Do Leaders Affect Ethical Conduct?

Giovanna d’Adda, Donja Darai, Nicola Pavanini, Roberto A. Weber†

April 15, 2016*

Abstract: We study whether leaders influence the unethical conduct of followers. To avoid selection issues present in natural environments, we use an experiment in which we create simple laboratory firms and assign leadership roles at random. In our first experiment, firms engage in competition and unethical behavior enhances firm earnings but produces a negative externality for all firms. We vary, by treatment, two instruments through which leaders can influence follower conduct—prominent statements to the group and the allocation of monetary incentives. We find that leaders influence the ethical conduct of followers both through their statements and through the use of incentives. Moreover, leaders who are likely to have acted dishonestly in a preliminary stage of the experiment are more likely to employ mechanisms to encourage dishonesty among followers. As a result, firms randomly assigned one of these unethical leaders are more likely to engage in misreporting. A second experiment finds that the above relationships are present, though weaker, when firms do not engage in direct competition.

Keywords: leadership, ethics, dishonesty, experiment    

                                                                                                               † Giovanna D'Adda, Politecnico di Milano, giovanna.dadda@polimi.it; Donja Darai, donja.darai@gmail.com; Nicola Pavanini, University of Zurich, nicola.pavanini@econ.uzh.ch; Roberto A. Weber, University of Zurich, roberto.weber@econ.uzh.ch. * The authors gratefully acknowledge financial support from the Swiss National Science Foundation (Project 100018-140571). We are also grateful for comments and suggestions from participants at several conferences and seminars. We are also greatly thankful to Nicola Gennaioli, an associate editor and four anonymous referees for helpful comments and suggestions.

1

1. Introduction

Responses to unethical conduct in organizational, political or social settings often focus on

the role of leaders in facilitating or encouraging such behavior. For example, analysis of

WorldCom’s dramatic collapse in 2002 often focuses on the role played by its founder,

chairman and CEO, Bernard Ebbers. Indeed, the SEC committee investigating WorldCom’s

collapse prominently noted that Ebbers “was the source of the culture, as well as much of the

pressure, that gave birth to this fraud” (Beresford, et al., 2003, p. 1). More recently, former

Volkswagen CEO Martin Winterkorn was accused of playing a role in the 2015 “Dieselgate”

emissions manipulation scandal through a “management style . . . that fostered a climate of

fear” (Cremer and Bergin, 2015). Conversely, a great deal of attention is also dedicated to the

idea than a leader who prioritizes ethics can create more ethical conduct within an

organization (Brown and Treviño, 2006).

However, despite the widespread belief that leaders play a critical role in producing

unethical conduct in groups or firms that they lead, there is little direct evidence in economic

or organizational research of such a relationship. This is not surprising, since there are serious

challenges to cleanly identifying the influence of leaders in fomenting unethical conduct. For

starters, unethical acts in the field, either by leaders or by those who follow them, are often

hidden from view, which makes the study of this relationship difficult. Moreover, even if a

researcher is able to document a relationship between leaders’ actions and followers’

unethical conduct, establishing causality is often impossible due to non-random selection of

leaders: when a corrupt firm has an unethical CEO, is the leader the source of the culture, or

is it the culture that led to the appointment of an unethical leader?1

Recognizing these identification problems, we employ a novel approach to study the

relationship between leadership and unethical conduct. In particular, we use laboratory

experiments, which allow us to exploit the high degree of control afforded by such

environments to avoid many of the problems present in more natural settings. In our

experiments, we employ a task that captures key features of much real-world unethical

conduct. In this task, individual “workers” in simple laboratory “firms” decide whether to lie

and hurt others outside the firm for personal material gain, with the knowledge that individual

lies are unlikely to be detected. We study the impact of leadership by giving some of these

                                                                                                               1  As an example, many people attribute a large part of the corruption in FIFA, football’s international governing body, to the actions of its former leader, Sepp Blatter, who, for example, allegedly directly oversaw the payment of large bribes. However, others argue that Sepp Blatter’s corrupt behavior is a symptom, rather than a cause, of FIFA’s corruption and that the ultimate cause lies with the organization’s structure (Bialik, 2015).  

2

laboratory firms “leaders” endowed with some of the influence channels typical of

organizational leaders—i.e., the ability to make public statements and control over the

distribution of financial rewards. An important feature of our design is that it randomly

assigns leaders to firms so that some firms have leaders more prone to personally engage in

unethical conduct. We then observe how workers in these firms behave, both in comparison

with firms randomly assigned more ethical leaders and with firms free of any influence by

leaders. In this manner, we are able to determine whether the type of leader—and the things

that leader does—affect the ethical conduct of the firms they lead.

Our study represents a novel contribution to the literature on leadership and the

ethical conduct of groups. Milgram (1963) provided shocking evidence that people’s

willingness to harm others can be influenced by statements from an authority figure,

supporting the idea that unethical leaders can facilitate unethical conduct by followers.

However, this precise relationship is often supported in the organizational literature by

evidence that is not entirely compelling, due to the identification problems outlined above.

For example, some of the strongest evidence comes from studies that use survey-based

instruments to measure followers’ perceptions of the degree to which a leader possesses

“ethical” characteristics (e.g., the Ethical Leadership scale; Brown, et al., 2005), and then

correlate this measure with other subjective measures of whether individuals within an

organization act ethically (Mayer, et al., 2012). While these studies often find positive

correlations between the two kinds of perceptions, the results must be interpreted cautiously

due to many possible interpretations, such as correlated bias across subjective measures.

Given the hidden nature of unethical conduct, other studies attempt to identify a

relationship with leadership by measuring observable behaviors that are potentially correlated

with unethical follower conduct, such as employee exit or excess costs (Detert, et al., 2007;

Burks & Krupka, 2012). Beekman, et al. (2013) use a proxy for corrupt leader behavior—the

misplacement of community resources under the leader’s control—and show that this

correlates with lower cooperation among community members, though not necessarily with

more unethical conduct. As with the studies discussed above, these results are more

suggestive of a relationship, since they ultimately only observe behavior that is imperfectly

correlated with unethical conduct. Moreover, even if such a relationship exists, it is difficult

to identify whether leaders are causal.2

                                                                                                               2 Other studies, not directly related to ethical conduct, show a relationship between leaders’ behaviors and the prevalence of cooperative behavior within a group (Kosfeld and Rustagi, 2015; Jack and Recalde, 2015). There  

3

To address concerns about causal identification, we follow other recent research that

relies on laboratory experiments to study leadership (Potters, et al., 2007; Güth, et al., 2007;

Brandts and Cooper, 2007; Hamman, et al., 2011; Brandts, et al., 2014). Like us, these studies

exogenously vary the abilities and characteristics of leaders to cleanly identify the effects

leaders have on a particular behavior of interest.

To address the concern that unethical conduct is hard to observe, we also follow other

experimental work that studies unethical conduct (Gneezy, 2005; Mazar, et al., 2008; Gino, et

al., 2009; Fischbacher and Föllmi-Heusi, 2013; Gibson, et al., 2013; Conrads et al., 2013,

2014). Specifically, we use a behavioral task in which workers have the opportunity to act

unethically—i.e., to tell a lie for profit in a manner that cannot be detected—but where our

ability to create numerous identical replications of the same situation makes inference

regarding the presence of unethical conduct possible. In the task, workers privately roll a die,

which provides their actual performance in a period, but have the opportunity to misreport the

actual outcome, with no possibility of lies being discovered. The extent to which subjects

inflate their reported die roll is the key indicator of unethical conduct in our study.

In our first experiment, firms compete with each other and the harm created by

misreporting is a negative externality for all firms. When a worker acts dishonestly, she

makes her firm better off but harms all other firms to a greater extent. Hence, aggregate

profits are highest, in expectation, when everyone in the industry acts honestly, but any

individual firm’s profits always increase with misreporting. This captures many situations in

which unethical conduct provides a firm an advantage over competitors, but where

widespread unethical conduct is damaging to the entire industry. For example, unethical

lending practices by an individual financial institution may allow it to generate more loans

and short-term profits, but may yield greater long-term regulatory burdens and higher risk of

financial instability. Moreover, domains involving high levels of competition, in which a win

for one party produces losses for another, are where one often encounters instances of

unethical conduct.3

At the level of a worker, this design also incorporates many key features of real-world

unethical conduct. Inflating the reported die roll is personally rewarding, but wrong according

                                                                                                                                                                                                                                                                                                                                                       are rare cases in which random variation in leadership exists in the field. These studies provide valuable insights into the influence and importance of leaders for economic outcomes—e.g., leaders’ gender and public good provision (Chattopadhyay and Duflo, 2004) and national leaders and economic growth (Jones and Olken, 2005). 3 Examples include competition over market share (Volkswagen’s 2015 emissions scandal), in financial markets (Toshiba’s 2008-2015 earnings manipulation) and in sports (organized doping in cycling and athletics).  

4

to most normative moral principles—it violates deontological rules (e.g., Kant, 1785) against

dishonesty and, by producing more social harm than welfare, it also violates utilitarian ethics

(Mill, 1863).4 Moreover, as is often the case outside the laboratory, unethical conduct, at the

individual level, is hidden from view, meaning there is no probability of sanctioning or

detection of individual unethical conduct. Hence, “good” behavior is likely the result of

ethical considerations, rather than fear of detection or punishment (Nagin, et al., 2002).

The main part of our experimental design consists of two stages. In the first,

participants perform the die-roll task once individually, with the possibility to benefit

personally from misreporting, which also harms others in a subjects’ experimental session.

This provides us a (noisy) measure of each individual’s tendency to misreport. In the second

stage, which is our primary focus, participants are randomly matched into four-person firms,

and one person within each firm is randomly selected to be the leader, while the remaining

three subjects are workers. The second stage of our experiment introduces, and varies by

treatment, the principal influence channels ascribed to leaders in the leadership literature.5

The first influence channel available to leaders is the ability to make public statements

to the workers in their firm. Organizational leadership is often associated with someone who,

from a position of prominence, can articulate a broad direction for the firm and motivate

employees in pursuit of that end (Kotter, 1990). In his seminal book on corporate executives,

Barnard (1938) notes that one of their fundamental functions “is to formulate and define the

purposes, objectives, ends, of the organization” (p. 231). Modern studies of and training in

corporate leadership similarly note the importance of a leader’s ability to motivate and

convince followers to pursue a particular direction, often through prominent and visible

speeches (Antonakis, et al., 2012; Lazear, 2012). We implement a simple form of this

                                                                                                               4 In contrast with other possible designs—as when a lie benefits the decision maker without affecting other subjects (Fischacher and Heusi-Föllmi, 2013; Gibson, et al., 2013)—self-interested dishonesty in our first experiment clearly produces negative net social impact. We chose this design specifically to make dishonesty inconsistent with utilitarian ethics. This design creates the possibility that highly utilitarian subjects may lie by reporting lower die rolls—though this only benefits others in the case where there is widespread dishonesty. While we cannot rule out this possibility, a priori, the data suggests that it is not a common motive.  5  Central to organizational leadership research is a distinction between “transactional” and “transformational” leaders (Burns, 1978), with the former effecting change through sanctions and rewards and the latter doing so through persuasion and influencing follower preferences (Bass, 1990). Of course, individuals who are not formal leaders—such as influential peers—might also exert a form of informal leadership influence if they make statements that are attended to by co-workers. Such influence is consistent with leadership in our experiment. Moreover, leaders also do many other things than what we allow them to do. For example, they set examples for followers through their actions (Hermalin, 1998) and select or exclude organizational members (Güth, et al., 2007). For simplicity, we focus on two specific leadership functions. The omission of other channels likely underestimates the degree to which leaders influence followers in more natural settings. Our experimental design can be naturally extended to include other such possible channels.

5

leadership function in our experiment, by providing leaders with a platform to send

statements to workers in their firm between periods of the task.

Second, corporate leaders often possess discretion over financial incentives. A central

theme in organizational economics is the use of monetary incentives as a means for

motivating employees to act in a manner desired by their employers. Such incentives take

many forms, from fixed wages, to precisely defined explicit performance contracts, to

subjectively determined performance bonuses at the discretion of supervisors (Baker, et al.,

1994; Prendergast, 1999). In principle, a leader could use any form of variable pay to

incentivize ethical or unethical behavior. In our experiments, we provide leaders with the

ability to distribute part of employees’ compensation in the form of a discretionary bonus.

To see how a leader might affect the behavior of followers, consider a worker with the

following expected utility function, facing a choice of how much to misreport an outcome:

𝐸 𝑢 = 𝐸 𝜋 𝑟, 𝐿! − 𝑐(𝑟, 𝐿!) (1)

Let 𝑟 = 0 correspond to honesty, while 𝑟 > 0 captures the degree to which the worker

engages in favorable misrepresentation of a performance measure, such as company earnings.

The first term on the right-hand side represents the expected monetary payoff to a worker

from her action, while the second represents an internal psychological cost, such as guilt,

from the chosen action. If 𝐸 𝜋 𝑟, 𝐿 and 𝑐(𝑟, 𝐿!) are both increasing in the degree of

misreporting, 𝑟, then a worker is financially better off by acting unethically, but also

experiences greater psychological disutility.

The terms, 𝐿! and 𝐿!, represent, respectively, the incentive and statement channels

through which a leader can influence workers. As an example, take the case of auditors at

Arthur Andersen, who were alleged to have frequently facilitated accounting violations at

audited firms, including Enron and WorldCom. Corporate leaders incentivized auditors to

keep clients satisfied (𝐿!), as a means for generating revenue from selling non-auditing

services to the same clients (Toffler and Reingold 2004). Hence, auditors could personally

benefit from performing lax audits. At the same time, leaders’ statements justifying more lax

auditing of large clients (𝐿!)—for instance, the head of U.S. operations urged auditors to

“empathize” with clients (Brown and Dugan 2002)—potentially led auditors to focus on the

perspective of the audited firm, thereby lowering the psychological costs of overlooking

auditing irregularities. Our experiments represent a very simplified form of these kinds of

channels through which leaders may influence the conduct of followers.

6

Previewing the results of our main experiment, we find that leaders influence the

ethical conduct of followers, both through their statements and through their use of

incentives. Workers report higher numbers when their leader makes statements encouraging

dishonest reporting and when the leader’s past behavior indicates that workers who report

high numbers will be rewarded. We identify these effects both from reduced-form estimates

of how leaders’ actions influence followers’ behaviors and from structural estimates of a

simple utility model, based on the one above, in which workers’ utilities depend on monetary

payoffs and on a psychological cost to acting unethically, both of which can be influenced by

the leader through different channels.

We also provide evidence that leaders who are, themselves, more dishonest in the first

stage of the experiment are more likely to use incentives and statements to encourage

dishonesty. As a consequence, being randomly assigned a less ethical leader ends up making

a firm’s workers act less ethically. Thus, in combination, we both show that unethical leaders

beget more unethical firms and we identify a causal channel, from a leader’s type, to the

leaders’ actions, to the influence that those actions have on followers.

Finally, we conduct a second experiment to test the importance of the specific

competitive environment in which our main experiment places laboratory firms, as a factor

driving the influence of leaders. We change the nature of the externality, to one in which the

harm for misreporting affects a passive third party, rather than all firms. While this

experiment yields similar qualitative results to those in the first experiment, they are smaller

in magnitude and statistically weaker. This suggests that the effects of an unethical leader on

unethical conduct may be stronger in environments with greater inter-firm competition.

The rest of the paper is structured as follows. In Section 2, we describe the design for

our main experiment and we present the results for this study in Section 3. Section 4 presents

our second experiment. Finally, we discuss our findings and conclude in Section 5.

2. Experiment 1: Design

Each experimental session consists of 20 participants interacting through computers.

The experiment comprises three stages, with the first two stages constituting the main part

(see Figure A.1 in the Appendix for an overview). In the first stage, subjects engage in an

individual reporting task, in which they can act unethically by inflating their die-roll score

and thereby obtain higher individual earnings at the expense of social welfare. The second

stage modifies the task into one in which subjects repeatedly compete in “firms,” with

varying degrees and forms of leadership, for 10 periods. Our primary focus is on how

7

leadership influences misreporting in the second stage. In the third stage, we elicit measures

of perceived social norms of conduct and also several individual-level characteristics.

Subjects receive the instructions for each stage separately, at the beginning of the stage, but

are informed about the overall structure of the experiment from the outset.6

2.1 First stage

In Stage 1, all 20 participants, 𝑖 ∈ 1, . . . , 20 , compete for a prize of varying size, 𝑉.

We introduce the possibility of cheating in the competition—which is undetectable and

profitable for an individual, but socially inefficient—as a way of studying unethical behavior.

At the beginning of the stage, each subject privately rolls a fair six-sided die, the

outcome of which, 𝑝! ∈ 1, 2, . . . , 6 , can be thought of as a subject’s actual realized type or

performance in that period. Subjects are instructed that it is this value, 𝑝!, which they are

supposed to report as their performance for the purposes of the competition. However, since

only a subject observes the die roll, the performance report, 𝑟!, can be any integer from 1 to 6.

Each subject receives a share of a prize, 𝑠!, and this share increases in that subject’s

own reported performance, 𝑟!, relative to the performance reported by other subjects,  𝑟!!. The

size of the total prize available,  𝑉 𝑟! , 𝑟!! , is a function of the average performance reported

by all subjects. Specifically, the total prize obtains its maximum possible value as long as the

average of all the 𝑟! is equal to or below the expected mean of 20 fair die rolls, or 3.5.

However, if the mean reported performance exceeds 3.5 then the size of the prize decreases

linearly with the average reported performance. Thus, misreporting negatively impacts the

size of the prize, and the prize is lowest when all subjects report the maximum possible

performance of 6. As we note earlier, this design feature captures the property that

widespread unethical conduct can harm a society or industry.

More precisely, a subject’s profit, 𝜋! , is the subject’s share of the prize, determined by

the ratio of own performance to total performance, multiplied by the total size of the prize:

𝜋! = 𝑠! 𝑟! , 𝑟!!  𝑉 𝑟! , 𝑟!! = !!!!! !!!"

!!!,!!!  𝑉 𝑟! , 𝑟!! (2)

with 𝑉 𝑟! , 𝑟!! =𝑎

𝑎 − 𝑏  !!! !!

!"!!!,!!!

!"− 𝜇

𝑖𝑓   !!!"!!!!"

≤ 𝜇

𝑖𝑓   !!!"!!!!"

> 𝜇 , (3)

                                                                                                               6 We provide experimental instructions in Appendix E.

8

where  𝜇 = 3.5, 𝑎 = 1250 and 𝑏 = 300. The parameters are chosen such that, under the

assumption of self-interest and no psychological cost to lying, the unique Nash equilibrium is

for all players to report the highest possible outcome of the die roll, 𝑟! = 6 (see Appendix D).

This yields a total prize of 𝑉 = 500, which is considerably lower than the maximal possible

total prize of 𝑉 = 1250.

After rolling the die privately, each subject enters his or her performance report on the

computer. While we cannot detect lying at the individual level, the distribution of reported

performance scores allows us to detect, statistically, the degree of misreporting (cf. Houser, et

al., 2012; Fischbacher and Föllmi-Heusi, 2013; Gino, et al., 2013). The task is performed

only once in Stage 1 and afterwards all subjects are informed about the average reported

performance across all subjects, the total size of the prize, and their own payoff for Stage 1.

2.2 Second stage

In Stage 2 we use the same task, repeated for 10 periods, but this time in the context

of competing “firms.” The 20 subjects in a session are randomly matched into five four-

person firms—consisting each of three workers and one leader (referred to as the

“supervisor”). Workers individually and privately each roll a die and report performance, as

in Stage 1. The function of the leaders varies by condition.

Similarly to Stage 1, firms compete for shares of a prize, 𝑉. Each firm,

𝑓 ∈ 1, 2,… , 5 , obtains a share, 𝑠!, based on the average reported performance by the three

workers in that firm, 𝑟! = 𝑟!,!!!!! 3, relative to the average reported performance in other

firms. As in Stage 1, 𝑉 is highest when the average of all firms’ reported performance levels

is no greater than 3.5, but decreases for higher average reported performance across the

industry. The profit obtained by each firm is then:

𝜋! = 𝑠! 𝑟! , 𝑟!!  𝑉 𝑟! , 𝑟!! = !!!!!

!!!  𝑉 𝑟! , 𝑟!! (4)

with 𝑉 𝑟! , 𝑟!! =𝑎

𝑎 − 𝑏  !!!

!!!

!− 𝜇

𝑖𝑓  !!!

!!!

!≤ 𝜇

𝑖𝑓  !!!

!!!

!> 𝜇

, (5)

where 𝜇 = 3.5, 𝑎 = 1250, and 𝑏 = 300.

Each subject in firm 𝑓 receives an individual share 𝑥!,! of the firm’s profit 𝜋!. The

leader always receives a share of one-fourth of the total firm profit, 𝑥!,! = 0.25. The shares

received by workers can vary across experimental conditions. In some of our conditions, each

9

of the four firm members receives an equal share of the firm’s profits, e.g., 𝑥!,! = 0.25  with

𝑖 ∈ 𝐿, 1, 2, 3 . If this is the case, each individual worker benefits from reporting the highest

possible performance level, or 6, regardless of what other subjects report (see Appendix D).

In other experimental conditions, described below, the firm supervisor can determine the

share of the firm’s profits received by each worker. In these cases, the supervisor’s reward

strategy can either strengthen or weaken the incentives to report performance of 6.

Roles of workers and supervisors are allocated randomly within each firm, and remain

fixed during all 10 periods of Stage 2. After each period, all subjects are informed about their

own profit, their firm’s profit, the average reported performance level of each of the five

firms, the overall average reported performance level, and the individual reported

performance levels and profits of all members of their own firm.

2.3 Leader conditions

We study four treatment conditions, in a 2 x 2 design (see Table 1), which vary the

instruments available to the leader in the second stage of the experiment. In particular, we

vary the ability of leaders to determine incentives, through performance bonuses, or to

articulate a direction for the firm, through prominent statements to workers.

Table 1: Overview of treatment conditions in Experiment 1

Treatment No incentive power Incentive power No public statements Inactive Leader Leader Incentives Only Public statements Leader Statements Only Leader

Note: For each condition, we conducted four sessions with 20 subjects each.

Public statements. In the two conditions with public statements, leaders have 90

seconds at the beginning of the period—before workers roll their dice and report their

performance—to send a typed message to the firm’s three workers. Leaders send statements

to their workers, but workers cannot reply to the leader or send messages to one another. In

the two conditions without public statements, leaders cannot send messages to workers.

Incentive power. In the two conditions in which leaders have no incentive power, all

four subjects in each firm receive the same share, 25 percent, of the firm profit in a period.

Thus, worker payoffs are independent of any actions of the leader. In the conditions with

leader incentive power, leaders distribute financial rewards among the workers. Specifically,

leaders observe the reported performance level of each of the three workers, and the resulting

10

firm profit, and then decide how to allocate 45% of the group profit among the three

workers.7 The leader has to allocate the entire 45% “bonus pool” among the three workers

and cannot keep any part of it. More precisely, of the profits received by the firm,  𝜋!, the

leader receives a fixed share of one-fourth, 𝑥!,! = 0.25. The three workers each receive a

guaranteed portion, 10 percent, plus a share of the remaining 45 percent of the firm profits,

allocated by the leader. A worker’s share of the firm’s payoff is thus, 𝑥!,! = 0.10+ 𝑦!,!,

where 𝑦!,! ∈ 0, 0.45 is the discretionary bonus allocated to that worker and 𝑦!,!!!!! =

0.45. A profit-maximizing leader will always prefer workers to report higher numbers.8 At

the end of the period, workers in all conditions are informed about the reported performance,

allocated rewards, and total earnings of all workers in their firm. Importantly, this means,

they can infer how the leader rewards reported performance.

The Leader condition provides leaders with both channels through which they can

influence the conduct of firm workers, i.e. the leader can send statements before every period

and allocate rewards at the end of each period. The Inactive Leader condition serves as a

control—there is still a person in the role of “supervisor” and this person receives a 25-

percent share of the firm’s profits, but this person cannot do any of the things that leaders do.

The two other conditions, Leader Statements Only and Leader Incentives Only, vary only in

the presence or absence of these two instruments. By eliminating one instrument and keeping

the other, we can study the relative importance and impact of these distinct potential

influence channels. Moreover, within each condition with an active leader (i.e., excluding the

Inactive Leader condition), we can test how different characteristics of and strategies

employed by leaders affect the degree of misreporting.

2.4 Third stage

Following Stage 2, we elicit subjects’ perceptions of social norms regarding the

appropriateness of inflating performance in the first and second stages of the experiment. We

                                                                                                               7 We choose this incentive mechanism for a few reasons. First, subjective allocation of rewards is commonplace in firms (e.g., discretionary monetary bonuses or fixed pay increases in contexts including financial firms and academic departments, the allocation of non-financial rewards such as desirable office space). Second, since we are interested in unethical conduct, which is often unenforceable in pay contracts, a discretionary bonus seems appropriate. Third, we wanted to provide leaders with flexibility in their ability to allocate rewards; this is limited if we provide one specific kind of incentive contract (e.g., a piece-rate or target-based scheme) and complicates the experiment if we introduce too many schemes. Note also, in combination with the ability to make statements, our firm leaders can, in principle, specify a large variety of ex ante performance contracts. 8 Consider the following illustrative examples: assume that the average reported performance by other firms is 3, then if firm workers all report 6 a leader receives a 63% higher payoff than if they report 3. Even if all other firms report 6, the leader still receives a 32% higher payoff if his workers report 6 than if they report 3.

11

follow the method of Krupka and Weber (2013), in which subjects are provided with a

description of a possibly unethical act, and then rate the “social appropriateness” of this

behavior, with an incentive to match the ratings provided by others.9 Finally, we collect

several psychological measures of personality traits (Protected Values toward acting

dishonestly, following Gibson, et al. (2013); Big Five 15-item version; Machiavellianism

MACH IV) as well as socio-demographic measures.

2.5 Procedural details

In total, 320 subjects participated in the experiment—80 subjects (20 firms) in each

condition. Appendix Table B.1 provides a summary of subject demographic characteristics.

The experiment was computerized using the software z-tree (Fischbacher, 2007) and subject

recruitment used ORSEE (Greiner, 2015). All sessions were conducted in January and

February 2013 at the Birmingham Experimental Economics Laboratory of the University of

Birmingham. Subjects’ payoffs were denominated in Experimental Currency Units (ECU),

which were converted into GBP at the rate of 40 ECU = 1 GBP. Each session lasted about

two hours, and subjects earned, on average, £ 19.94, including a show-up fee of £ 2.50.

3. Results

In presenting our results, we first discuss subjects’ behavior in Stage 1, to get a sense of

the prevalence of dishonesty. Then, we turn to our main research question—whether leaders’

characteristics and actions influence workers’ performance reports in Stage 2.

3.1 Individual behavior in Stage 1

Table 2 shows the average reported performance in Stage 1, by condition. Recall, that

Stage 1 is identical across all conditions; instructions for Stage 2 are only distributed after

Stage 1. Therefore, unsurprisingly, individual behavior in the first stage does not differ

significantly between the four conditions (two-sided Kruskal-Wallis test p=0.85).10 In all

conditions, we find evidence of misreporting. Overall, the mean reported score is                                                                                                                9 For instance, we ask subjects to rate how socially appropriate it is for a subject who rolled a 1 in the first stage of the experiment to report a higher number. We incentivize their answer by giving them an extra £ 0.5, if their answer matches the answer of a randomly chosen other subject in their session. 10 Mean reported performance differs little for subjects subsequently assigned to be leaders (4.43) and workers (4.52). For comparison, the mean reported die-roll outcomes of the “externality treatment” in Fischbacher and Föllmi-Heusi (2013) is 4.18, which is lower than what we find (t-test, p=0.10). We also investigate whether individual characteristics, measured at the end of the experiment, predict Stage 1 responses. Regression analysis shows that reported performance in the first stage is significantly higher if the subject is male, older, an economics student, or scores lower on the Big Five conscientiousness dimension (Table B.3 in the Appendix).

12

approximately 4.5, which is higher than the empirically expected mean score, of 3.5. The

distribution of reported performance levels is also highly right-skewed: 26% of observations

are 3 or lower, while 77% are 4 and above, with the median at 5 and mode at 6. Table B.2 in

the Appendix shows that, in every condition, the frequency of scores of 1 (6) is significantly

lower (higher) than the expected frequency of 1/6. Note, however, that many reported scores

are still well below the individual payoff-maximizing report of 6.

Table 2: Average reported performance in Stage 1

Inactive Leader Leader Leader

Incentives Only Leader

Statements Only Total

4.48 (0.17)

4.49 (0.17)

4.59 (0.17)

4.44 (0.15)

4.50 (0.08)

N=80 N=80 N=80 N=80 N=320

Note: Standard errors in parentheses

3.2 Do unethical leaders produce unethical groups?

We next consider behavior in Stage 2, when subjects performed the task collectively

in firms, sometimes with active leaders. We begin by analyzing whether unethical leaders

influence misreporting among their workers. For this, we construct a measure of a leader’s

propensity to act honestly. Specifically, we use a binary variable, dishonest leader, which

indicates whether a leader’s reported Stage 1 performance equals 6. Since the frequency of

scores of 6 in Stage 1, equal to 37.5%, is more than twice as high as the expected frequency

of 1/6 (see Appendix Table B.2), we expect this variable to be correlated with misreporting.

That is, conditional on observing a reported performance of 6 in Stage 1, it is more likely that

an individual misreported than otherwise. According to this classification, 19 of the 60

leaders in the three conditions with active leaders are coded as dishonest.11

Figure 1 shows average reported group performance over time, based on the type of

leader assigned to a firm. For example, the line marked by circles indicates average reports in

firms with inactive leaders. The distance of these mean scores from 3.5 provides us with an

estimate of the degree of misreporting. We divide the observations from the active leader

                                                                                                               11 The results are similar if we, instead, define the indicator variable as those leaders who reported performance of 5 or 6 in Stage 1 or who reported a Stage 1 performance higher than the average performance reported in their session.  Of course, these are all noisy measures of dishonesty—e.g., they misclassify some honest leaders who actually obtained scores of 6 in Stage 1. Our goal is to use a measure that captures a tendency toward high scores, which likely captures a greater proportion of those who misreport. To the extent there is noise in our classification, the results will likely underestimate the impact of an unethical leader.  

13

conditions, which we pool here, depending on whether the leader is classified as dishonest or

not. The graph also includes, as “Period 0,” the average Stage 1 performance of leaders and

workers in the respective conditions. In all cases, there is a general tendency for the mean

report to be greater than 3.5.12

Figure 1: Average reported performance over time by leader type

In the Inactive Leader condition, average reported performance is fairly stable over

time, with only a slight upward trend, and the average reported performance is generally

similar to that from Stage 1. Thus, while there is some evidence of misreporting in the

absence of active leaders, its incidence does not change much over time—with the exception,

perhaps, of the last period—and remains relatively modest in magnitude.

In contrast, reported performance is higher and increases more strongly over time in

conditions with active leaders. This increase in reported performance is stronger for groups

with leaders classified as dishonest. Considering the average reported performance by a

worker across all periods, the mean report is higher for those led by an active leader classified                                                                                                                12 Since, in Stage 2, we have multiple observations per worker, we can also explore the tendency to misreport at the individual level. Across all conditions the mean of the 10 reported die rolls significantly differs from the expected mean of 3.5 for 54.17% of all workers (i.e., comparisons using t-tests yield significance levels of p<0.1). Of those, only 4.62% report a lower mean than 3.5, which is roughly what one would expect with chance, and the lowest mean is 2.1. This suggests that the tendency to decrease performance reports to benefit others is not widespread in our experiment.

3.5

44.

55

5.5

Mea

n re

porte

d pe

rform

ance

0 2 4 6 8 10Period

Inactive Leader Active Leader (other)Active Leader (dishonest)

Bars indicate standard errors; Period 0 corresponds to mean report by workers in Stage 1

14

as dishonest (4.82), than for those led by other active leaders (4.54) or without an active

leader (4.28), and both comparisons are statistically significant (p = 0.032 and p = 0.003,

respectively, two-sided Mann-Whitney tests).13 Moreover, the distributions also differ

substantially between workers led by dishonest leaders and other workers (see Appendix

Figure A.2). For example, the frequency of reports of 6 is 34% with inactive leaders, 36%

with active leaders not classified as dishonest, and 53% with dishonest active leaders.

Table 3 tests the relationship between leaders’ honesty and followers’ behavior.

Model 1 regresses the reported performance of workers in Stage 2 on whether there is an

active leader (i.e., whether the condition is Leader, Leader Incentives Only or Leader

Statements Only) and on whether the active leader is also a dishonest leader. That is, the

coefficient for “active leader” measures the effect of having an active leader who is not

dishonest, while “dishonest leader” measures the additional effect of having a dishonest

leader. We also control for time trends and for a worker’s propensity to misreport, measured

by the worker’s reported die roll in Stage 1. Given the high frequency of reports at the upper

boundary of 6, we use a tobit specification in all analyses of workers’ reported performance.

However, the results are similar using either linear or ordered probit specifications.14

Consistent with Figure 1, model 1 shows that simply having an active leader produces

slightly higher reported performance than an inactive leader, but this effect is not statistically

significant. However, having a dishonest leader has a large and statistically significant

additional effect on reported performance.15 Model 2 introduces interaction effects to allow

for different time trends across the conditions. The indicator variables for active and

dishonest leaders are now both statistically insignificant, but there is a positive time trend in

all conditions and, consistent with Figure 1, the time trend for active leaders is significantly

more positive than that for inactive leaders. The time trend for dishonest leaders is large in

magnitude, indicating that mean reports increase by 0.118 more per period than those for

groups led by active leaders who are not dishonest. While this coefficient is not statistically

                                                                                                               13 As a placebo, we would not expect dishonest leaders to affect reported performance in the Inactive Leader condition. Indeed, average reported performance is very similar for dishonest (4.25) and other (4.29) leaders. 14  In Appendix C, we report linear and ordered probit models for all tobit specifications employed in the paper. Note that only one observation is right censored in model 4—in only one group did all workers report 6 in all periods—suggesting that the tobit specification is unnecessary; comparison with the linear specification in Appendix Table C.3a reveals almost identical results.  15  Recall that the coefficient for dishonest leader identifies the marginal effect of having an active dishonest leader over merely having an active leader. The sum of coefficients for dishonest leader and active leader provides the effect of a dishonest leader, relative to an inactive leader. This statistic is positive and statistically significant in models 1 (χ2(1) = 12.69, p < 0.001), 3 (F1,237=8.29, p = 0.004) and 4 (F1,77 = 8.94, p = 0.004).    

15

significant, tests of the restriction that 𝛽dishonest leader + 𝑡  𝛽dishonest leader*period equals zero are

rejected at p < 0.05 for all t ≥ 3 (χ2(1) ≥ 4.36). Hence model 2 provides statistical support for

the persistent differences that arise for dishonest leaders after three periods in Figure 1.

Finally, models 3 and 4 test the effects of the different leader types when aggregating the

data, respectively, at the level of the individual or group. In both cases, dishonest leaders

produce greater misreporting than either inactive leaders or than other active leaders and the

coefficients are at least marginally statistically significant.

Table 3: Leader honesty and reported performance

Dependent variable Reported Stage 2 performance (1) (2) (3) (4)

Unit of analysis: Worker in a period Worker mean

across periods

Group mean across periods

Active leader 0.280 (0.235)

-0.176 (0.282)

0.235 (0.164)

0.263* (0.153)

Dishonest (active) leader 0.887***

(0.292) 0.275

(0.426) 0.384** (0.187)

0.319* (0.168)

Period 0.151*** (0.024)

0.063** (0.026)

Active leader * Period 0.084** (0.038)

Dishonest leader * Period 0.118 (0.075)

(Mean) Stage 1 report by worker(s)

0.366*** (0.058)

0.366*** (0.059)

0.193*** (0.040)

0.099 (0.060)

Constant 2.343*** (0.327)

2.814*** (0.316)

3.430*** (0.224)

3.822*** (0.318)

Number of observations 2400 2400 240 80 Obs. left censored (≤ 1) 142 142 0 0 Obs. right censored (≥ 6) 952 952 16 1 Number of workers, firms 240, 80 240, 80 240, 80 - , 80 Log Likelihood -3766.09 -3757.09 -315.57 -69.40

Wald / F χ2(4) = 84.9 (p < 0.001)

χ2(6) = 83.6 (p < 0.001)

F3,237 = 10.3 (p < 0.001)

F3,77 = 3.63 (p = 0.017)

Notes: Active leader equals 1 in all conditions with an active leader; active (dishonest) leader equals 1 when the group has an active leader and this leader reported a 6 in Stage 1. All models report tobit regressions. Models 1 and 2 are panel specifications that use a worker’s performance report in a period as an observation and include worker random effects, with bootstrapped standard errors clustered at the firm level. Model 3 uses the average report by a worker across all periods, with standard errors clustered at the firm level. Model 4 uses the average report within a firm across all periods, with robust standard errors. * p < 0.1; ** p < 0.05; *** p < 0.01.

16

3.3 How do leaders affect workers’ honesty?

We next investigate more precisely how the strategies employed by leaders affect

workers’ behavior. We begin with the ability to distribute financial incentives and then

analyze the role of statements.

Table 4: Reported performance and leader’s incentive use

Dependent variable Reported Stage 2 performance (1) (2) (3) (4) (5) Bonus / perf. shares corr. up to previous period (𝜌!"!!)

1.394***

(0.162) 1.181*** (0.174)

1.389*** (0.223)

1.324*** (0.214)

1.123*** (0.223)

Communication condition * 𝜌!"!!

0.017 (0.314)

Dishonest leader * 𝜌!"!! 0.204 (0.356)

Dishonest worker * 𝜌!"!! 0.624 (0.456)

Stage 1 report by worker 0.380*** (0.080)

0.319*** (0.077)

0.380*** (0.082)

0.386*** (0.082)

0.369*** (0.079)

Previous period performance report 0.236*

(0.141)

Previous period bonus share -0.712 (1.476)

Previous period performance report * bonus share 0.184

(0.403)

Constant 3.458*** (0.394)

2.646*** (0.565)

3.459*** (0.399)

3.413*** (0.415)

3.553*** (0.392)

Number of observations 1080 1080 1080 1080 1080 Obs. left censored (≤ 1) 51 51 51 51 51 Obs. right censored (≥ 6) 444 444 444 444 444 Number of workers, firms 120, 40 120, 40 120, 40 120, 40 120, 40 Log Likelihood -1667.79 -1654.36 -1667.79 -1667.65 -1666.31

Wald χ2(2) = 83.7 (p < 0.001)

χ2(5)=118.6 (p < 0.001)

χ2(3) = 89.7 (p < 0.001)

χ2(3) = 86.9 (p < 0.001)

χ2(3) = 82.0 (p < 0.001)

Notes: Data from periods 2 through 10 of both conditions in which leaders allocate bonuses (Leader, Leader Incentives Only). Bonus / performance shares correlation up to previous period (𝜌!"!!) is the overall correlation between performance shares and bonus shares demonstrated by the leader up through the prior period. All models report panel tobit regressions that use a worker’s performance report in a period as an observation and include worker random effects, with bootstrapped standard errors clustered at the firm level. * p < 0.1; ** p < 0.05; *** p < 0.01.

In conditions with leader incentive power, leaders can choose how to distribute

rewards among workers, after observing their reported performance. Leaders can either

reward high or low reported performance or they can distribute the bonus independently of

workers’ performance reports. If workers use their leader’s past behavior to form

17

expectations of how rewards will be distributed, this can affect the expected profitability of

inflated performance reports. Therefore, we construct a measure, for each firm in period 𝑡, of

the cumulative correlation between workers’ performance shares and their bonus shares

through period 𝑡 − 1. Specifically, let 𝑠!"# = 𝑟!"# 𝑠!"#!!!! represents worker 𝑖’s share of the

total performance reports in a period, and let 𝑏!"# = 𝑏!"# 𝑏!"#!!!! be the share of the total

bonus received by worker 𝑖. Then, for a leader 𝑓 in period 𝑡, 𝜌!" is the Pearson correlation for

all 𝑠!"# and 𝑏!"# up through period 𝑡. This measures leaders’ use of incentives to encourage

dishonesty up to a particular period: 𝜌!" > 0 indicates that a leader provided a greater share

of the bonus to workers reporting high relative performance, 𝜌!" < 0 indicates that the leader

provided a greater share for low relative performance and 𝜌!" = 0 indicates that the leader

assigned bonus shares independently of relative performance.16

Table 4 studies how the leader’s prior incentive use affects workers’ performance

reports. The dependent variable is a worker’s reported performance in a period and the key

independent variable is the cumulative correlation between performance shares and bonus

shares awarded by the leader, through the previous period (𝜌!"!!). As before, we employ a

tobit specification due to the large number of observations at the upper boundary of 6.17 All

models control for a worker’s performance report in Stage 1. Model 1 shows that

performance reports are strongly positively impacted by how the leader distributed bonuses in

the past. Model 2 controls for the worker’s personal experience with performance and bonus

shares in the previous period, by including the worker’s lagged performance report, lagged

share of bonus received, and their interaction. Workers who reported larger numbers in the

previous period are likely to report higher numbers again, but their share of the bonus

received and its interaction with their own bonus add little explanatory power to the

cumulative correlation. Models 3, 4 and 5 test whether the performance-bonus share

correlation has differential effects when, respectively, leaders can send messages, a firm’s

                                                                                                               16 The variable is defined as long as both 𝑠!"# and 𝑏!"# exhibit some variation in periods 1  . . . 𝑡 (i.e., 𝑏!"# ≠ 1 3 for some 𝑖, 𝑡 and 𝑏!"# ≠ 1 3 for some 𝑖, 𝑡). Since this is not always satisfied, we set 𝜌!" = 0 for undefined cases, reflecting the fact that workers at that point have no information on how the leader distributes bonuses. This is sensible if the lack of variation occurs only because 𝑏!"# = 1 3 for all periods (while 𝑠!"# ≠ 1 3 in at least one period), since this indicates the leader allocated bonuses equally even though performance shares varied. The cases in which 𝑠!"# = 1 3, for all 𝑖, 𝑡, are harder to interpret, since this means workers have no information on how the leader rewards high relative performance reports. If we drop all such observations from our analysis (5 percent of cases), the results are unchanged. 17 Tables C.4a and C.4b in Appendix C report analogous linear and ordered probit specifications. The only substantive difference is that previous period performance report has a marginally statistically significant coefficient in the tobit and ordered probit versions of model 2, which is highly significant in the linear model.  

18

leader is classified as dishonest and a worker is similarly classified as dishonest (i.e., the

worker reported a score of 6 in Stage 1). None of these interactions yields any significant

differences. Most importantly, across all regressions, we see that how the leader distributed

bonuses in the past strongly affects reported performance.

We next explore the effect of leaders’ statements on workers’ reported performance.

To this end, we conducted content analysis, relying on classifications of the statements by

three independent coders. We initially developed eight categories that distinguish between

messages encouraging high and low reported performance. Specifically, the categories

request high and request low identify specific requests for high or low performance numbers,

praise high and praise low identify praise for having reported high or low numbers, dishonest

and honest identify direct appeals to either honesty or dishonesty, and bonus high and bonus

low identify instances (in the condition with leader incentives) in which the leader states that

bonuses will be awarded for either high or low performance. Coders observed the statement

made by a particular leader in a period, and then identified which categories applied to that

particular statement. Where there was disagreement, we use the median coder’s rating.18

Using these category codings, we construct a measure, 𝜔!" ∈ −1,1 , of the tendency

of a leader to make statements exerting upward (request high, praise high, dishonest, bonus

high) versus downward (request low, praise low, honest, bonus low) pressure on workers’

reported performance in a period. Specifically, we sum the number of high categories

satisfied by a leader’s statements in a period, subtract the number of low categories satisfied

in that period, and divide this sum by the total number of either positive or negative

categories possible in that firm’s condition.19

                                                                                                               18 Table B.4 in the Appendix reports all message categories, their definitions, and frequencies. The first eight categories are the ones that we believed, ex ante, would be the most relevant, so we focus our analysis on these. For completeness, we also included additional categories, not used in the current analysis, that identify messages referring to other groups’ performance, to the size of the prize or to the group’s share of the prize, and residual categories for messages containing apologies, jokes, general encouragement, or miscellaneous messages. These additional categories generally have little statistically significant effect on reported performance (see Table B.5 in the Appendix), with two exceptions. First, references to other groups have a marginally significantly positive effect on performance reports. Second, humor has a larger and statistically stronger positive relationship, though it is reasonable that the causality might run the other way (high performance reports elicit humorous comments from the leader). Since we did not have strong priors regarding these message categories’ relationships with misreporting, we exclude them from our primary analysis.  19  That is, 𝜔!" = 𝑘!"#

!!"! − 𝑘!"#!"#!!!!

!!!! 𝐾, where 𝑘!"#

!!"! ∈ {request highft, praise highft, dishonestft, bonus highft}, 𝑘!"#!"# ∈ {request lowft, praise lowft, honestft, bonus lowft} and 𝐾 = 4 in the Leader condition and 𝐾 = 3 in the Leader statements only condition. Hence, a score of 𝜔!" = 1 (𝜔!" = −1) indicates that the leader used all of the forms of content to encourage high (low) numbers and none to encourage low (high) numbers.

19

Table 5: Reported performance and leader’s statement use

Dependent variable Reported Stage 2 performance (1) (2) (3) (4) (5) Leader’s encouragement of high performance reports (𝜔!")

2.116***

(0.496) 2.058*** (0.483)

1.450*** (0.494)

1.310*** (0.408)

2.031*** (0.509)

Incentive condition * 𝜔!" 1.668* (0.868)

Dishonest leader * 𝜔!" 2.374*** (0.925)

Dishonest worker * 𝜔!" 0.243 (0.947)

Stage 1 report by worker 0.410*** (0.079)

0.333*** (0.081)

0.403*** (0.076)

0.435*** (0.077)

0.396*** (0.094)

Previous period performance report 0.255***

(0.095)

Constant 3.375*** (0.379)

2.629*** (0.472)

3.459*** (0.399)

3.204*** (0.368)

3.438*** (0.428)

Number of observations 1200 1080 1200 1200 1200 Obs. left censored (≤ 1) 59 50 59 59 59 Obs. right censored (≥ 6) 565 533 565 565 565 Number of workers, firms 120, 40 120, 40 120, 40 120, 40 120, 40 Log Likelihood -1756.76 -1531.14 -1750.33 -1744.61 -1756.65

Wald χ2(2) = 44.0 (p < 0.001)

χ2(3) = 51.9 (p < 0.001)

χ2(3) = 54.6 (p < 0.001)

χ2(3) = 60.1 (p < 0.001)

χ2(3) = 45.0 (p < 0.001)

Notes: Data from both conditions in which leaders send statements (Leader, Leader Statements Only). Leader’s encouragement of high performance reports indicates the relative use of statement categories encouraging high (1) versus low (-1) performance reports. Models 1, 3 and 4 use all periods of data; model 2 uses periods 2 through 10. All models report panel tobit regressions that use a worker’s performance report in a period as an observation and include worker random effects, with bootstrapped standard errors clustered at the firm level. * p < 0.1; ** p < 0.05; *** p < 0.01.

Table 5 reports random-effects tobit regressions of workers’ reported performance in

a period on the degree to which the leader’s statements at the start of the period encourage

high performance reports (𝜔!"). As before, we employ a tobit specification and control for

the worker’s reported score in Stage 1.20 Model 1 shows that statements encouraging the

reporting of high numbers leads workers to do so. In model 2, we additionally control for a

worker’s tendency to misreport in the previous period, which does not change the results.

Model 3 tests whether the degree to which statements influence followers differs between the

                                                                                                               20 In Appendix C, Tables C.5a and C.5b report linear and ordered probit specifications of the same models. The results do not change in either case. Appendix Table B.6 analyzes the relationship between individual categories and workers’ reported performance. The signs of all relationships are what one would expect and all categories except for Praise high and Praise low demonstrate statistically significant relationships. The strongest statistical relationships appear to be for Request high, Request low, Honest and Reward low.

20

two conditions in which leaders made statements: Leader and Leader Statements Only. The

variable, Incentive condition * 𝜔!", identifies whether statements by leaders who could also

set incentives (in the Leader condition) had a differential effect. The results indicate that

statements by such leaders were more effective, though the effect even for leaders without

this ability remains positive and statistically significant.21 Model 4 studies whether statements

from leaders classified as dishonest were differentially effective. This is the case: statements

from dishonest leaders have a greater impact than statements from other leaders. While we

are cautious, given the post hoc nature of these two findings, they suggest that statements

requesting unethical conduct can have complementary effectiveness with the characteristics

of and tools available to a leader. Finally, model 5 tests whether statements directed toward

workers who reported 6 in Stage 1 were more effective, finding no relationship.

Figure 2 provides a graphical representation of the above relationships, using the firm

as the unit of observation. The vertical axis in each graph indicates the impact a leader has on

the reports by workers. Specifically, it shows the average performance reports by all three

workers in a firm across all 10 periods of Stage 2 minus the three workers’ reported scores in

Stage 1. The horizontal axes show the leaders’ tendencies to employ the above two

instruments—incentives and statements—to encourage high performance reports over the

course of the experiment. In Figure 2A, this corresponds to the correlation between the

performance shares and the bonus shares awarded by a leader across all 10 periods of Stage 2

(𝜌!!"); in Figure 2B, it is the leader’s average tendency to use statements eliciting high

performance reports (𝜔! = 𝜔!"!"!!! 10). Both graphs provide visual evidence of the

statistical relationships in Tables 4 and 5, aggregated at the level of the firm.22

An additional interesting observation arises when comparing the two graphs. In

Figure 2A, there is considerable dispersion in how leaders award bonuses across Stage 2,

from a minimum of -0.885 to a maximum of 0.997. The mean and median values are both

close to zero (𝜌!!"!"#$ = 0.134; 𝜌!!"!"#$%& = 0), with the former not statistically different from

zero (𝑡!" = 1.566,𝑝 = 0.125). Moreover, 16 of 40 leaders (40 percent) have negative values

of 𝜌!!". Figure 2B, however, reveals a different pattern. First, there is considerably less

dispersion (𝜔!!"# = −0.25;  𝜔!!"# = 0.6), which might be due to the coarse method in which

                                                                                                               21 Based on this finding, we also studied whether leaders with incentive power employ different communication strategies than those without. The frequencies of request high, request low, praise high, praise low, dishonest and honest do not differ much by leader type (never by more than 9.5%) and are never statistically significant. 22   Appendix Figure A.3A and A.3B provide analogous graphs using the actual average Stage 2 reported performance, rather than the change in reported performance. The relationships are similar.  

21

this variable is constructed from a few noisy categories. However, there is also a clear

positive tendency in leaders’ statement use scores (𝜔!!"#$%& = 0.271;  𝜔!!"#$ = 0.25, 𝑡!" =

7.66,𝑝 < 0.001), with 36 out of 40 leaders (90 percent) exhibiting a tendency to employ

more messages requesting high, rather than low, performance reports.

Figure 2A. Effect of leader’s incentive use on workers’ performance reports

Figure 2B. Effect of leader’s statement use on workers’ performance reports

-3-2

-10

12

3Av

g. w

orke

r rep

orte

d pe

rform

ance

cha

nge

-1 -.5 0 .5 1Overall leader incentive use encouraging high reports

Leader Incentives Only Leader

-3-2

-10

12

3Av

g. w

orke

r rep

orte

d pe

rform

ance

cha

nge

-1 -.5 0 .5 1Avg. leader statement use encouraging high reports

Leader Statements Only Leader

22

3.4 Structural estimates of leader influence

To complement the above reduced-form analysis of leaders’ influence on followers,

we estimated a structural model in the spirit of Equation 1. This also allows us to explore

possible heterogeneity in workers’ responses to leaders’ actions.  

Our design has 𝑖 = 1, . . , 𝐼 workers in 𝑓 = 1, . . ,𝐹 firms, each of them providing a

performance report in each of 𝑡 = 0, . . ,10 periods, where 𝑡 = 0 corresponds to Stage 1 and

𝑡 ≥ 1 to Stage 2. Once a worker rolls the die, she can choose to report an outcome, 𝑟!"#.

Assume that the worker does so with the objective of maximizing an expected utility function

based on that in Equation 1, with:

𝐸 𝑢(𝑟!"# , 𝐿) = 𝐸 𝜋(𝑟!"# , 𝐿!) − 𝑐(𝑟!"# , L!). (6)

We rewrite this general utility function with one specifically suited to our experimental data:

𝐸 𝑢(𝑟!"# , 𝐿) = 𝜆! 0.25+ 𝛾!Δ𝑌!"#$ 𝑆!"#𝑉!" + 𝛼! + 𝛽!𝑀!"#$ 𝐶!"#$ + 𝜀!"#$, (7)

using the subscript, 𝑗, to denote the case in which 𝑟!"# = 𝑗 ∈ 1, . . ,6 . Hence, 𝑉!" is the

expected total prize available for all firms in the session in the case where the worker reports

a performance of 𝑗 and group 𝑓 receives an expected group share of this prize, 𝑆!"#. The

parameter, 𝜆!, captures how much the worker cares about her own monetary payoff. We

assume that expectations, where necessary, are formed by each worker believing that all the

other workers will behave as observed in the previous period. Hence, given a performance

report 𝑟!"# by worker, 𝑖, then 𝑉!" and 𝑆!"# follow immediately from this assumption.

The leader’s use of incentives is captured by Δ𝑌!"#$, which represents the expected

departure for the worker from a 25% share of the group’s earnings from reporting

performance, 𝑗, based on the worker’s expectation of how the leader will distribute this share.

As in our reduced-form analysis, we assume that workers look back at the history of the

leader’s allocation of the bonus in prior periods, and compute an expected share of the total

group reward for each possible performance report. Hence, 𝛾! measures the extent to which

the worker’s expectations of her own earnings are influenced by the leader’s distribution of

bonuses in the past. When 𝛾! equals zero, for example, the worker might not attend to how

the leader has distributed bonuses in the past when forming expectations.

The second term models how the leader’s statements influence workers’ perceptions

of the appropriateness of misreporting. Let 𝐶!"#$ measure of the degree to which reporting

23

outcome 𝑟!"# = 𝑗 is unethical. Then, 𝛼! + 𝛽!𝑀!"#$ represents the degree to which the worker

experiences psychological (dis)utility from taking this act. Our design does not allow us to

observe directly the degree to which a worker is lying, which is the difference between 𝑟!"#

and the worker’s actual performance in that period. Hence, we proxy “dishonesty” as the

deviation of a worker's cumulative reported performance up to period 𝑡 from 3.5. For each

worker/firm/period/reported outcome we compute a measure of “cumulative dishonesty:”  

𝐶!"#$ =0, 𝑖𝑓   !!"#

!!!!!!! ≤ 3.5

!!"#!!!

− 3.5!!!! , 𝑖𝑓   !!"#

!!!!!!! > 3.5

(8)

This index will be 0 if the average of the worker's reported rolls up to period 𝑡 is less than

3.5, assuming that there is no clear psychological impact from reporting outcomes lower than

or equal to the expected value of a die roll. This index will increase the higher is the worker's

average of reported outcomes above 3.5. Hence, 𝐶!"#$ is a simple measure of how much the

worker’s reports, 𝑟!"!,… , 𝑟!"#, reflect dishonestly high reporting.

The parameter, 𝛼!, identifies how much the worker cares, intrinsically, about 𝐶!"#$. A

worker who dislikes appearing dishonest will have 𝛼! < 0. This intrinsic weight can be

modified by a leader’s statements. Specifically, let,

𝑀!"#$ =0,                                                                            𝑖𝑓  𝑙𝑒𝑎𝑑𝑒𝑟  𝑠𝑒𝑛𝑑𝑠  𝑛𝑜  /  𝑎𝑚𝑏𝑖𝑔𝑢𝑜𝑢𝑠  𝑠𝑡𝑎𝑡𝑒𝑚𝑒𝑛𝑡1− 0.4 𝑟!"# − 1 ,                                                              𝑖𝑓  𝑙𝑒𝑎𝑑𝑒𝑟  𝑟𝑒𝑞𝑢𝑒𝑠𝑡𝑠  𝑙𝑜𝑤  𝑟𝑒𝑝𝑜𝑟𝑡𝑠−1+ 0.4 𝑟!"# − 1 ,                                                      𝑖𝑓  𝑙𝑒𝑎𝑑𝑒𝑟  𝑟𝑒𝑞𝑢𝑒𝑠𝑡𝑠  ℎ𝑖𝑔ℎ  𝑟𝑒𝑝𝑜𝑟𝑡𝑠

         (9)  

capture the leader’s use of statements in a period to influence reports. If the leader sends no

message or an ambiguous message, then 𝑀!"#$ = 0 and a worker who intrinsically cares

about appearing honest (𝛼! < 0) will experience disutility 𝛼!𝐶!"#$. If the leader makes a

request for either high or low numbers,23 then we define 𝑀!"#$ = 1 when the worker follows

this request (i.e., 𝑟!"# = 1 in response to a low request, 𝑟!"# = 6 in response to a high request)

and 𝑀!"#$ = −1 when the worker makes the most contrary report (i.e., 𝑟!"# = 6 in response to

a low request, 𝑟!"# = 1 in response to a high request). The term, 𝛽!, then identifies the extent

                                                                                                               23  We define a leader requesting high (low) reports in a period if the leader uses more of any of the message categories that we earlier classified as “high” (“low”) request than of any the categories that correspond to “low” (“high”) requests, i.e., 𝜔!" > 0 (𝜔!" < 0). A leader who makes a statement with the same number of both types of categories or with neither type of category (𝜔!" = 0) is classified as sending no or an ambiguous message.  

24

to which a worker’s concern for appearing dishonest, when 𝐶!"#$ > 0, is influenced by the

leader’s statements.24

We estimated the model in Equation 7 as a two-component finite mixture model using

maximum likelihood. We construct the following log-likelihood function:

𝐿 𝑝,𝜃 = log   𝑝! 𝑃𝑟 𝑟!"#$ = 𝑗|𝜃! ,𝑋!"#$!!"#$

!!!! (10)

where 𝑃𝑟 𝑟!"#$ = 𝑗|𝜃! ,𝑋!"#$ is the logit probability of reporting a roll 𝑗 based on the utility

function specified above, conditional on the 𝑘 type-specific parameters 𝜃! = 𝜆! ,𝛼! , 𝛾! ,𝛽! ;

𝑝! are the type probabilities, 𝑋!"#$ = Δ𝑌!"#$ , 𝑆!"# ,𝑉!" ,𝑀!"#$ ,𝐶!"#$ are the variables identified

from the data as described above, and 𝑑!"#$ are dummies for actual reported performance.

Table 6 reports the results.25 The model estimates 74% of workers as Type 1, who

care about their expected revenue (𝜆 > 0) and dislike (the appearance of) lying (𝛼 < 0). This

type is responsive both to expectations of how the leader will distribute the bonus (𝛾 > 0)

and to the leader’s statements (𝛽 > 0). Hence, consistent with our reduced form analysis, a

large number of workers are influenced by their leaders through both channels. The second

type, consisting of 26%, places relatively greater weight on own revenue and little weight on

appearing dishonest. Presumably, these are subjects who exhibit a high propensity to report

high scores. They are not influenced by how the leaders distribute incentives, but are

sensitive to the leader’s statements in the same manner as Type 1.

These estimates confirm our earlier observations that workers respond to leaders.

Leaders influence a majority of workers both through how they reward reported performance

and through their statements encouraging either high or low performance. The results also

suggest that the different channels available to a leader might influence different types of

workers. In particular, the incentive channel may be more influential toward workers

                                                                                                               24  Suppose, for instance, that the leader requests a high number, then a worker reporting a performance of 6 will experience 𝛽!(𝐶!!!" − 0.6𝐶!!!") higher utility from reporting a 6 than a 5. If, instead, the leader requests a low number, then the worker experiences a −𝛽!(𝐶!!!" − 0.6𝐶!!!") utility impact from reporting a 6 than a 5.  25 If we estimate a one-component model, the main variables of interest, 𝛾 and 𝛽, which measure leader influence,  are both positive and statistically significant, consistent with the estimates in Table 6. However, 𝛼 is positive and statistically significant, suggesting that subjects want to be seen as reporting high numbers. Instead, we interpret this as the model’s attempt to capture a disproportionately high frequency of 6 reports (in combination with the presence of lower reports) than can be captured by the utility component alone. The two-component model allows for heterogeneity in propensity to report 6, which better accommodates the data. Moreover, we find a significantly higher value of the likelihood with two types than one type (LR test statistic: χ2(5) = 252.8, p < 0.001). We also experimented with 3 components, finding a third type with zero probability, which suggests the bimodal distribution is sufficient to capture workers’ preferences in the data.

25

intrinsically inclined to care about being or appearing dishonest.

Table 6. Estimates for worker utility model

Type 1 Type 2

Type probability 0.740*** (0.012)

0.260*** (0.012)

𝜆 - Own revenue 0.054*** (0.005)

0.542*** (0.015)

𝛼 – (Dis)utility from lying -0.235*** (0.018)

-0.023 (0.024)

𝛾 – Weight on leader incentive use 0.071*** (0.017)

0.000 (0.011)

𝛽 – Weight on leader statements 0.515*** (0.014)

0.972*** (0.019)

Number of observations 14,400 Number of workers 240 Log likelihood -3,653.91 Note: Parameter estimates of model in Equation 7 using data from all periods in all conditions with active leaders. An observation is a worker’s decision of whether to make a particular performance report in a period. Standard errors in parentheses. * p < 0.1; ** p < 0.05; *** p < 0.01.

3.5 Strategies employed by unethical leaders

We next investigate whether unethical leaders are more likely to employ strategies

that encourage dishonest reporting. For this, we return to the earlier classification of leaders

as “dishonest” who reported performance of 6 in Stage 1. We then compare dishonest with

other leaders’ use of statements and incentives to encourage dishonesty. For incentive use, we

compare the overall correlation between bonus shares and performance shares exhibited by a

leader across all 10 periods, 𝜌!!". Dishonest leaders exhibit more positive correlation

between bonus and performance shares (0.269) than other leaders (0.043). For statement use,

we use the average tendency to make statements requesting high, rather than low

performance reports, 𝜔! = 𝜔!"!"!!! 10. Again, we find a greater tendency to encourage

dishonesty among dishonest leaders (0.351) than other leaders (0.201). Appendix Figure A.4

provides graphical representations of how the different types of leaders used incentives and

statements, both on aggregate and over time. The differences are similarly large in magnitude

and they are statistically significant for statements (Mann-Whitney: z = 2.182, p = 0.03) but

not for incentive use (Mann-Whitney: z = 1.080, p = 0.28). If we look only at the first period,

the difference in incentive use is much larger (0.73, z = 3.380, p = 0.007), but this difference

shrinks quickly over time. The difference is fairly constant across periods for statement use.

26

These patterns are confirmed in the regression analysis in Table 7, which studies the

behavior of a leader in a period. The first two models study incentive use, with the

independent variable the correlation between performance and bonus shares in a period—

rather than cumulative through a period.26 We observe a greater tendency for dishonest

leaders to use incentives to encourage dishonesty, but this is not significant in model 1. In

model 2, the difference is initially large and significant, but declines across periods. Models 3

and 4 study statement use, employing our earlier measure of a leader’s tendency in a period

to use messages encouraging high or low performance reports. Models 3 and 4 confirm that

dishonest leaders make more statements encouraging dishonesty. This difference does not

decrease over time, though there is a slight increasing tendency for more statements

encouraging dishonesty across periods for both types of leaders.

Table 7: Leaders’ strategies by leader type

Dependent variable: Incentive use in period

(𝜌!"! ) Statement use in period

(𝜔!") (1) (2) (3) (4)

Dishonest leader 0.152 (0.134)

0.579*** (0.188)

0.150*** (0.055)

0.176** (0.086)

Period 0.032* (0.016) 0.023***

(0.005)

Dishonest leader * period -0.077*** (0.025) -0.005

(0.014)

Constant 0.034 (0.083)

-0.140 (0.123)

0.201*** (0.042)

0.074 (0.050)

Number of observations 400 400 400 400 Number of firms 40 40 40 40 R2 0.013 0.040 0.044 0.078

Wald χ2(1) = 1.30 (p = 0.254)

χ2(3) = 11.1 (p = 0.011)

χ2(1) = 7.45 (p = 0.006)

χ2(3) = 30.3 (p < 0.001)

Notes: Models 1 and 2 include data from both conditions in which leaders allocate bonuses (Leader, Leader Incentives Only). Incentive use in period refers to the correlation between performance shares and bonus shares in a period. Models 3 and 4 include data from all periods of both conditions in which leaders send statements (Leader, Leader Statements Only). Statement use in period is the tendency to use message categories encouraging high (1) or low (-1) performance reports. All models report panel linear regressions that use a leader’s action in a period as an observation and include firm random effects, with standard errors clustered at the firm level. * p < 0.1; ** p < 0.05; *** p < 0.01.

                                                                                                               26  As discussed in footnote 16, we use a value of zero when either the bonus share or performance share exhibits no variation. This is sensible in cases where performance varies but the bonus share does not (i.e., cases in which the leader could have rewarded high or low performers but chose not to do so). In Appendix Table B.7 we repeat the analysis in Table 7, dropping all periods in which the performance shares of the three workers exhibited no variation. The results are very similar.  

27

Hence, we find evidence that leaders who are more dishonest are more likely to use

strategies that encourage followers to act dishonestly, particularly through their statements. In

combination with the analysis in Section 3.3 and the structural analysis in Section 3.4, which

both show that these leader strategies influence performance reports, this provides a basis for

the relationship that we identified in Section 3.2 between a leader’s type—dishonest or

other—and worker’s performance reports.

3.6 The aggregate effects of leadership

We have seen that leaders influence the conduct of followers. We next focus on the

overall effect of leadership, as an institutional feature. Our experimental conditions, which

exogenously vary the presence and abilities of leaders, allow us to do so. Figure 4 shows the

average reported performance across periods in each of the four conditions. This graph does

not draw a distinction between dishonest and other leaders, but instead shows aggregate

performance reports across all firms in a condition.

The figure indicates that both conditions in which leaders make public statements tend

to yield higher misreporting than when leaders are inactive or can only determine incentives.

Average reported performance when leaders can only make statements to workers largely

overlaps that of groups in the Leader condition, where they also control incentives. In

contrast, both the level and the trend of group performance in the Leader Incentives Only

condition follow closely that of groups in the Inactive Leader condition.27 Regression

analysis (see Appendix Table B.8) confirms that these aggregate treatment arise with the

leaders’ abilities to make statements to followers. Hence, while we saw earlier that leaders

can influence followers through their ability to make statements as well as determine

incentives, the overall effects of leadership seem to affect outcomes primarily through the

former. While surprising, this is consistent with our earlier observation that leaders—even

leaders we do not classify as dishonest—are more likely to encourage dishonesty through

their statements than through their incentive use, and that such use of statements increases

over time (see Appendix Figure A.4). Hence, at least in our experiment, the extent to which

leadership encourages dishonesty seems to be more strongly driven by leaders’ statements.

Of course, as these are post hoc observations and interpretations, we note them cautiously.                                                                                                                27 Two-sided Mann-Whitney tests show that reported performance is significantly different between the Inactive Leader and the Leader Statements Only and Leader conditions (p=0.000 and p=0.000), as well as between the Leader Incentives Only and the Leader Statements Only and Leader conditions (p=0.000 and p=0.000). But the difference between the Inactive Leader and Leader Incentives Only condition is insignificant (p=0.734) as well as the difference between the Leader Statements Only and the Leader condition (p=0.144).

28

Figure 4: Average reported performance over time by leader type

Finally, we also explore whether the effects of leadership extend to a broader change

in perceptions of appropriate and inappropriate conduct. At the end of the experiment, we

elicited workers’ social norms regarding the appropriateness of misreporting the outcome of

the die roll in Stage 1. We used the incentivized norm elicitation method introduced by

Krupka and Weber (2013). At the end of each session, subjects were asked to rate the

appropriateness of overstating one’s performance in Stage 1, on a scale from “very

inappropriate” (1) to “very appropriate” (4). Subjects were rewarded if their answer matched

that of a randomly drawn other participant in the session. This procedure captures social

perceptions of appropriateness, a central component of injunctive social norms. Consistent

with leaders having some form of persistent influence, being in any condition with active

leaders yielded perceived norms of conduct that are more lax with respect to misreporting

than being in the Inactive Leader condition (see Appendix Figure A.5).28 Thus, exposure to a

leader who has some channel through which to exert influence on the group changes workers’

perceptions regarding the appropriateness of acting unethically in other environments, where

the leader is not active. This suggests that leaders do more than change behavior by exerting                                                                                                                28 We do not find a significant difference between the appropriateness ratings given by workers who were led by a dishonest leader compared to those led by an honest leader. This might be because the incentives were to match the choice provided by another participant in the session, rather than in one’s firm.

3.5

44.

55

5.5

Mea

n re

porte

d pe

rform

ance

0 2 4 6 8 10Period

Inactive Leader Leader StatementsLeader Incentives Leader (Statements & Incentives)

Bars indicate standard errors; Period 0 corresponds to mean report by workers in Stage 1

29

direct influence; they may also change norms and values.

4. Experiment 2: Absence of intergroup competition

An important feature of our main experiment is that widespread inflation of

performance reports creates a negative externality for all firms and workers, and decreases

the total surplus. As we note earlier, this corresponds to many settings—such as several of

our motivating examples—involving competition and where high levels of unethical conduct

destroy aggregate value. However, it is also worthwhile to investigate to what extent leaders’

effects on followers’ ethical conduct extend to other kinds of environments. For this purpose,

we conducted a second experiment.

4.1 Experimental design

The design closely follows that of the first experiment, apart from the fact that the

negative externality from cheating is not borne by the other groups, but by a passive third

party. Each session now features 20 active participants and 5 inactive participants, with roles

fixed throughout the experiment. Inactive participants do not make any decisions, but their

payoffs are determined by active participants’ reported die roll outcomes.

In Stage 1, each inactive participant is randomly matched with 4 active ones. Each

participant begins with 50 ECU. As before, active participants each privately roll a die and

report the outcome. They each receive a payoff equal to 𝜋!!"#$%& = 50+ 5(𝑟! − 3.5). The

latter term corresponds to the amount that each active participant takes from her matched

inactive participant. Hence, each inactive participant, who did not roll a die, receives a payoff

determined by the reported scores of the four matched active participants: 𝜋!"#$%!&' = 50−

5(𝑟! − 3.5)!!!! . Thus, for example, reporting 𝑟! = 6 means that participant 𝑖 receives a

payoff of 62.5, but if all four active participants report this outcome the inactive participant

receives a payoff of zero. At the end of Stage 1, each subject observes the average reported

performance by all subjects, her own payoff and, if active, her own reported performance.

In Stage 2, firms are again comprised of one supervisor and three workers, randomly

drawn and assigned to firm roles from the set of active participants. Workers again roll dice

to determine performance and are free to enter any performance report. In each period, the

firm is also randomly matched with an inactive participant. As in our main experiment, the

firm’s total payoff is determined by the sum of the three workers’ performance reports. Now,

however, the firm benefits only by taking from the matched passive participant. Specifically,

30

a firm’s payoff in a period is, 𝜋! = 200+ 20   𝑟!!!!! 3 − 3.5 , meaning that the firm

benefits by 20 ECU for each unit increase in the performance report. Conversely, the inactive

participant’s payoff is 𝜋!"#$%!&' = 50− 20   𝑟!!!!! 3 − 3.5 , so that the firms’ benefits

from inflated performance reports come directly from the passive third party.29

We included only two leadership conditions from the first experiment: Inactive

Leader and Leader. In the former, leaders are passive observers and all firm members receive

equal shares of the firm’s payoff. In the latter, as before, leaders make statements to workers

at the beginning of the period and distribute 45 percent of the total payoff as a bonus between

the three workers.

Information at the end of a period is similar to that in our first experiment. After each

period in Stage 2, each active participant observes the reported individual performance levels

and payoff of each worker in her firm; the average performance reported by all firms in that

period and in all previous periods; her firm’s payoff; and the payoff of the inactive participant

matched with her firm for that period. Inactive participants only observe their own payoff and

the average reported performance level for each firm in the current and previous periods.

We conducted the experiment at the Laboratory for Experimental and Behavioral

Economics at the University of Zurich. In total, we conducted three sessions (15 firms, 60

active participants and 15 inactive participants) of the No Leader condition and six sessions

(30 firms, 120 active participants and 30 inactive participants) of the Leader condition.

Sessions were conducted in English, using only subjects proficient in English. At the end of

the session, subjects’ payoffs were converted into cash at a rate of 15 ECU = 1 Swiss Franc.

4.2 Results

We again classify active leaders as “dishonest” or “other” based on whether their

reported Stage 1 score is 6. This yields 11 dishonest leaders and 19 not classified as such.

Figure 5 shows the effects of these different kinds of leaders on reported performance,

providing similar data to Figure 1. The ordering of the three different kinds of firms is similar

to that in Experiment 1: workers in groups with dishonest leaders again exhibit the highest

                                                                                                               29 While we attempted to vary as little as possible relative to our main experiment, some changes were hard to avoid. For example, this payoff system no longer makes misreporting inefficient. Attempting to build this inefficiency into the design would require the gain for the firm exceeding the loss for the inactive participant, and either leaving this participant with negative payoffs or having this participant start off with more wealth (per capita) than the firm members. Our design potentially makes misreporting more attractive than in our first experiment, which we cautiously note is consistent with our data.

31

performance reports, on average, across all 10 periods (5.17), followed by groups with other

leaders (4.94). The lowest performance reports tend to be for firms with No Leader (4.81).

The distributions also show the same patterns as in Experiment 1, with a shift toward higher

reports for firms with dishonest leaders (see Appendix Figure A.6). For example, the

frequency of performance reports of 6 across all 10 periods is 66% for firms with leaders

classified as dishonest, 54% for firms with other leaders and 48% under No Leader.30

Figure 5: Average reported performance over time by leader type (Experiment 2)

Table 8 provides tobit regressions analyzing the effects of different leader types,

analogous to Table 3 for Experiment 1. The magnitude of the coefficients for dishonest

leaders is always positive and is similar in size to those in Table 3; but they are less precisely

estimated and, therefore, not statistically significant.31 Nevertheless, relative to the No Leader

control, we observe statistically significantly higher performance reports in the firms led by

dishonest leaders. For example, the sum of the first two coefficients in model 1, which

                                                                                                               30 Interestingly, the overall levels of dishonesty tend to start off higher in Experiment 2 than in Experiment 1, but increase less over time. While we are cautious to make too much of comparisons between the experiments, a natural interpretation is that efficiency concerns lead workers to misreport less initially in Experiment 1, but that competition in Experiment 1 creates an incentive for firms to “outperform” each other over time. 31 We again employ a tobit specification given the rate of censoring at 6. We provide analogous linear and ordered probit results in Appendix Tables C.7a and C.7b. Results are again substantively similar.

44.

55

5.5

6M

ean

repo

rted

perfo

rman

ce

0 2 4 6 8 10Period

Inactive Leader Active Leader (other)Active Leader (dishonest)

Bars indicate standard errors; Period 0 corresponds to mean report by workers in Stage 1

32

captures the effect of active and dishonest leaders relative to No Leader, is significantly

different from zero (𝜒! 1 = 5.26,𝑝 = 0.02).32 Thus, as with Experiment 1, we find

evidence, though weaker, that unethical leaders yield groups that tend to act unethically.

Table 8: Leader honesty and reported performance (Experiment 2)

Dependent variable Reported Stage 2 performance (1) (2) (3) (4)

Unit of analysis: Worker in a period Worker mean

across periods

Group mean across periods

Active leader 0.533 (0.429)

0.785* (0.470)

0.242 (0.238)

0.218 (0.298)

Dishonest (active) leader 0.844

(0.621) 0.595

(0.611) 0.315

(0.320) 0.173

(0.240)

Period 0.060*** (0.022)

0.080*** (0.026)

Active leader * Period -0.047 (0.047)

Dishonest leader * Period 0.047 (0.076)

(Mean) Stage 1 report by worker(s)

0.602*** (0.115)

0.602*** (0.115)

0.263*** (0.062)

0.247** (0.111)

Constant 2.871*** (0.477)

2.760*** (0.488)

3.676*** (0.271)

3.666*** (0.457)

Number of observations 1350 1350 135 45 Obs. left censored (≤ 1) 56 56 0 0 Obs. right censored (≥ 6) 737 737 28 2 Number of workers, firms 135, 45 135, 45 135, 45 - , 45 Log Likelihood -1722.8 -1722.3 -184.8 -42.8

Wald / F χ2(4) = 39.1 (p < 0.001)

χ2(6) = 41.8 (p < 0.001)

F3,132 = 8.04 (p < 0.001)

F3,42 = 3.73 (p = 0.018)

Notes: Active leader equals 1 in Leader condition; active (dishonest) leader equals 1 when the group has an active leader and this leader reported a 6 in Stage 1. All models report tobit regressions. Models 1 and 2 are panel specifications that use a worker’s performance report in a period as an observation and include worker random effects, with bootstrapped standard errors clustered at the firm level. Model 3 uses the average report by a worker across all periods, with standard errors clustered at the firm level. Model 4 uses the average report within a firm across all periods, with robust standard errors. * p < 0.1; ** p < 0.05; *** p < 0.01.

We also again find evidence that leaders’ actions influence followers. Tables B.9,

                                                                                                               32 This comparison, between groups with no leaders and groups with dishonest leaders, is also statistically significant in model 2 (𝜒! 1 = 5.04, 𝑝 = 0.02) and is marginally statistically significant in model 3 (𝐹(!,!"#) =3.20, 𝑝 = 0.08) and model 4 (𝐹(!,!") = 2.36, 𝑝 = 0.13).

33

B.10 and B.11, in the Appendix, provide analogous results to the reduced-form analysis

reported in Tables 4 and 5 and the structural estimates in Table 6. For brevity, Table 9

summarizes the principal coefficients measuring the effects of leaders’ actions on followers’

reported performance. For comparison, the table also presents these coefficients as a

percentage of the corresponding coefficients for Experiment 1. While the statistical

significance of the relationships is weaker for Experiment 2, the coefficients generally have

the same sign and are at least 60% of the size of the Experiment 1 coefficients. The only

major difference appears to be in the structural estimates for 𝛾, which measures workers’

sensitivity to how the leader distributes incentives. Both Types 1 and 2 demonstrate

insensitivity to how the leader distributes incentives in Experiment 2, while this was only the

case for Type 2 in Experiment 1.33

Table 9: Summary of how leaders’ actions affect followers’ reports (Experiment 2)

Coefficient (Std. error)

Size relative to Exp. 1

coefficient

Incentive use Bonus / perf. shares corr. up to previous period (𝜌!"!!) compare to model 1 in Table 4

0.920***

(0.292) 66.0%

𝛾 – Weight on leader incentive use for Type 1 (0.636) compare to Type 1 (0.740) in Table 6

0.012 (0.015) 16.9%

𝛾 – Weight on leader incentive use for Type 2 (0.364) compare to Type 2 (0.260) in Table 6

0.007 (0.011) ---

Statement use Leader’s encouragement of high performance reports (𝜔!") compare to model 1 in Table 5

1.270

(0.976) 60.0%

𝛽 – Weight on leader statements for Type 1 (0.636) compare to Type 1 (0.740) in Table 6

0.429*** (0.027) 83.3%

𝛽 – Weight on leader statements for Type 2 (0.364) compare to Type 2 (0.260) in Table 6

0.791*** (0.055) 81.4%

Notes: Presents primary coefficients of interest from Tables B.9, B.10, B.11 in Appendix B; these tables present analogous estimates for Experiment 2 to those in Tables 4, 5 and 6 for Experiment 1. Standard errors in parentheses. The second column presents the coefficient from Experiment 2 as a percentage of the coefficient from Experiment 1; “---” is reported for the case in which the Experiment variable was estimate as 0.000. * p < 0.1; ** p < 0.05; *** p < 0.01.

                                                                                                               33 Appendix Figures A.7A and A.7B show scatterplots of the relationships between a leader’s overall uses of incentives and statements to encourage dishonesty and the average change in workers’ reports between Stage 1 and Stage 2 (analogous to Figure 2A and 2B). Both graphs reveal positive relationships, though they are weak for incentive use.

34

Finally, we study whether leaders we classify as dishonest are more likely to employ

incentives and statements in a manner that encourages high performance reports. The results

are qualitatively similar to those in Experiment 1: leaders we classify as dishonest tend to use

incentives and statements in a manner that encourages more misreporting.34 For example,

when we re-estimate models 1 and 3 of Table 7, using Experiment 2 data, the analogous

coefficients for “dishonest leader” are, respectively, 0.191 (132% of the analogous

Experiment 1 coefficient) and 0.094 (63%) and both are at least marginally statistically

significant (p < 0.11).

To summarize, our second experiment studies the effect of leadership on followers’

ethical conduct in a different environment confronting firms than in our main experiment. In

this study, there is no competition between firms and misreporting does not harm other firms.

Nevertheless, we find similar effects of leadership, though smaller and statistically weaker, to

those in our first experiment.

5. Conclusion

We use controlled laboratory experiments to study how leaders influence the

unethical conduct of followers. As we note in the Introduction, the ability of researchers to

address such questions using natural data is complicated by the hidden nature of unethical

conduct, as well as by endogeneity issues that make causal inferences regarding the effects of

leadership difficult. The laboratory environment allows us to address many of these concerns,

while employing a task that incorporates features of real-world unethical conduct.35

Our main finding is that unethical leaders produce unethical behavior on the part of

followers. Even though we only observe Stage 1 unethical conduct imperfectly, a noisy

classification of leader “types” based on this behavior has explanatory power for how much

workers misreport their performance when in groups with active leaders. Two features of our                                                                                                                34  Appendix Figures A.8A and A.8B show the average tendency to use incentives and statements in ways that encourage misreporting, separately for leaders we classify as dishonest and for those we do not. Table B.12 in the Appendix reports regression analysis similar to Table 7.  35  In fact, three of the authors of this paper attempted unsuccessfully to study a causal relationship between real leaders’ ethical characteristics and followers’ conduct. Specifically, we developed a simple task (juggling a ball with alternating feet) and gave amateur football players incentives similar to those in our first experiment: they could inflate their own score by misreporting it, thereby obtaining higher monetary earnings for their own team at the expense of other teams. We attempted to study the causal influence of leaders through priming—half the players were read task instructions by their coach and the other half by an experimenter—and tested whether priming in this task allowed us to detect an effect of leaders rated as less ethical (using the Ethical Leadership scale). In a pilot with 14 teams / leaders, we found no effect, while also discovering very little evidence of score misreporting. There are many reasons for a null effect—from a weak priming manipulation, to an unreliable measure of leader ethical traits, to a task in which intrinsic motivations limit dishonest reporting. These highlight the challenges of attempting to identify the relationships we study here in the field.  

35

experiments make this finding especially important. First, leaders are appointed at random,

which means that the effect of unethical leaders on workers is causal—the selection and

endogeneity issues present in the field are eliminated in our laboratory setting. Second,

followers are never informed of the leader’s Stage 1 behavior. Thus, the effect of leaders

must be through what actions they take in their functions as leaders—making statements to

workers and distributing incentives. Indeed, our analysis reveals direct connections between

these actions by leaders and followers’ unethical conduct. We also show that leaders who are

more likely to have misreported high performance in the first stage are more likely to employ

strategies that encourage misreporting, particularly statements.

Our experimental design also allows us to compare the relative importance of how

leaders use statements and incentives to influence followers. The effect of leaders through

statements appears at least as strong as the influence through incentives, confirming other

findings that the effect of leaders’ communication can be as strong in shaping worker

behavior as reliance on incentives (Brandts & Cooper, 2007; Brandts, Cooper & Weber,

2015; Antonakis, et al., 2016). Importantly, leaders tend to employ more statements

requesting higher “performance” and this produces greater dishonesty among followers.

While the ability to distribute incentives also has an effect on unethical conduct, leaders rely

less on this mechanism, and its aggregate effect is therefore weaker. Of course, our study

includes only one incentive mechanism, and it is possible that alternative, more powerful

incentives might produce stronger overall effects on dishonesty.36 It is also the case that our

study does not identify precisely why statements affect follower conduct. Moreover, the

effectiveness of these two mechanisms may also depend on the environment confronting a

group—the effects of leaders on ethical conduct appear stronger when firms are engaged in

competition than when they are not. Hence, while our experiments produce evidence that

leaders, the channels of influence at their disposal and the environment confronting firms can

jointly influence follower conduct, future work is necessary to fully understand all of these

relationships. Put differently, while this paper shows that leaders can affect ethical conduct,

more work is necessary to understand precisely when and why they have such effects.37

Our results have potentially important policy implications for the design of

organizations and for the selection of leaders and the definition of their responsibilities. Most

                                                                                                               36  We thank two anonymous reviewers for raising this point.  37  For example, Schweitzer, et al. (2004), demonstrate that goal setting can influence unethical conduct. A natural channel for leadership influence may be through the setting of performance targets. Alternatively, a leader’s statements may make such targets more salient, by focusing workers’ attention on them.  

36

importantly, we document a causal relationship between unethical leaders and the unethical

behavior of those they lead. This suggests that the moral qualifications of a corporate leader

can be important, beyond his or her managerial or technical abilities. Moreover, leaders’

statements are a central channel through which such influence occurs and can be as strong an

influence as financial rewards. Thus, rather than focusing solely on incentives and monitoring

as a way to foment more ethical conduct, the “tone at the top” may play an important role.

Importantly, our structural analysis suggests that different workers may respond differently to

different channels of leadership influence, suggesting that workers’ characteristics can be an

important element for whether they can be led to act unethically. Our findings also

underscore the importance of governance and institutional checks on leaders’ authority and

their internal communication, since leaders can spread unethical behavior within an

organization at very little cost, if they wish.

Our study also indicates that the selection of “good” leaders and their use of

strategies, including public statements, to encourage ethical conduct may positively effect

followers’ ethical behavior. Thus, both mechanisms present potential opportunities for

leaders and organizations interested in curbing unethical conduct. The trick appears, at least

partly, to get leaders to use these techniques at their disposal—particularly, to employ public

statements that discourage unethical conduct. Our results suggest that a key approach may be

emphasizing the search for leaders who act ethically themselves and are likely to employ

these instruments to encourage ethical, rather than unethical, employee conduct.

37

References

Antonakis, J., G. d’Adda, R. Weber and C. Zehnder (2016). “On the economic value of

charismatic leadership.” Unpublished manuscript.

Antonakis, J., M. Fenley, and S. Liechti (2012), “Learning charisma: Transform yourself

into the person others want to follow”, Harvard Business Review, 90, 127-147.

Baker, G., R. Gibbons, and K. J. Murphy (1994), “Subjective Performance Measures in

Optimal Incentive Contracts”, Quarterly Journal of Economics, 109, 1125-1156.

Barnard, C.I. (1938). The Functions of the Executive. Cambridge, MA: Harvard University

Press.

Bass, B.M. (1990), Bass and Stogdill’s handbook of leadership: Theory, research, and

managerial applications (3rd ed.), New York: Free Press.

Beekman, G., E. Bulte, and E. Nillesen (2013), “Corruption and economic activity: Micro

level evidence from rural Liberia”, European Journal of Political Economy.

Beresford, D.R., N.B. Katzenbach, and C.B.J. Rogers (2003), “Report of Investigation by the

Special Investigative Committee of the Board of Directors of WorldCom, Inc”,

Clinton, Miss.: Worldcom Incorporated.

Bialik, C. (2015). “How FIFA’s structure lends itself to corruption.” Fivethirtyeight.com,

May 27, 2015. http://fivethirtyeight.com/features/how-fifas-structure-lends-itself-to-

corruption/ (accessed 03/21/2016).

Brandts, J. and D. Cooper (2007), “It’s What You Say Not What You Pay: An Experimental

Study of Manager-Employee Relationships in Overcoming Coordination Failure”,

Journal of the European Economic Association, 5, 1223-1268.

Brandts, J., D. Cooper, and R. Weber (2015), “Legitimacy, Communication, and Leadership

in the Turnaround Game”, Management Science, 61(11): 2627-2645.

Brown, K. and I.J. Dugan (2002), “Arthur Andersen's Fall From Grace Is a Sad Tale of Greed

and Miscues”, Wall Street Journal.

Brown, M.E., and L.K. Treviño. "Ethical leadership: A review and future directions." The

Leadership Quarterly 17.6 (2006): 595-616.

Brown, M.E., L.K. Treviño, and D.A. Harrision (2005), “Ethical leadership: A social learning

perspective for construct development and testing”, Organizational Behavior and

Human Decision Processes, 97, 117-134.

38

Burks, S.V. and E. Krupka (2012), “A Multimethod Approach to Identifying Norms and

Normative Expectations Within a Corporate Hierarchy: Evidence from the Financial

Services Industry”, Management Science, 58, 203-217.

Burns, J.M. (1978), Leadership, New York: Harper & Row.

Chattopadhyay, R. and E. Duflo (2004), “Women as Policy Makers: Evidence from a

Randomized Policy Experiment in India”, Econometrica, 72, 1409-1443.

Conrads, J., B. Irlenbusch, R. M. Rilke, A. Schielke, and G. Walkowitz (2014), “Honesty in

Tournaments”, Economic Letters, 123, 90-93.

Conrads, J., B. Irlenbusch, R. M. Rilke, and G. Walkowitz (2013), “Lying and Team

Incentives”, Journal of Economic Psychology, 34, 1-7.

Cremer, A. and T. Bergin (2015). “Fear and respect: VW’s culture under Winterkorn.”

Reuters.com, Saturday October 10, 2015. http://www.reuters.com/article/us-

volkswagen-emissions-culture-idUSKCN0S40MT20151010 (accessed 16/03/2016).

Detert, J.R., L.K. Treviño, E.R. Burris, and M. Andiappan (2007) “Managerial Modes of

Influence and Counterproductivity in Organizations: A Longitudinal Business-Unit-

Level Investigation”, Journal of Applied Psychology, 92, 993-1005.

Fischbacher, U. (2007), “z-Tree: Zurich Toolbox for Ready-Made Economic Experiments”,

Experimental Economics, 10, 171-78.

Fischbacher, U. and F. Föllmi-Heusi (2013), “Lies in Disguise – An Experimental Study on

Cheating”, Journal of the European Economic Association, 11, 525-47.

Gibson, R., C. Tanner, and A. F. Wagner (2013), “Preferences for Truthfulness:

Heterogeneity Among and Within Individuals”, American Economic Review, 103,

532-48.

Gino, F., S. Ayal, and D. Ariely (2009), “Contagion and Differentiation in Unethical

Behavior: The Effect of One Bad Apple on the Barrel”, Psychological Science, 20,

393-398.

Gino, F., E. Krupka, and R.A. Weber (2013), “License to Cheat: Voluntary Regulation and

Ethical Behavior”, Management Science, 59, 2187-2203.

Gneezy, U. (2005), “Deception: The Role of Consequences”, American Economic Review,

95, 384-394.

Greiner, B. (2015), “Subject pool recruitment procedures: Organizing experiments with

ORSEE.” Journal of the Economic Science Association, 1(1): 1-12.

39

Güth, W., M.V. Levati, M. Sutter, and E. Van der Heijden (2007), “Leading by example with

and without exclusion power in voluntary contribution experiments”, Journal of

Public Economics, 91, 1023-1042.

Hamman, J. R., R. Weber and J. Woon (2011), “An Experimental Investigation of Electoral

Delegation and the Provision of Public Goods”, American Journal of Political

Science, in press.

Hermalin, B. E. (1998), “Toward an Economic Theory of Leadership: Leading by Example”,

American Economic Review, 88, 1188-1206.

Houser, D., S. Vetter, J. Winter (2012), “Fairness and Cheating”, European Economic

Review, 56, 1645-55.

Jack, B.K. and M.P. Recalde (2015), “Leadership and the voluntary provision of public

goods: Field evidence from Bolivia”, Journal of Public Economics, 122, 80-93.

Jones, B. F. and B. A. Olken (2005), “Do Leaders Matter? National Leadership and Growth

Since World War II”, The Quarterly Journal of Economics, 120, 835-864.

Kant, I. (1785/1964), “Groundwork of the Metaphysic of Morals”, (H. J. Paton, Trans.). New

York: Harper & Row. (Original work published 1785)

Kosfeld, M. and D. Rustagi (2015), “Leader Punishment and Cooperation in Groups:

Experimental Field Evidence from Commons Management in Ethiopia”, American

Economic Review, 105, 747-83.

Kotter, J.P. (1990), “Force for change: How leadership differs from management”, Simon and

Schuster.

Krupka, E.L. and R.A. Weber (2013), “Identifying Social Norms Using Coordination Games:

Why Does Dictator Game Sharing Vary?”, Journal of the European Economic

Association, 11, 495-524.

Lazear, E. P. (2012), “Leadership: A Personnel Economics Approach”, Labour Economics,

19, 92-101.

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, 151-

171.

Mazar, N., O. Amir, and D. Ariely (2008), “The Dishonesty of Honest People: A Theory of

Self-Concept Maintenance”, Journal of Marketing Research, 45, 633-44.

Milgram, S. (1963), “Behavioral Study of Obedience”, Journal of Abnormal and Social

Psychology, 67, 371-78.

40

Mill, J.S. (1863), “Utilitarianism”, London: Parker, Son and Bourn.

Nagin, D., J. Rebitzer, S. Sanders and L. Taylor (2002), “Monitoring, motivation and

management: The determinants of opportunistic behavior in a field experiment”, The

American Economic Review, 92, 850-873.

Potters, J., M. Sefton, and L. Vesterlund (2007), “Leading-by-example and signaling in

voluntary contribution games: an experimental study”, Economic Theory, 33, 169-

182.

Prendergast, C. (1999), “The Provision of Incentives in Firms”, Journal of Economic

Literature, 37, 7-63.

Schweitzer, M. E., L. Ordóñez, and B. Douma. “Goal setting as a motivator of unethical

behavior.” Academy of Management Journal 47.3 (2004): 422-432.

Toffler, B.L. and J. Reingold (2004), Final Accounting: Ambition, Greed and the Fall of

Arthur Anderson, Currency/Doubleday.

41

A Appendix Figures

Figure A.1: Overview of the experimental design (Experiment 1)

42

Figure A.2: Distribution of reported performance in Stage 2 by Leader Type

0.2

.4.6

0 2 4 6 0 2 4 6 0 2 4 6

Inactive leader Active leader (other) Active leader (dishonest)D

ensi

ty

Workers' reported performance (Stage 2, all periods)Graphs by Leader Type

0.2

.4.6

0 2 4 6 0 2 4 6 0 2 4 6

Inactive leader Active leader (other) Active leader (dishonest)

Den

sity

Workers' reported performance (Stage 2, periods 6-10)Graphs by Leader Type

43

Figure A.3A. Effect of leader’s incentive use on workers’ performance reports

Figure A.3B. Effect of leader’s statement use on workers’ performance reports

12

34

56

Avg.

wor

ker r

epor

ted

perfo

rman

ce

-1 -.5 0 .5 1Overall leader incentive use encouraging high reports

Leader Incentives Only Leader

12

34

56

Avg.

wor

ker r

epor

ted

perfo

rman

ce

-1 -.5 0 .5 1Avg. leader statement use encouraging high reports

Leader Statements Only Leader

44

Figure A.4A: Leader’s dishonesty and incentive use

Figure A.4B. Leader’s dishonesty and incentive use across periods

0.1

.2.3

.4In

cent

ive u

se e

ncou

ragi

ng h

igh

repo

rts

Other leader Dishonest leaderBars indicate standard errors

-.4-.2

0.2

.4.6

Ince

ntive

use

enc

oura

ging

dish

ones

ty

0 2 4 6 8 10Period

Dishonest leader Other leaderBars indicates standard errors

45

Figure A.4C: Leader’s dishonesty and statement use

Figure A.4D. Leaders’ dishonesty and statement use across periods

0.1

.2.3

.4St

atem

ent u

se e

ncou

ragi

ng h

igh

repo

rts

Other leader Dishonest leader

Bars indicate standard errors

0.2

.4.6

Stat

emen

t use

enc

oura

ging

dish

ones

ty

0 2 4 6 8 10Period

Dishonest leader Other leaderBars indicate standard errors

46

Figure A.5: Effect of leadership on social appropriateness of misreporting

Notes: Mean responses to question, “Suppose that, in Stage 1, someone rolled a 1 on his or her die. How appropriate would it have been for that participant to report a higher number than 1 in Stage 1 of the experiment, when earnings were based on individual performance level and on the individual performance level of other participants?” and selected from four possible responses: “very socially inappropriate” (1), “somewhat socially inappropriate” (2), “somewhat socially appropriate” (3), and “very socially appropriate” (4). Subjects were incentivized to provide most common response provided by other subjects, following Krupka and Weber (2013). We also asked three additional questions, with a focus on norms of behavior in Stage 2 and perceptions of specific group members (see instructions in Appendix E). Pair-wise Mann-Whitney tests reveal significant differences between the Inactive Leader and any other condition: Inactive Leader vs. Leader Statements Only (p=0.0003), Inactive Leader vs. Leader Incentives Only (p=0.013), and Inactive Leader vs. Leader (p=0.007).

12

34

soci

al a

ppro

pria

tene

ss

Inactive Leader Statements Incentives Leadercondition

mean 95% c.i.

47

Figure A.6: Distribution of reported performance in Stage 2 by Leader Type (Experiment 2)

0.2

.4.6

0 2 4 6 0 2 4 6 0 2 4 6

Inactive leader Active leader (other) Active leader (dishonest)

Den

sity

Workers' reported performance (Stage 2, all periods)Graphs by Leader Type

48

Figure A.7A. Effect of leader’s incentive use on workers’ performance reports (Experiment 2)

Figure A.7B. Effect of leader’s statement use on workers’ performance reports

(Experiment 2)

-10

12

3Av

g. w

orke

r rep

orte

d pe

rform

ance

cha

nge

-1 -.5 0 .5 1Avg. leader incentive use encourating dishonesty

-10

12

3Av

g. w

orke

r rep

orte

d pe

rform

ance

cha

nge

-.2 0 .2 .4 .6Avg. leader statement use encourating dishonesty

49

Figure A.8A: Leader’s dishonesty and incentive use (Experiment 2)

Figure A.8B: Leader’s dishonesty and statement use (Experiment 2)

0.2

.4.6

Ince

ntive

use

enc

oura

ging

hig

h re

ports

Other leader Dishonest leaderBars indicate standard errors

0.1

.2.3

Mes

sage

use

enc

oura

ging

hig

h re

ports

Other leader Dishonest leader

Bars indicate standard errors

50

B Appendix Tables (Additional Data and Analyses) Note: Robustness of tobit models to linear and ordered profit specifications are all provided in Appendix C.

Table B.1: Summary statistics (Experiment 1)

Variable All Inactive Leader

Leader Statements

Only

Leader Incentives

Only Leader

Age 20.51 (2.96)

21 (3.51)

19.72 (1.99)

20.49 (3.28)

20.65 (2.54)

Male (%) 56.15 (.50)

50 (.50)

57.14 (.49)

60 (.49)

49.75 (.50)

Economics student (%)

40.62 (.49)

42.5 (.50)

31.25 (.47)

48.75 (.50)

40 (.49)

Disposable monthly income (GBP)

251 (175)

265 (171)

257 (178)

283 (204)

200 (131)

Big five conscientiousness

50 (9.29)

51.67 (8.83)

49.18 (10.23)

48.67 (9.67)

50.48 (7.99)

Table B.2: Differences with expected distribution in Stage 1

Reported performance (in percent) Condition N 1 2 3 4 5 6 mean

Inactive Leader 80 3.8%*** 11.3% 10.0%* 20.0% 18.8% 36.3%††† 4.48

Leader 80 1.3%*** 8.8%** 21.3% 18.8% 8.8%** 41.3%††† 4.49

Statements 80 1.3%*** 8.8%** 12.5% 30.0%††† 17.5% 30.0%††† 4.44

Incentives 80 2.5%*** 10.0%* 13.8% 16.3% 15.0% 42.5%††† 4.59

Total 320 2.2%*** 9.7%*** 14.4% 21.3%†† 15.0% 37.5%††† 4.50

Stars (crosses) refer to significance levels of one-sided binomial probability test that the observed frequency is smaller (larger) than the expected frequency of 16.7%. *(†) 10%-level, **(††) 5%-level, and *** (†††) 1%-level.

51

Table B.3: Individual characteristics and performance in Stage 1

Dependent variable Stage 1 performance

(1)

Economics student

1.358*** (0.284)

Male

0.658* (0.373)

Age

0.042** (0.019)

Big Five conscientiousness score

-0.147*** (0.042)

Constant

5.027*** (1.449)

Number of Obs 225

Log-likelihood -360.738

Notes: Random effects tobit regression. Bootstrapped standard errors, clustered at the session level, in parentheses. * significant at 10%; ** significant at 5%; *** significant at 1%. The number of observations is 300 instead of 320 (full sample), since we could not elicit the Big Five scores in one session because of technical problems.

 

52

Table B.4: Message categories' definitions and use

Message type Definition Frequency

Request high Supervisor requests an individual worker, several workers, or the entire group to report a high number, i.e., a number that is higher than 3.5 on average.

.57

Request low Supervisor requests an individual worker, several workers, or the entire group to report a low number, i.e., a number that is lower than or equal to 3.5 on average.

.17

Praise high Supervisor praises an individual, several individuals, or the entire group for reporting a high number, i.e., a number that is higher than 3.5 on average.

.16

Praise low Supervisor praises an individual, several individuals, or the entire group for reporting a low number, i.e., a number that is lower than or equal to 3.5 on average.

.05

Reward high Supervisor refers to how the bonus will be distributed in a way that gives workers a financial incentive to report a high number, i.e., a number that is higher than 3.5 on average.

.19

Reward low Supervisor refers to how the bonus will be distributed in a way that gives workers a financial incentive to report a low number, i.e., a number that is lower than or equal to 3.5 on average.

.03

Dishonest Supervisor refers to dishonest behavior in the message. .39

Honest Supervisor refers to honest behavior in the message. .12

Ref. other groups Supervisor refers to other groups, or other groups’ behavior in a previous period. .42

Ref. prize Supervisor refers to the size of the prize in a previous period. .24

Ref. earnings Supervisor refers to the group’s earnings in a previous period. .31

Humor Supervisor makes a joke or the message is ironic or humorous. .13

Apology Supervisor apologizes for his or her messages or bonus distribution in a previous period. .04

Encouragement Supervisor sends a message that includes a form of general encouragement. .36

Miscellaneous Supervisor sends a message that does not belong in any of the other categories. .07

No message Supervisor sends no message. .07

Note: Frequency is defined as number of messages per group and period divided by the total number of messages, which is 400 messages for the Leader and Leader Statements Only conditions. The bonus categories refer only to observations of the Leader condition with a total number of 200 messages. The messages were classified by three independent coders, paid students, according to the provided categories.

53

Table B.5: Relationships between secondary statement categories and workers’ reports

Dependent variable Reported Stage 2 performance (1) (2) (3) (4) (5) (6) Total no. of characters sent by leader during period

0.001 (0.001)

Ref. other groups 0.408* (0.083)

Ref. prize -0.147 (0.221)

Ref. earnings 0.221 (0.235)

Humor 0.878** (0.356)

Encouragement -0.021 (0.208)

Stage 1 report by worker 0.427*** (0.084)

0.424***

(0.083) 0.430***

(0.083) 0.427***

(0.084) 0.440***

(0.085) 0.426***

(0.083)

Constant 3.763*** (0.464)

3.678*** (0.410)

3.860*** (0.427)

3.766*** (0.433)

3.673*** (0.424)

3.850*** (0.413)

Number of Obs 1200 1200 1200 1200 1200 1200 Obs. left censored (≤ 1) 59 59 59 59 59 59 Obs. right censored (≥ 6) 565 565 565 565 565 565 Number of workers, firms 120, 40 120, 40 120, 40 120, 40 120, 40 120, 40 Log likelihood -1793.00 -1790.25 -1792.84 -1792.40 -1787.69 -1793.14

Wald χ2(2)=26.0 (p<0.001)

χ2(2)=30.9 (p<0.001)

χ2(2)=28.5 (p<0.001)

χ2(2)=25.9 (p<0.001)

χ2(2)=34.3 (p<0.001)

χ2(2)=23.3(p<0.001)

Notes: Data from all periods of both conditions in which leaders send statements (Leader, Leader Statements Only). All models report panel tobit regressions that use a worker’s performance report in a period as an observation and include worker random effects, with bootstrapped standard errors clustered at the firm level. * p < 0.1; ** p < 0.05; *** p < 0.01.

 

54

Table B.6: Relationships between primary statement categories and workers’ reports

Dependent variable Reported Stage 2 performance (1) (2) (3) (4) (5)

Request high 1.026*** (0.270) 0.786***

(0.260)

Request low -1.412*** (0.419) -1.315***

(0.389)

Praise high 0.285 (0.277) 0.223

(0.253)

Praise low -0.586 (0.434) -0.294

(0.399)

Dishonest 0.456*** (0.176) 0.135

(0.195)

Honest -0.866*** (0.306) -0.697**

(0.301)

Reward high 0.931* (0.540)

0.522 (0.386)

Reward low -2.829** (1.219)

-1.749** (0.870)

Stage 1 report by worker 0.414*** (0.074)

0.430***

(0.083) 0.412***

(0.084) 0.430***

(0.083) 0.411***

(0.073)

Constant 3.523*** (0.379)

3.769*** (0.409)

3.831*** (0.410)

3.769*** (0.409)

3.642*** (0.375)

Number of Obs 1200 1200 1200 1200 1200 Obs. left censored (≤ 1) 59 59 59 59 59 Obs. right censored (≥ 6) 565 565 565 565 565 Number of workers, firms 120, 40 120, 40 120, 40 120, 40 120, 40 Log likelihood -1749.66 -1780.41 -1783.40 -1780.41 -1739.42

Wald χ2(2)=47.5 (p<0.001)

χ2(2)=30.6 (p<0.001)

χ2(2)=35.1 (p<0.001)

χ2(2)=30.6 (p<0.001)

χ2(9)=57.0 (p<0.001)

Notes: Data from all periods of both conditions in which leaders send statements (Leader, Leader Statements Only). All models report panel tobit regressions that use a worker’s performance report in a period as an observation and include worker random effects, with bootstrapped standard errors clustered at the firm level. Results are very similar if models 4 and 5 are estimated using only data from Leader condition, in which Reward high and Reward low are relevant. * p < 0.1; ** p < 0.05; *** p < 0.01.

55

Table B.7: Leaders’ incentive use

(dropping periods with zero performance or bonus share variation)

Dependent variable: Incentive use in period

(𝜌!"! ) (1) (2)

Dishonest leader 0.226 (0.163)

0.723*** (0.208)

Period 0.044** (0.017)

Dishonest leader * period -0.097*** (0.033)

Constant 0.043 (0.097)

-0.186 (0.122)

Number of observations 313 313 Number of firms 39 39 R2 0.023 0.053

Wald χ2(1) = 1.91 (p = 0.167)

χ2(3) = 14.1 (p = 0.002)

Notes: Data from both conditions in which leaders allocate bonuses (Leader, Leader Incentives Only). Incentive use in period refers to the correlation between performance shares and bonus shares in a period. Both models report panel linear regressions that use a leader’s action in a period as an observation and include firm random effects, with bootstrapped standard errors clustered at the firm level. * p < 0.1; ** p < 0.05; *** p < 0.01.

56

Table B.8: Reported performance by condition

Dependent variable: Reported Stage 2 performance Probit Linear Ordered probit (1) (2) (3) (4) (5) (6)

Statements 0.759*** (0.255)

0.021 (0.327)

0.497*** (0.160)

0.177 (0.204)

0.388*** (0.151)

0.035 (0.175)

Incentives -0.100 (0.282)

-0.303 (0.347)

0.006 (0.182)

-0.172 (0.217)

-0.045 (0.168)

-0.162 (0.185)

Statements * Incentives

0.361 (0.390)

0.299 (0.573)

0.101 (0.243)

0.105 (0.335)

0.162 (0.236)

0.132 (0.293)

Period 0.151*** (0.024)

0.063** (0.026)

0.082*** (0.013)

0.037** (0.014)

0.076*** (0.012)

0.032** (0.011)

Period * Statements 0.138*** (0.048) 0.058**

(0.025) 0.066*** (0.023)

Period * Incentives 0.038 (0.054) 0.032

(0.032) 0.022 (0.027)

Period * Statements * Incentives 0.014

(0.095) -0.001 (0.048) 0.007

(0.045) Stage 1 report by worker s

0.362***

(0.057) 0.362***

(0.057) 0.180***

(0.036) 0.180***

(0.036) 0.175***

(0.034) 0.176***

(0.035)

Constant 2.359*** (0.322)

2.830*** (0.318)

3.000*** (0.215)

3.248*** (0.223)

Number of observations 2400 2400 2400 2400 2400 2400

Obs. left cens. (≤ 1) 142 142 Obs. right cens. (≥ 6) 952 952 No. of workers, firms 240, 80 240, 80 240, 80 240, 80 240, 80 240, 80 Log Likelihood / R2 -3763.9 -3753.2 0.081 0.085 -3508.0 -3497.6 Wald χ2(5)=142.2

(p < 0.001) χ2(8)=140.6 (p < 0.001)

χ2(5)=137.1 (p < 0.001)

χ2(8)=148.3 (p < 0.001)

χ2(5)=116.2 (p < 0.001)

χ2(8)=123.8 (p < 0.001)

Notes: Data from all periods and all conditions. Statements refers to conditions in which leaders sent statements at the beginning of each period. Incentives refers to conditions in which leaders distributed bonuses at the end of each period. Models 1 and 2 report tobit regressions with bootstrapped standard errors clustered at the firm level. Models 3 and 4 report linear regressions with standard errors clustered at the firm level. Models 5 and 6 report ordered probit regressions with standard errors clusterd at the firm level. All models employ a panel structure with worker random effects. * p < 0.1; ** p < 0.05; *** p < 0.01.

57

Table B.9: Reported performance and leader’s incentive use (Exp. 2)

Dependent variable Reported Stage 2 performance (1) (2) (3) (4) Bonus / perf. shares corr. up to previous period (𝜌!"!!)

0.920***

(0.292) 0.943*** (0.273)

0.821*** (0.395)

1.058*** (0.258)

Dishonest leader * 𝜌!"!! 0.306 (0.587)

Dishonest worker * 𝜌!"!! -0.442 (0.884)

Stage 1 report by worker 0.700*** (0.169)

0.700*** (0.172)

0.695*** (0.170)

0.733*** (0.170)

Previous period performance report 0.032

(0.168)

Previous period bonus share 0.005 (0.017)

Previous period performance report * bonus share -0.001

(0.004)

Constant 3.608*** (0.636)

3.427*** (0.897)

3.612*** (0.636)

3.449*** (0.628)

Number of observations 810 810 810 810 Obs. left censored (≤ 1) 33 33 33 33 Obs. right censored (≥ 6) 483 483 483 483 Number of workers, firms 90, 30 90, 30 90, 30 90, 30 Log Likelihood -964.28 -964.23 -964.15 -964.05

Wald χ2(2) = 26.0 (p < 0.001)

χ2(5) = 28.8 (p < 0.001)

χ2(3) = 29.8 (p < 0.001)

χ2(3) = 41.1 (p < 0.001)

Notes: Data from periods 2 through 10 Leader condition. Bonus / performance shares correlation up to previous period (𝜌!"!!) is the overall correlation between performance shares and bonus shares demonstrated by the leader up through the prior period. All models report panel tobit regressions that use a worker’s performance report in a period as an observation and include worker random effects, with bootstrapped standard errors clustered at the firm level. * p < 0.1; ** p < 0.05; *** p < 0.01.

58

Table B.10: Reported performance and leader’s statement use (Experiment 2)

Dependent variable Reported Stage 2 performance (1) (2) (3) (4) Leader’s encouragement of high performance reports (𝜔!")

1.270

(0.976) 1.612

(1.025) 1.963

(1.336) 0.955

(1.222)

Dishonest leader * 𝜔!" -1.724 (1.809)

Dishonest worker * 𝜔!" 1.172 (1.568)

Stage 1 report by worker 0.699*** (0.152)

0.667*** (0.171)

0.701*** (0.152)

0.679*** (0.159)

Previous period performance report 0.022

(0.088)

Constant 3.608*** (0.575)

3.826*** (0.756)

3.625*** (0.569)

3.696*** (0.621)

Number of observations 900 810 900 900 Obs. left censored (≤ 1) 37 33 37 37 Obs. right censored (≥ 6) 523 483 523 523 Number of workers, firms 90, 30 90, 30 90, 30 90, 30 Log Likelihood -1101.36 -965.32 -1100.24 -1100.95

Wald χ2(2) = 27.8 (p < 0.001)

χ2(3) = 19.0 (p < 0.001)

χ2(3) = 25.4 (p < 0.001)

χ2(3) = 30.9 (p < 0.001)

Notes: Data from Leader condition. Leader’s encouragement of high performance reports indicates the relative use of statement categories encouraging high (1) versus low (-1) performance reports. Models 1 and 3 use all periods of data; model 2 uses periods 2 through 10. All models report panel tobit regressions that use a worker’s performance report in a period as an observation and include worker random effects, with bootstrapped standard errors clustered at the firm level. * p < 0.1; ** p < 0.05; *** p < 0.01.

59

Table B.11. Estimates for worker utility model (Experiment 2)

Type 1 Type 2

Type probability 0.636*** (0.020)

0.364*** (0.020)

𝜆 - Own revenue 0.149*** (0.010)

1.677*** (0.039)

𝛼 – (Dis)utility from lying 0.007 (0.031)

-3.332*** (0.115)

𝛾 – Weight on leader incentive use 0.012 (0.015)

0.007 (0.011)

𝛽 – Weight on leader statements 0.429*** (0.027)

0.791*** (0.055)

Number of observations 8,100 Number of workers 135 Log likelihood -1,642.42 Note: Parameter estimates of model in Equation 7 using data from all periods in Leader condition. An observation is a worker’s decision of whether to make a particular performance report in a period. Standard errors in parentheses. * p < 0.1; ** p < 0.05; *** p < 0.01.

60

Table B.12: Leaders’ strategies by leader type (Experiment 2)

Dependent variable: Incentive use in period

(𝜌!"! ) Statement use in period

(𝜔!") (1) (2) (3) (4)

Dishonest leader 0.201 (0.123)

0.194 (0.202)

0.094** (0.048)

0.163** (0.067)

Period -0.023*

(0.014) -0.007 (0.006)

Dishonest leader * period 0.001 (0.038) -0.013

(0.009)

Constant 0.177*** (0.074)

0.302*** (0.096)

0.063** (0.026)

0.104*** (0.031)

Number of observations 300 300 300 300 Number of firms 30 30 30 30 R2 0.030 0.044 0.050 0.087

Wald χ2(1) = 2.65 (p = 0.103)

χ2(3) = 6.26 (p = 0.100)

χ2(1) = 3.88 (p = 0.049)

χ2(3) = 11.7 (p < 0.001)

Notes: All models include data from all 10 periods of Leader condition. Incentive use in period refers to the correlation between performance shares and bonus shares in a period. Statement use in period is the tendency to use message categories encouraging high (1) or low (-1) performance reports. All models report panel linear regressions that use a leader’s action in a period as an observation and include firm random effects, with standard errors clustered at the firm level. * p < 0.1; ** p < 0.05; *** p < 0.01.

61

C Robustness of Tobit Specifications under Linear and Ordered Probit Analyses

Table C.3a: Leader honesty and reported performance (linear regression)

Dependent variable Reported Stage 2 performance (1) (2) (3) (4)

Unit of analysis: Worker in a period Worker mean

across periods

Group mean across periods

Active leader 0.263* (0.153)

0.006 (0.190)

0.263* (0.154)

0.263* (0.156)

Dishonest (active) leader 0.333**

(0.168) 0.104

(0.241) 0.333* (0.169)

0.308* (0.166)

Period 0.082*** (0.013)

0.037** (0.014)

Active leader * Period 0.047** (0.022)

Dishonest leader * Period 0.042 (0.036)

(Mean) Stage 1 report by worker(s)

0.178*** (0.037)

0.178*** (0.037)

0.178*** (0.037)

0.101 (0.061)

Constant 3.007*** (0.221)

3.254*** (0.226)

3.458*** (0.217)

3.814*** (0.323)

Number of observations 2400 2400 240 80 Number of workers, firms 240, 80 240, 80 240, 80 - , 80 R2 0.068 0.071 0.126 0.125

Wald / F χ2(4) = 83.4 (p < 0.001)

χ2(6) = 88.9 (p < 0.001)

F3,79 = 10.7 (p < 0.001)

F3,76 = 3.49 (p = 0.02)

Notes: Active leader equals 1 in all conditions with an active leader; active (dishonest) leader equals 1 when the group has an active leader and this leader reported a 6 in Stage 1. All models report linear GLS or OLS regressions. Models 1 and 2 are panel specifications that use a worker’s performance report in a period as an observation and include worker random effects, with standard errors clustered at the firm level. Model 3 uses the average report by a worker across all periods, with standard errors clustered at the firm level. Model 4 uses the average report within a firm across all periods, with robust standard errors. * p < 0.1; ** p < 0.05; *** p < 0.01.

62

Table C.3b: Leader honesty and reported performance (ordered probit)

Dependent variable Reported Stage 2 performance (1) (2) (3) (4)

Unit of analysis: Worker in a period Worker mean

across periods

Group mean across periods

Active leader 0.152 (0.141)

-0.079 (0.163)

0.248 (0.178)

0.488* (0.278)

Dishonest (active) leader 0.414**

(0.177) 0.118

(0.218) 0.413** (0.202)

0.559* (0.290)

Period 0.076*** (0.012)

0.032*** (0.011)

Active leader * Period 0.043** (0.020)

Dishonest leader * Period 0.058 (0.037)

(Mean) Stage 1 report by worker(s)

0.177*** (0.036)

0.178*** (0.036)

0.212*** (0.043)

0.191* (0.108)

Number of observations 2400 2400 240 80 Number of workers, firms 240, 80 240, 80 240, 80 - , 80 Log pseudolikelihood -3510.85 -3401.82 -816.78 -310.88

Wald χ2(4) = 74.8 (p < 0.001)

χ2(6) = 76.9 (p < 0.001)

χ2(3) = 32.2 (p < 0.001)

χ2(3) = 11.1 (p = 0.01)

Notes: Active leader equals 1 in all conditions with an active leader; active (dishonest) leader equals 1 when the group has an active leader and this leader reported a 6 in Stage 1. All models report ordered probit regressions. Models 1 and 2 are panel specifications that use a worker’s performance report in a period as an observation and include worker random effects, with standard errors clustered at the firm level. Model 3 uses the average report by a worker across all periods, with standard errors clustered at the firm level. Model 4 uses the average report within a firm across all periods, with robust standard errors (note that in this specification there are 58 categories for 80 observations; it is reported only for completeness). Robust standard errors, clustered at the group level in Models 1, 2 and 3, in parentheses. * p < 0.1; ** p < 0.05; *** p < 0.01.

63

Table C.4a: Reported performance and leader’s incentive use (linear regression)

Dependent variable Reported Stage 2 performance (1) (2) (3) (4) (5) Bonus / perf. shares corr. up to previous period (𝜌!"!!)

0.847***

(0.097) 0.573*** (0.098)

0.877*** (0.130)

0.855*** (0.119)

0.779*** (0.143)

Communication condition * 𝜌!"!!

-0.083 (0.199)

Dishonest leader * 𝜌!"!! -0.018 (0.216)

Dishonest worker * 𝜌!"!! 0.149 (0.257)

Stage 1 report by worker 0.206*** (0.050)

0.139*** (0.037)

0.207*** (0.050)

0.206*** (0.050)

0.202*** (0.051)

Previous period performance report 0.279***

(0.087)

Previous period bonus share -0.042 (0.920)

Previous period performance report * bonus share 0.081

(0.209)

Constant 3.598*** (0.267)

2.565*** (0.385)

3.597*** (0.270)

3.602*** (0.268)

3.628*** (0.271)

Number of observations 1080 1080 1080 1080 1080 Number of workers, firms 120, 40 120, 40 120, 40 120, 40 120, 40 R2 0.115 0.205 0.115 0.115 0.116

Wald / F χ2(2)=107.5 (p < 0.001)

χ2(5)=278.3 (p < 0.001)

χ2(3)=108.7 (p < 0.001)

χ2(3)=107.5 (p < 0.001)

χ2(3)=102.7 (p < 0.001)

Notes: Data from periods 2 through 10 of both conditions in which leaders allocate bonuses (Leader, Leader Incentives Only). Bonus / performance shares correlation up to previous period (𝜌!"!!) is the overall correlation between performance shares and bonus shares demonstrated by the leader up through the prior period. All models report panel linear regressions that use a worker’s performance report in a period as an observation and include worker random effects, with standard errors clustered at the firm level. * p < 0.1; ** p < 0.05; *** p < 0.01.

64

Table C.4b: Reported performance and leader’s incentive use (ordered probit)

Dependent variable Reported Stage 2 performance (1) (2) (3) (4) (5) Bonus / perf. shares corr. up to previous period (𝜌!"!!)

0.727***

(0.087) 0.611*** (0.095)

0.727*** (0.115)

0.694*** (0.104)

0.594*** (0.114)

Communication condition * 𝜌!"!!

-0.002 (0.169)

Dishonest leader * 𝜌!"!! 0.093 (0.188)

Dishonest worker * 𝜌!"!! 0.303 (0.228)

Stage 1 report by worker 0.192*** (0.047)

0.160*** (0.042)

0.192*** (0.047)

0.195*** (0.048)

0.187*** (0.046)

Previous period performance report 0.119*

(0.070)

Previous period bonus share -0.330 (0.654)

Previous period performance report * bonus share 0.081

(0.174)

Number of observations 1080 1080 1080 1080 1080 Number of workers, firms 120, 40 120, 40 120, 40 120, 40 120, 40 Log pseudolikelihood -1552.06 -1539.07 -1552.06 -1551.94 -1550.69

Wald χ2(2) = 84.7 (p < 0.001)

χ2(5)=160.8 (p < 0.001)

χ2(3) = 88.3 (p < 0.001)

χ2(3) = 86.3 (p < 0.001)

χ2(3) = 84.7 (p < 0.001)

Notes: Data from periods 2 through 10 of both conditions in which leaders allocate bonuses (Leader, Leader Incentives Only). Bonus / performance shares correlation up to previous period (𝜌!"!!) is the overall correlation between performance shares and bonus shares demonstrated by the leader up through the prior period. All models report panel ordered probit regressions that use a worker’s performance report in a period as an observation and include worker random effects, with standard errors clustered at the firm level. * p < 0.1; ** p < 0.05; *** p < 0.01.

65

Table C.5a: Reported performance and leader’s statement use (linear regression)

Dependent variable Reported Stage 2 performance (1) (2) (3) (4) (5) Leader’s encouragement of high performance reports (𝜔!")

1.133***

(0.212) 0.971*** (0.172)

0.807*** (0.231)

0.784*** (0.204)

1.195*** (0.239)

Incentive condition * 𝜔!" 0.782** (0.332)

Dishonest leader * 𝜔!" 0.929*** (0.314)

Dishonest worker * 𝜔!" -0.163 (0.404)

Stage 1 report by worker 0.191*** (0.049)

0.115*** (0.040)

0.187*** (0.047)

0.199*** (0.048)

0.201*** (0.060)

Previous period performance report 0.281***

(0.043)

Constant 3.671*** (0.246)

2.768*** (0.195)

3.677*** (0.227)

3.613*** (0.239)

3.622*** (0.280)

Number of observations 1200 1080 1200 1200 1200 Number of workers, firms 120, 40 120, 40 120, 40 120, 40 120, 40 R2 0.106 0.180 0.117 0.115 0.108

Wald χ2(2) = 46.0 (p < 0.001)

χ2(3)=152.6 (p < 0.001)

χ2(3) = 54.6 (p < 0.001)

χ2(3) = 71.0 (p < 0.001)

χ2(3) = 46.3 (p < 0.001)

Notes: Data from both conditions in which leaders send statements (Leader, Leader Statements Only). Leader’s encouragement of high performance reports indicates the relative use of statement categories encouraging high (1) versus low (-1) performance reports. Models 1, 3 and 4 use all periods of data; model 2 uses periods 2 through 10. All models report linear regressions that use a worker’s performance report in a period as an observation and include worker random effects, with standard errors clustered at the firm level. * p < 0.1; ** p < 0.05; *** p < 0.01.

66

Table C.5b: Reported performance and leader’s statement use (ordered probit)

Dependent variable

Reported Stage 2

performance

(1) (2) (3) (5) (5) Leader’s encouragement of high performance reports (𝜔!")

1.032***

(0.201) 1.009*** (0.194)

0.714*** (0.216)

0.662** (0.186)

0.995*** (0.223)

Incentive condition * 𝜔!! 0.812** (0.380)

Dishonest leader * 𝜔!" 1.127*** (0.388)

Dishonest worker * 𝜔!" 0.107 (0.409)

Stage 1 report by worker 0.192*** (0.043)

0.156*** (0.041)

0.189*** (0.041)

0.206*** (0.042)

0.185*** (0.050)

Previous period performance report 0.114**

(0.048)

Number of observations 1200 1080 1200 1200 1200 Number of workers, firms 120, 40 120, 40 120, 40 120, 40 120, 40 Log pseduolikelihood -1622.60 -1410.97 -1615.76 -1610.47 -1622.50

Wald χ2(2) = 48.8 (p < 0.001)

χ2(3) = 65.1 (p < 0.001)

χ2(3) = 63.5 (p < 0.001)

χ2(3) = 77.4 (p < 0.001)

χ2(3) = 48.8 (p < 0.001)

Notes: Data from both conditions in which leaders send statements (Leader, Leader Statements Only). Leader’s encouragement of high performance reports indicates the relative use of statement categories encouraging high (1) versus low (-1) performance reports. Models 1, 3 and 4 use all periods of data; model 2 uses periods 2 through 10. All models report ordered probit regressions that use a worker’s performance report in a period as an observation and include worker random effects, with standard errors clustered at the firm level. * p < 0.1; ** p < 0.05; *** p < 0.01.

67

Table C.7a: Leader honesty and reported performance (Exp. 2, linear regression)

Dependent variable Reported Stage 2 performance (1) (2) (3) (4)

Unit of analysis: Worker in a period Worker mean

across periods

Group mean across periods

Active leader 0.186 (0.192)

0.371* (0.218)

0.186 (0.194)

0.197 (0.191)

Dishonest (active) leader 0.175

(0.248) 0.080

(0.249) 0.175

(0.251) 0.164

(0.256)

Period 0.012 (0.010)

0.030** (0.014)

Active leader * Period -0.034 (0.021)

Dishonest leader * Period 0.017 (0.025)

(Mean) Stage 1 report by worker(s)

0.195*** (0.049)

0.195*** (0.049)

0.195*** (0.049)

0.232** (0.109)

Constant 3.843*** (0.221)

3.743*** (0.224)

3.909*** (0.225)

3.737*** (0.479)

Number of observations 1350 1350 135 45 Number of workers, firms 135, 45 135, 45 135, 45 - , 45 R2 0.052 0.053 0.116 0.141

Wald / F χ2(4) = 28.6 (p < 0.001)

χ2(6) = 35.0 (p < 0.001)

F3,44 = 8.53 (p < 0.001)

F3,41 = 3.65 (p = 0.02)

Notes: Active leader equals 1 in Leader condition; active (dishonest) leader equals 1 when the group has an active leader and this leader reported a 6 in Stage 1. All models report linear GLS or OLS regressions. Models 1 and 2 are panel specifications that use a worker’s performance report in a period as an observation and include worker random effects, with standard errors clustered at the firm level. Model 3 uses the average report by a worker across all periods, with standard errors clustered at the firm level. Model 4 uses the average report within a firm across all periods, with robust standard errors. * p < 0.1; ** p < 0.05; *** p < 0.01.

68

Table C.7b: Leader honesty and reported performance (Exp. 2, ordered probit)

Dependent variable Reported Stage 2 performance (1) (2) (3) (4)

Unit of analysis: Worker in a period Worker mean

across periods

Group mean across periods

Active leader 0.244 (0.248)

0.381 (0.243)

0.241 (0.216)

0.357 (0.289)

Dishonest (active) leader 0.395

(0.339) 0.266

(0.319) 0.288

(0.310) 0.226

(0.456)

Period 0.027** (0.011)

0.039*** (0.013)

Active leader * Period -0.026 (0.021)

Dishonest leader * Period 0.024 (0.034)

(Mean) Stage 1 report by worker(s)

0.283*** (0.065)

0.284*** (0.065)

0.241*** (0.055)

0.364** (0.167)

Number of observations 1350 1350 135 45 Number of workers, firms 135, 45 135, 45 135, 45 - , 45 Log pseudolikelihood -1602.7 -1602.1 -405.5 -152.5

Wald χ2(4) = 28.3 (p < 0.001)

χ2(6) = 33.4 (p < 0.001)

χ2(3) = 27.6 (p < 0.001)

χ2(3) = 9.97 (p = 0.02)

Notes: Active leader equals 1 in Leader condition; active (dishonest) leader equals 1 when the group has an active leader and this leader reported a 6 in Stage 1. All models report ordered probit regressions. Models 1 and 2 are panel specifications that use a worker’s performance report in a period as an observation and include worker random effects, with standard errors clustered at the firm level. Model 3 uses the average report by a worker across all periods, with standard errors clustered at the firm level. Model 4 uses the average report within a firm across all periods, with robust standard errors (note that in this specification there are 58 categories for 80 observations; it is reported only for completeness). Robust standard errors, clustered at the group level in Models 1, 2 and 3, in parentheses. * p < 0.1; ** p < 0.05; *** p < 0.01.

69

D Mathematical Appendix

Here, we note the incentive properties of the Stage 2 task in Experiment 1. The incentives for misreporting in Experiment 2 are stronger (as the negative externality from doing so does not affect a worker’s own firm).

First stage (individual task)

As in Section 2, we denote the player’s reported performance level with 𝑟! and the true outcome of the die roll with 𝑝!. Then, a player’s profit is given by:

𝜋!(𝑟! , 𝑟!!) = 𝑠!(𝑟! , 𝑟!!)𝑉(𝑟! , 𝑟!!)

with 𝑠!(𝑟! , 𝑟!!) =!!

!!! !!!!!!!!,!!!

,

and 𝑉(𝑟! , 𝑟!!) =𝑎,                                                                                      if   !!

!!!!!

≤ 𝜇  ,

𝑎 − 𝑏!!! !!!!!

!!!,!!!

!− 𝜇 ,  if   !!!

!!!!

> 𝜇  .

The incentive for each player i to report a higher performance level than the one drawn from the random distribution is denoted by Δ = 𝜋!(𝑟! + 1)− 𝜋!(𝑟!) and depends on the other players' reported performance levels. Therefore we have to distinguish between the following cases: (i) Assume all other players' reported performance levels are such that

!!!!!!!

≤ 𝜇 holds, then reporting a higher performance can either (a) increase the reported average performance such that it is higher than the expected performance level 𝜇 or (b) the average reported performance remains lower than the expected performance. (ii) Assume all

players' reported performance levels are such that !!!!!!!

> 𝜇 holds, independent of 𝑝!.

We can show that for the parameter choice of a=1250, b=300, for the distribution of 𝑟! ∈ {1,… , 6}, and for N=20 players, the incentive to deviate is positive in all cases. Hence, we have no interior equilibrium and all players will report the highest possible performance level of 𝑟! = 6.

Second stage (group task)

Each of the firms 𝑓 ∈ {1,… ,𝐹}, consists of n players 𝑖 ∈ {1,… ,𝑛}, with 𝑛 − 1 of them being workers and one being the leader. As in stage 1, each worker randomly draws a performance level, which is private information to each worker. The leaders are inactive players. Each worker is then asked to report her performance level 𝑟!. The firm’s performance

corresponds to the average performance reported by the three workers, 𝑟! =  !!

!!!!!!!!!

. Each

player's payoff is a share 𝑥!,! of the firm's payoff with 𝑥!,! ∈ [0.1,… , (1−!!)− 𝑛 − 2 0.1],

70

since a worker can receive any share between 0.1 and ((1− !!)− 𝑛 − 2 0.1) of the firm’s

payoff in the conditions, in which leaders have the power to give financial incentives. They can freely distribute a share of ((1− !

!)− 𝑛 − 1 0.1) of the firm’s payoff among the

workers. The leader receives in all conditions a share of !!- of the firm’s payoff and in the

Inactive Leader and the Leader Statements Only condition each worker receives !!-th as well.

A worker’s payoff is therefore:

𝜋!(𝑟! , 𝑟!!) = 𝑥!,!  𝑠! 𝑟! , 𝑟!! 𝑉(𝑟! , 𝑟!!)

with 𝑠!(𝑟! , 𝑟!!) =!!

!!! !!!!!!!!,!!!

,

and 𝑉(𝑟! , 𝑟!!) =𝑎,                                                                                              if  

!!!!!!

!≤ 𝜇  ,

𝑎 − 𝑏!!! !!!!!

!!!,!!!

!− 𝜇 ,  if  

!!!!!!

!> 𝜇  .

The incentive for each player i to report a higher performance level than the one drawn from the random distribution depends on the other players' reported performance levels. The incentive is denoted by Δ = 𝜋!(𝑟! +

!!!!

)− 𝜋!(𝑟!). Note that reporting a one unit

higher performance level results in an increase of !!!!

of the average reported performance level of this player's firm. Therefore, we have to distinguish between the following cases: (i)

Assume all firms' reported performance levels are such that !!!

!!!

!≤ 𝜇 holds, then reporting

a higher performance can either (a) increase the reported average performance that it is higher than the expected performance level 𝜇 or (b) the average reported performance remains lower than the expected performance. (ii) Assume all other firms’ average reported performance

levels are such that !!!

!!!

!> 𝜇 holds independent of the average reported performance level

of this group 𝑟!.

We can show that for our parameter choice of a=1250, b=300, for the boundaries of the distribution of 𝑟! ∈ {1,… , 6}, for F=5 groups, and n=4 players per group, the incentive to deviate is positive in all cases. Hence, we have no inner equilibrium and all players will report the highest possible performance level of 𝑟! = 6.

71

E Instructions (Experiment 1)  

Thank you for participating in today’s experiment.

I will read through a script to explain to you the nature of today’s experiment as well as how to navigate the computer interface with which you will be working. I will use this script to make sure that the information given in all sessions of this experiment is the same.

In addition to a GBP 2.50 payment that you receive for your participation, you will be paid money that you accumulate from the decision tasks that will be described to you in a moment. The exact amount you receive will be determined during the experiment and will depend on your decisions and the decisions of others. You will be paid privately, in cash, at the conclusion of the experiment.

All monetary amounts you will see in this experiment will be denominated in ECUs or Experimental Currency Units. We will convert ECUs into GBP at the rate of

35 ECUs = 1 GBP.

If you have any questions during the experiment, please raise your hand and wait for an experimenter to come to you.

Please do not talk, exclaim, or try to communicate with other participants during the experiment.

Do not use the computer in a way not specified by these instructions or by the experimenters.

Participants intentionally violating the rules may be asked to leave the experiment with only their participation payment.

All numerical examples in these instructions are used simply to provide examples and do not represent any hints or suggestions for how you should make your decisions during the experiment.

The experiment

The experiment consists of two stages. There are 20 participants in today’s experiment. You and all other 19 participants will take part in a decision task in each of these two stages What happens in each stage will not affect the procedures or your earnings in the other stage. You will receive separate instructions for each stage, once we reach that stage of the experiment.

INSTRUCTIONS ON STAGE 1

In the first stage of the experiment all participants will participate together in a single decision task. This task will only be performed once.

The task

72

In the task, each participant in the experiment will roll a die privately and individually to determine his or her “individual performance.” A participant’s performance represents how well the participant performed the task. In this task, higher scores represent better performance.

A 6-sided die and a cup were provided to each participant before the start of the experiment. Each of you will roll the die inside the cup privately to determine your individual performance. You should roll the die in a way that no other participant observes the outcome of your die roll. The experimenter will now demonstrate how you can roll the die so that only you can observe the outcome. Afterward, please try rolling the die several times, to practice this procedure and to verify that you received a fair die.

After rolling the die, each participant will report his or her individual performance on the computer screen. Note that there is no way for anyone, neither the experimenter nor other participants, to verify the actual individual performance (die roll) of any participant. Therefore, all payoffs will be based on the level of individual performance that is reported by each participant. Your payoff in the task will be determined by the performance you report and by the performance reported by others.

Payoffs

Each participant’s payoff for the task then depends on the performance level reported by that participant and on the performance level reported by all other participants. In particular, the reported individual performance levels of all participants determine the size of the total prize available for performing the task and each participant’s reported individual performance level determines that participant’s share of the prize. This share, multiplied by the size of the total prize, is then that participant’s payoff for the task.

Specifically, each participant’s payoff is calculated as follows:

1. Size of the prize: The size of the total prize depends on the average reported individual performance level of all participants, including you. The prize is largest, equal to 1250 ECU, when the average reported performance level is not higher than the expected average performance level from rolling a die, or 3.5. If everyone rolls their die and reports the observed performance level, one would expect an average score of 3.5 (this is the average of the numbers, 1, 2, 3, 4, 5 and 6). If the average reported performance level is 3.5 or lower, the total size of the prize will be fixed at 1250 ECU. But a higher average reported performance level will lower the size of the prize. The table below shows how the prize varies with the average reported performance level. The average reported performance level is calculated by adding the performance levels reported by all participants and then dividing this sum by the number of participants, which is equal to 20

73

Average reported performance level of all participants Size of the prize Prize in ECU

1 100% 1250

1.5 100% 1250

2 100% 1250

2.5 100% 1250

3 100% 1250

3.5 (expected) 100% 1250

4 88% 1100

4.5 76% 950

5 64% 800

5.5 52% 650

6 40% 500

The exact size of the prize is determined by the following formula:

That is, for every unit by which the average reported performance level is above 3.5, the size of the prize decreases by 24 percent.

If  the  average  reported  performance  level  is  less  than  or  equal  to  3.5  (expected  average  performance  level):    

Prize  =  1250  

If  the  average  reported  performance  level  is  greater  than  3.5  (expected  average  performance  level):  

Prize  =  1250  –  300  x  (average  reported  performance  level  –  3.5)  

74

2. Your share of the prize: Each participant’s share of the total prize is determined by the ratio of that participant’s reported performance level and the sum of reported performance levels of all participants.

Share  of  the  prize =Own  reported  performance  level

Sum  of  reported  performance  levels  of  all  participants

Thus, for any total size of the prize, your share of the prize is greater when your reported performance level is greater.

3. Your payoff: Each participant’s payoff is calculated by multiplying the size of the prize

by that participant’s share of the prize.

Your  payoff = (Size  of  the  prize)  ×  (Your  share  of  the  prize)

Are there any questions about how the size of the prize, shares of the prize, or payoffs will be determined in stage 1? Remember that if you have a question at any point during the experiment, you should raise your hand and wait for the experimenter to come to you.

Feedback

At the end of Stage 1, you will see a screen that shows you:

• Your reported performance level • The average reported performance among all participants • The size of the prize • Your share of the prize • Your payoff

Examples

We will now go through some examples to make sure that it is clear to everyone how payoffs are determined.

Example 1:

Suppose that your reported performance level is 5, the sum of all participants’ performance levels, including yours, is 100, and the average performance level of all 20 participants is 5.

In this case:

• The total prize is 800 ECU (i.e., 1250 ECU – 300 ECU (5 – 3.5) = 800 ECU). The prize is smaller than 1250 ECU because the average performance level of all participants (5) is greater than 3.5

75

• Your share of the prize is 5/100 = 0.05. • Your payoff for this task is then 800 ECU x 0.05 = 40 ECU.

Example 2:

Now suppose that your reported performance level is 3, the sum of all participants’ performance levels, including yours, is 70, and the average performance level of all participants is 3.5.

In this case:

• The total prize is 1250 ECU because the average performance level of all participants (3.5) is equal to 3.5.

• Your share of the prize is 3/70≅0.043. • Your payoff for this task is then 1250 ECU x 0.043 ≅ 53.8 ECU.

Note that in this and following examples we round numbers to one decimal place for calculations. However, the computer will calculate numbers exactly.

Practice questions

Before we begin with Stage 1, in which you will perform the above task one time, we will first ask you to answer some practice questions about payoffs. This is done to make sure that everyone understands how payoffs are calculated.

Please click “OK” on your screen now, to see the practice questions, and try to answer them. If you have a question or are confused, please raise your hand and wait for the experimenter to come to you.

As soon as everyone has answered the practice questions correctly, Stage 1 will start.

{PRACTICE QUESTION STAGE 1, on screen}

Please answer the questions below.

If you need help, the instructions contain detailed explanations of how to determine each answer. For example, to determine the size of the prize in the first question, please refer to the section of the instructions titled “Size of the Prize.”

If you have a question, please raise your hand.

Suppose a participant reports her performance level to be 4, the sum of the reported performance levels of all 20 participants is 80, and the average performance level of all participants is 4.

76

What is the prize in ECU?

What is that participant’s share of the prize?

What is that participant’s payoff in stage 1?

INSTRUCTIONS ON STAGE 2

The second stage of the experiment consists of 10 periods.

Groups and roles

At the beginning of the stage you will be randomly matched with three other participants. You and these participants will form a group consisting of four participants. The computer will randomly decide who is matched together, using randomly assigned ID numbers. The matching process is not affected by anything that happened in stage 1. At no time will your true identity be revealed to the other participants with whom you are matched, nor will you ever know the identity of these participants. You remain matched with the same three participants for all 10 periods of stage 2.

All groups will participate in a decision task in every period of stage 2.

The four participants in each group will be randomly assigned to roles by the computer. The computer will randomly select one member of each group to be the supervisor of the group. The remaining three participants will perform the task for the group, as workers.

We will now describe how the workers perform the task, how the prize and each group’s share of the prize are determined, the function of the supervisor, and how supervisor and worker payoffs are determined in stage 2.

The task for workers

The task will be different for the supervisors and for the workers. We will first describe the task for workers and will later describe the role of supervisors.

Like in stage 1, each participant in the role of a worker will privately and individually roll a 6-sided die inside a cup to determine his or her individual performance level. After rolling the die each worker will then report his or her individual performance level on the computer screen.

Then, the computer will add up all reported performance levels within each group to determine the average reported performance for each group. The average reported performance level of a group represents how well the workers in that group report having collectively performed the task.

A group’s payoff

The group’s payoff in a period depends on the group’s reported average performance level and on the average reported performance levels of all of the other groups.

77

In particular, the average reported performance level across all groups determines the size of the total prize available for performing the task. Each group’s performance level determines that group’s share of the prize. This share, multiplied by the total prize, is that group’s payoff. This group payoff is then divided between all four members of that group, the supervisor and the three workers.

Specifically, each group’s total payoff is calculated as follows:

1. Size of the prize: The size of the total prize depends on the average of the reported group performance levels of all of the groups, including your group. The prize is largest, equal to 1250 ECU, when the average reported group performance level of all groups is not higher than the expected average performance level from rolling a die, or 3.5. If all workers in a group roll their die and report the observed performance level, one would expect an average group performance score of 3.5. If this is the case for all groups, then one would also expect an average reported group performance level of 3.5 across all groups. If the average reported group performance level is 3.5 or lower, the total size of the prize will be fixed at 1250 ECU. But a higher average reported group performance level will lower the size of the prize.

The table below shows how the prize varies with the average reported group performance level of all groups. This table is the same as in stage 1. The average reported group performance level of all groups is calculated by adding up the average reported performance levels reported by all groups, and then dividing it by the number of groups, which is equal to 5.

78

Average reported group performance level of all groups Size of the prize Prize in ECU

1 100% 1250

1.5 100% 1250

2 100% 1250

2.5 100% 1250

3 100% 1250

3.5 (expected) 100% 1250

4 88% 1100

4.5 76% 950

5 64% 800

5.5 52% 650

6 40% 500

The exact size of the prize is determined by the following formula:

That is, for every unit by which the average reported group performance level is above 3.5, the size of the prize decreases by 24 percent.

If  the  average  reported  group  performance  level  is  less  than  or  equal  to  3.5  (expected  average  performance  level):    

Prize  =  1250  

If  the  average  reported  group  performance  level  is  greater  than  3.5  (expected  average  group  performance  level):  

Prize  =  1250  –  300  x  (average  reported  group  performance    

level  of  all  groups  –  3.5)  

79

2. Your group’s share of the prize: Each group’s share of the total prize is determined by the ratio of that group’s average reported performance level and the sum of the average reported group performance levels of all groups. Group′s  share  of  the  prize

=Group′s  average  reported  performance  level

Sum  of  average  reported  group  performance  levels  of  all  groups

Thus, for any total size of the prize, your group’s share of the prize is greater when your group’s reported average performance level is greater.

3. Your group’s payoff: Each group’s total payoff is calculated by multiplying the size of the prize by that group’s share of the prize.

Your  group!s  payoff = (Size  of  the  prize)  𝑥  (Your  group!s  share  of  the  prize)

Notice that the ways in which the size of the prize and each group’s share of the prize are calculated in essentially the same way as in stage 1.

The role of the supervisor

One person in each group is in the role of supervisor. The same person in each group will be supervisor in all 10 periods of stage 2.

The group supervisor does not perform the task, but receives the same information as the group workers. In particular, the supervisor also observes the reported individual performance level of each worker in the group and the average reported group performance levels of all groups.

{LEADER INCENTIVES ONLY & LEADER CONDITIONS: The supervisor will also determine how part of the group’s payoff is divided among the workers. Specifically, after finding out the group’s payoff for a period, the supervisor will divide 45% of this total group payoff among the workers, as a “bonus.” That is, 45% of the total group payoff in a round is set aside as a bonus for the supervisor to distribute among the workers.

Bonus = 0.45  𝑥  Group!s  payoff

The supervisor must distribute this entire bonus amount among the three workers.

To divide the bonus among the workers, the supervisor will enter a share of the bonus, as a percentage, that is to be received by each worker. For instance,

• The supervisor can give the entire bonus to one worker, by specifying that this worker receive 100 percent of the bonus and the other two workers each receive 0 percent.

• Alternatively, the supervisor can distribute the bonus equally among the three workers, by specifying that each worker receive 33.3 percent of the bonus.

80

• Or, the supervisor can distribute the bonus equally among only two workers, by specifying that each of these two workers receive 50 percent of the bonus and the other worker receives 0 percent.

• The supervisor can distribute the bonus in any other way among the three workers, by specifying three numbers that add up to 100 percent.

The supervisor cannot give any of the bonus to him or herself. The supervisor’s own payoff is not affected by how he or she distributes the bonus among the three workers.}

{LEADER STATEMENTS ONLY CONDITION: At the beginning of each period, before workers roll their die and report their performance, the supervisor in each group will communicate with the workers in his or her group. The supervisor will have 90 seconds to send messages to his or her group workers. Messages will be sent via a chat box on the computer screen. The group workers cannot send messages to each other or to the supervisor.

When sending messages, please do not provide any information that could reveal your identity or try to elicit the identity of others and avoid using any offensive language in your messages.}

Supervisor and worker payoffs

In a period, each of the four group members – both supervisors and workers – will receive portions of the group’s share of the prize. Specifically, the supervisor’s and workers’ portions are determined as follows:

1. Supervisor’s payoff: In each group, the supervisor receives one fourth (25 percent) of the group’s payoff. That is,

Supervisor!s  payoff = 0.25  𝑥  Group!s  payoff

{NO LEADER & LEADER STATEMENTS ONLY CONDITIONS:

2. Worker’s payoffs: In each group, each of the three workers will also receive one-fourth (25 percent) of the group’s payoff. That is,

Each  worker!s  payoff = 0.25  𝑥  Group!s  payoff

The total amount received by the three workers is 75% (25% + 25% + 25%) of the group’s payoff.}

{LEDAER INCENTIVES ONLY & LEADER CONDITIONS:

2. Worker’s payoffs: In each group, each of the three workers will receive one-tenth (10 percent) of the group’s payoff, plus any amount that is allocated to that worker by the supervisor as part of the bonus.

81

                             Each  worker!s  payoff = 0.10  𝑥  Group!s  payoff                                                                                      +  bonus  awarded  by  supervisor  

Recall that the supervisor can divide 45% of the group’s payoff among the workers, as a bonus, in any way that he or she desires. Each worker will therefore receive the guaranteed 10% share that each worker receives, plus a bonus between 0% and 45% of the group’s payoff. A worker can therefore receive, in total, any amount between 10% and 55% percent of the group’s total payoff, depending on how the supervisor chooses to divide the 45% bonus.

The total amount received by the three workers is 75% (10% + 10% + 10% + 45%) of the group’s payoff.}

Are there any questions about how the size of the prize, shares of the prize, or supervisor’s or workers’ payoffs will be determined in stage 2? Remember that if you have a question at any point during the experiment, you should raise your hand and wait for the experimenter to come to you.

Feedback

After each period, each member of your group, the supervisor and the three workers, will find out:

• The reported individual performance levels of each worker in your group • The average group performance level of your group • The average group performance level of each group • The size of the prize • Your group’s share of the prize • Your group’s payoff • The payoff received by each member of your group, including you. {LEADER

INCENTIVES ONLY & LEADER CONDITIONS: This will include the fixed 10 percent share and the additional bonus share, determined by the supervisor, received by each worker in your group.}

At the end of the experiment, you will be informed about your total earnings, i.e., the sum of your payoffs for stage 1 and for all 10 periods of stage 2. After the experiment your total earnings in ECU will be converted into GBP and will be paid privately, in cash, together with your participation payment.

Examples

We will now go through some examples to make sure that it is clear to everyone how payoffs are determined.

Example 1:

82

Suppose that your group’s reported average performance level is 5, the sum of average performance levels of all groups, including your group’s average performance level, is 25, and the average performance level of all groups is 5.

{LEADER INCENTIVES ONLY & LEADER CONDITIONS: Suppose that the group supervisor decides to divide the 45% bonus evenly among the group workers. That is, each group worker receives 15% of the group’s payoff as part of the bonus, in addition to the fixed 10% share, or 25% in total.}

In this case:

• The total prize is 800 ECU (i.e., 1250 ECU – 300 ECU (5 – 3.5) = 800 ECU). The prize is smaller than 1250 ECU because the average performance level of all groups (5) is greater than 3.5

• Your group’s share of the prize is 5/25 = 0.2.

• Your group’s payoff for this task is 0.2 x 800 ECU = 160 ECU.

• The supervisor’s payoff is 0.25 x 160 = 40 ECU.

• {NO LEADER & LEADER STATEMENTS ONLY CONDITIONS: Each worker’s payoff is 0.25 x 160 = 40 ECU.}

• {LEADER INCENTIVES ONLY & LEADER CONDITIONS: The available share for the supervisor to distribute among workers as a bonus is 0.45 x 160 ECU = 72 ECU.

• Each worker’s payoff is 0.10 x 160 ECU + 72 ECU / 3 = 16 ECU + 24 ECU = 40 ECU.

Note that the first part of a worker’s payoff is the fixed 10 percent that each worker receives and the second part is the portion that is allocated by the supervisor as a bonus.}

Example 2:

Now suppose that your group’s average reported performance level is 3, the sum of average performance levels of all groups, including your group’s average performance level, is 17.5, and the average performance level of all groups is 3.5. {LEADER INCENTIVES ONLY & LEADER CONDITIONS: Suppose that the group supervisor decides to give the entire 45% bonus to one of the three workers and to give none of the bonus to the other two workers.}

In this case:

• The total prize is 1250 ECU because the average performance level of all groups (3.5) is equal to 3.5.

• Your group’s share of the prize is 3/17.5≅0.17.

• Your group’s payoff is 0.17 x 1250 ECU = 212.5 ECU.

83

• The supervisor’s payoff is 0.25 x 212.5 ECU ≅ 53.1 ECU.

• {NO LEADER & LEADER STATEMENTS ONLY CONDITIONS: Each worker’s payoff is 0.25 x 212.5 ECU ≅ 53.1ECU.}

• {LEADER INCENTIVES ONLY & LEADER CONDITIONS: The available share for the supervisor to distribute among workers as a bonus is 0.45 x 212.5 ECU≅ 95.6 ECU.

• One worker receives a payoff of 0.10 x 212.5 ECU + 95.6 ECU ≅ 116.9ECU.

• The other two workers receive a payoff of 0.10 x 212.5 ECU + 0 ≅ 21.3 ECU.

Note that the first part of a worker’s payoff is the fixed 10 percent that each worker receives and the second part is the portion that is allocated by the supervisor as a bonus.}

Practice questions

Before we begin with Stage 2, in which you will perform the above task in groups for 10 periods, we will first ask you to answer some practice questions about payoffs. This is done to make sure that everyone understands how payoffs are calculated.

Please click “OK” on your screen now, to see the practice questions, and try to answer them. If you have a question or are confused, please raise your hand and wait for the experimenter to come to you.

As soon as everyone has answered the practice questions correctly, Stage 2 will start.

{PRACTICE QUESTION STAGE 2, on screen}

Please answer the questions below.

If you need help, the instructions contain detailed explanations of how to determine each answer. For example, to determine the size of the prize in the first question, please refer to the section of the instructions titled “Size of the Prize.”

If you have a question, please raise your hand.

1) Suppose a participant in the role of a worker reports her performance level to be 4, and the average performance level of that participant’s group is 4. The sum of average reported group performance levels of all 5 groups is 20 and the average reported group performance level of all groups is 4. {LEADER INCENTIVES ONLY & LEADER CONDITIONS: Suppose the supervisor decides to allocate the entire bonus of 45% to that participant.} What is the prize in ECU? What is the group’s share of the prize? What is the group’s payoff in that period? What is the group supervisor’s payoff in that period? (Hint Button: This is one fourth (0.25) of the group’s payoff.)

84

{NO LEADER & LEADER STATEMENTS ONLY CONDITIONS: What is a group worker’s payoff in that period? (Hint Button: This is one fourth (0.25) of the group’s payoff.)} {LEADER INCENTIVES ONLY & LEADER CONDITIONS: What is the amount that can be allocated by the group supervisor to the group workers in that period as a bonus? (Hint Button: This is 45 percent (0.45) of the group’s payoff.) What is the payoff of the group worker receiving the entire bonus in that period? (Hint Button: The worker receives 10 percent plus 45 percent, or 55 percent (0.55), of the group’s payoff.)}

2) At the end of a period in stage 2, who is informed about a participant’s reported performance level in that period? ☐ all participants in the experiment ☐ only the participant’s group supervisor. ý all members of the participant’s group.

INSTRUCTIONS ON STAGE 3

Stage 2 of the experiment is complete. Before concluding the experiment, you will complete several questionnaires in which you will have the possibility to earn additional ECUs.

We will next ask you and all other participants to evaluate different possible choices one might have made in stages 1 and 2 of the experiment. Specifically, we will describe a choice that a participant in the experiment might have made, and you should decide whether making that choice would be “socially appropriate” and “consistent with moral or proper social behavior” or “socially inappropriate” and “inconsistent with moral or proper social behavior.” By socially appropriate, we mean behavior that most people agree is the “correct” or “ethical” thing to do. Another way to think about what we mean is that, if someone were to make a socially inappropriate choice, then someone observing this behavior might get angry at the person who made the choice for acting in that manner.

In each of your responses, we would like you to evaluate what constitutes socially appropriate or inappropriate behavior. To give you an idea, consider the following example.

Someone is at a local cafe. While there, the person notices that someone has left a wallet at one of the tables. How appropriate would it be to take the wallet?

Very socially

inappropriate

Somewhat socially

inappropriate

Somewhat socially

appropriate

Very socially

appropriate

☐ ☐ ☐ ☐

85

If this were the situation we asked you about in the study, you would indicate the extent to which you believe taking the wallet would be "socially appropriate" and "consistent with moral or proper social behavior" or "socially inappropriate" and "inconsistent with moral or proper social behavior". Recall that by socially appropriate we mean behavior that most people agree is the "correct" or "ethical" thing to do.

For example, suppose you thought that taking the wallet was very socially inappropriate. Then, you would indicate your response by selecting the first box.

For each choice you make, we will compare your response to the response of one other participant. If you give the same response as the one provided by the selected other participant, then you will receive an additional ECU 17.5 (GBP 0.50). This amount will be paid to you, along with your other earnings, at the conclusion of the experiment.

For instance, in the example situation above, if your response had been "somewhat socially inappropriate," then you would receive an additional ECU 17.5 (GBP 0.50) if this was also the response provided by a randomly selected other participant in today’s session. Otherwise you would not receive any additional money for this question.

Are there any questions about this task? Once we continue, you will see detailed descriptions of the choices you are to evaluate and instructions on how to proceed.

If you have any questions at any time, please raise your hand and wait for the experimenter to come to you.

After you have answered all questions you will see how much how have earned. Then we will start another questionnaire. For answering theses further questions we will pay you an additional ECU 70 (GBP 2.00). After completing the questionnaire you will be informed about your total payoff and receive your payment.

On computer screen:

Question 1: For the following question, we will compare your response to the response of one randomly selected other participant in the session. You will receive the additional payment if your response is the same as the one provided by the randomly selected participant in the session.

Suppose that, in Stage 1, someone rolled a 1 on his or her die. How appropriate would it have been for that participant to report a higher number than 1 in Stage 1 of the experiment, when earnings were based on individual performance level and on the individual performance level of other participants?

Very socially

inappropriate

Somewhat socially

inappropriate

Somewhat socially

appropriate

Very socially

appropriate

☐ ☐ ☐ ☐

86

Question 2: For the following question, we will compare your response to the response of one randomly selected other participant in the session. You will receive the additional payment if your response is the same as the one provided by the randomly selected participant in the session.

Suppose that, in Stage 2, a Worker rolled a 1 on his or her die. How appropriate would it have been for that Worker to report a higher number than 1 in Stage 2 of the experiment, when earnings were based on the group performance level of the Worker’s group and on the performance level of other groups?

Very socially

inappropriate

Somewhat socially

inappropriate

Somewhat socially

appropriate

Very socially

appropriate

☐ ☐ ☐ ☐

Question 3: For the following question, we will compare your response to the response of one randomly selected other person in your group. That is, you will receive the additional payment if your response is the same as the one provided by one randomly selected other member of your group.

Suppose that, in Stage 2, a Worker rolled a 1 on his or her die. How appropriate would it have been for that Worker to report a higher number than 1 in Stage 2 of the experiment, when earnings were based on the group performance level of the Worker’s group and on the performance level of other groups?

Very socially

inappropriate

Somewhat socially

inappropriate

Somewhat socially

appropriate

Very socially

appropriate

☐ ☐ ☐ ☐

{WORKERS:

Question 4: For the following question, we will compare your response to the response provided in Question 3 by the Supervisor in your group. That is, you will receive the additional payment if your response is the same as the one provided by the Supervisor in your group when answering Question 3. For your information, we present Question 3 again, below.}

{SUPERVISORS:

87

Question 4: For the following question, we will compare your response to the response provided in Question 3 by one randomly selected other Supervisor in today’s session. That is, you will receive the additional payment if your response is the same as the one provided by one randomly selected other Supervisor when answering Question 3. For your information, we present Question 3 again, below. }

Suppose that, in Stage 2, a Worker rolled a 1 on his or her die. How appropriate would it have been for that Worker to report a higher number than 1 in Stage 2 of the experiment, when earnings were based on the group performance level of the Worker’s group and on the performance level of other groups?

Very socially

inappropriate

Somewhat socially

inappropriate

Somewhat socially

appropriate

Very socially

appropriate

☐ ☐ ☐ ☐

Elicitation of Protected Values (on computer screen)

1) People inside a firm can sometimes make more money for themselves and for their co-workers by misreporting information to people outside the firm. That is, to benefit their firm, people may have an incentive to report having done better at their job (more sales, higher profits) than is actually true.

What is your opinion on people engaging in this kind of misreporting? Please choose the appropriate category. That is:

very immoral ☐ ☐ ☐ ☐ ☐ ☐ ☐ very moral

not at all praiseworthy ☐ ☐ ☐ ☐ ☐ ☐ ☐ very praiseworthy

not at all blameworthy ☐ ☐ ☐ ☐ ☐ ☐ ☐ very blameworthy

not at all outrageous ☐ ☐ ☐ ☐ ☐ ☐ ☐ very outrageous

not at all acceptable ☐ ☐ ☐ ☐ ☐ ☐ ☐ very acceptable

2) People often have an opportunity to inflate how well they appear to have performed at their job. Some view such modification as a violation of truthfulness, others regard it as acceptable protection of personal interests. What do you think about the value of truthfulness in such a situation?

Truthfulness is something... ... that one should not sacrifice, no matter what the (material or other) benefits.

88

very strongly disagree ☐ ☐ ☐ ☐ ☐ ☐ ☐ very strongly agree

... for which I think it is right to make a cost-benefit analysis.

very strongly disagree ☐ ☐ ☐ ☐ ☐ ☐ ☐ very strongly agree

... that cannot be measured in monetary terms. very strongly disagree ☐ ☐ ☐ ☐ ☐ ☐ ☐ very strongly agree

... about which I can be flexible if the situation demands it. very strongly disagree ☐ ☐ ☐ ☐ ☐ ☐ ☐ very strongly agree