spent an average USD 40 per trip on bribes, extortion and protection payments accounting to 13 percent of total cost of the trip. Two major routes were studied and data was collected between November 2005 and July 2006. The authors use plausible exogenous changes in number of checkpoints to show market structure affects the level of illegal payments. They find that a reduction in number of checkpoints for bribes resulted in an increase in average bribes at the remaining points along a route.

an

[44] Jin GZ, Kato A: Price, quality, and reputation: Evidence from an online field experiment. The RAND Journal of Economics 2006, 37:983–1005.

d

M

Jin and Kato in this online field experiment look at internet auctions to study the link between price, quality, seller claims and seller reputation. The purchase and grading procedures of ungraded baseball cards were designed to mimic actual practices in the market. Also to avoid changes in value of cards due to players’ performance, the authors restricted the purchase of baseball cards from December 2001 to March 2002. They find some buyers in the market are not only misled by incredible claims of quality, but also pay higher prices for lower quality and are often defrauded. They also find online seller reputation is effective for identifying good-faith sellers; high-claim sellers often target less-experienced buyers and reputable sellers provide better quality conditional on completed auctions.

te

[45] Pierce L, Snyder JA: Unethical demand and employee turnover. Journal of Business Ethics 2015, doi: 10.1007/s10551-013-2018-2.

Ac ce p

Pierce and Snyder examine how the unethical behavior of employees are influenced by their organizations. They use data from over three million vehicle emissions tests in the northeastern US to find strong evidence of ethical spillovers from firms to individuals. The results suggest that employee behavior can be influenced by managers using formal norms and incentives, but employees have persistent ethics that limits the magnitude of influence. The authors find that these spillovers are strongest at corporate chains and large facilities, and weakest for large-volume inspectors. [46] Simonsohn U: Just post it: The lesson from two cases of fabricated data detected by statistics alone. Psychological Science 2013, 24:1875–1888. Simonsohn in this article describes two cases of suspected academic fraud exclusively using statistical analysis of reported means and standard deviations. Analyzing raw data of published results confirmed his initial suspicions of data fabrication and suspected fraud. He argues that availability of raw data makes the detection of fraud easier and more diagnostic, thus making the task of fabrication more difficult and intimidating. The author investigates the following two articles – Sanna, Chang, Miceli and Lundberg (2011) and Smeesters and Liu (2011).

Page 21 of 22

[47] Simmons JP, Nelson LD, Simonsohn U: False-positive psychology: Undisclosed flexibility in data collection and analysis allows presenting anything as significant. Psychological Science 2011, 22:1359–1366.

ip t

Simmons et al. argue that empirical work in psychology is very easy to publish “statistically significant” evidence that can contain any hypothesis. They show that flexibility in data collection, analysis and reporting dramatically increases false-positive rates, which is in spite of a low rate of false-positive findings (≤ 0.05). The authors suggest an effective disclosure-based simple and low cost solution to the problem, which involves six concrete requirements for authors and four guidelines for reviewers.

cr

[48] Bennett VM, Pierce L, Snyder JA, Toffel MW: Customer-driven misconduct: How competition corrupts business practices. Management Science 2013, 59: 1725-1742.

an

us

The authors exploit variation in local competition among vehicle emissions testing facilities to demonstrate that the demand for fraudulent leniency from customers pressures expert inspectors to illegally pass gross polluting cars. The paper argues that under intense competition for customers, firms will be driven to cross ethical and legal boundaries when dishonesty is a valued element of quality for customers. The authors use individual-level test and inspector data, combined with discontinuous pass thresholds and geographically isolated small-town markets, make causal arguments about competition and fraud.

M

[49] Mayer DM, Nurmohamed S, Treviño LK, Shapiro DL, Schminke M: Encouraging employees to report unethical conduct internally: It takes a village. Organizational Behavior and Human Decision Processes 2013, 121:89–103.

Ac ce p

te

d

Using two field studies and a laboratory study, Mayer et al. examine how supervisory ethical leadership within an organization is associated with employees reporting unethical conduct. One field study was conducted in a multinational company headquartered in the US, with participants who were newly-hired employees responsible for selling food to new and current customers in their assigned delivery areas. In the second field study, the authors sample 33,756 employees of 16 manufacturing and technology companies with the question if an employee has observed unethical behavior in the past; a total of 6,554 employees responded ‘yes’ to the question. They find that both supervisor and coworker have an influence on employees’ internal whistle-blowing. [50] Derfler-Rozin R, Moore C, Staats BR: Enhancing ethical behavior through task variety. London Business School Working Paper. Derfler-Rozin et al. use data from the field and a laboratory experiment to argue that task design affects rule breaking in the workplace. For the field study, they use data from the home loan applications-processing operations of a Japanese bank collected between June 2007 and December 2009. In order to capture rule breaking, the authors examine whether employees take a lunch break that exceeds the length permitted under bank regulation and test it against task variety. They find that task variety supports rule compliance through deliberative thinking, when conceptualized in terms of increasing number of switches between different types of subtasks. [51] Rees-Jones A: Loss aversion motivates tax sheltering: Evidence from U.S. tax returns. University of Pennsylvania Working Paper. Rees-Jones uses data from the 1979-190 IRS Panel of Individual Tax Returns to show evidence that loss aversion affects how taxpayers treat losses differently from gains in tax sheltering decisions. He shows evidences of bunching and shifting in the distribution of tax returns near zero that cannot be explained by structural elements of the tax code and is consistent with prospect theory. Although legal and illegal sheltering methods cannot be fully disentangled, the identified behavioral patterns almost certainly reflect widespread dishonesty motivated by loss aversion, given past evidence on base rates of tax fraud.

Page 22 of 22