No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Loss Averse Agents and Lenient Supervisors in Performance Appraisal
Abstract
A consistent empirical literature shows that in many organizations supervisors systematically overrate their employees’ performance. Such leniency bias is at odds with the standard principal-agent model and has been explained with causes that range from social interactions to fairness concerns and to collusive behavior between the supervisor and the agent. We show that the principal-agent model, extended to consider loss-aversion and reference-dependent preferences, predicts that the leniency bias is comparatively less detrimental to effort provision than the severity bias. We test this prediction with a laboratory experiment where we demonstrate that failing to reward deserving agents is significantly more detrimental than rewarding undeserving agents. This offers a novel explanation as to why supervisors tend to be lenient in their appraisals.
... Leniency bias may also be a result of luck: Bol and Smith (2011) show that when evaluations are related to an objective measure in which an uncontrollable stochastic factor (luck) plays a role, supervisors evaluations are higher when the agent has bad luck (this factor decreases the agent's objective measure), but are not lower if he has good luck. Other justifications also come from behavioral considerations, such as loss-aversion, from the agent's side (Marchegiani et al. 2016). All these potential explanations are excluded in my setting. ...
... 10 Bol (2011) shows that it is positively associated with employees' performance. From a comparative aspect, Marchegiani et al. (2016) show that failing to reward a deserving agent under a severe contract is significantly more detrimental to effort provision than rewarding an undeserving agent under a lenient contract. The effect of evaluation bias involves severe bias that is negatively associated with the performance (cf. ...
... For instance,Bol (2011) shows that leniency bias has a positive effect on motivating employees within a large bank in the Netherlands. The empirical evidence also documents the negative correlation between evaluation biases involving severity bias and the effort provision (see,Ahn et al. 2010;Bol 2011;Trapp and Trapp 2019;Marchegiani et al. 2016). ...
My dissertation investigates the design of information policy in three different types of strategic interactions: Chapter 1: "Performance Evaluation Design in Dynamic Incentive Contracts", examines how to motivate employees in an organization by strategically disclosing evaluations. Consider a continuous-time principal-agent setting where the agent exerts effort to generate output and the principal subjectively evaluates and pays for the agent's performance. If the principal can commit, what type of evaluation system should she implement? I show that the optimal evaluation system is not based directly on the output, but rather on an adjusted evaluation of output. In particular, it assigns inflated evaluations when the agent's continuation value is low and deflated evaluations when the continuation value is high. Adding such adjustment into evaluations allows the principal to recalibrate the agent's continuation value, which improves the contract by capturing gains from concavification that are not feasible in contracts directly based on output. As a result, adjusting evaluations also induces weakly higher volatility in the agent's continuation value even though the agent is risk-averse. Moreover, I show that additional contractual possibilities such as leaving the firm for an outside option, promotion, and reciprocity in output, could result in strengthening different evaluation biases at the optimum. My results help explain evaluation biases that have been empirically observed in appraisal systems. Chapter 2: "Persuasion of Interacting Receivers", investigates how the consideration of realistic features could complicate the structure of information policy in situations of strategic interaction. Its main contribution is to propose a general framework that allows a formal characterization for several relevant features and provides a tractable way to design optimal information policy in such settings. Specifically, I consider a setting where a sender could communicate privately with multiple receivers before they engage in a one-shot strategic interaction. To understand optimal information policy, standard approaches would focus on signals that recommend actions. However, if there are realistic features outside the scope of the standard model, such a focus could be suboptimal. I consider the following four features: (i) the equilibrium selection rule may be different from the sender-preferred one, (ii) receivers may have private information, (iii) receivers may have non von Neumann--Morgenstern utilities, and (iv) receivers may have heterogeneous priors. I establish a generalized obedience principle. In this version, the sender recommends actions and conjectures. I further provide a sufficient condition under which it is without loss of generality to reduce the messages involved in any signal from continuum to countably many. The construction provides a tractable way to compute optimal signals. I also apply my result to study information policy design in two applications in which the equilibrium selections differ from the sender-preferred selection and receivers may be privately informed. Chapter 3: "Derandomized Persuasion Mechanisms", investigates when focusing only on information policy that either fully reveals or (partially) pools the underlying states without adding extra noise is without loss of generality. I consider a setting where one sender can communicate with several privately informed receivers through a persuasion mechanism before the receivers play a game. I show that for any potentially randomized persuasion mechanism, under certain conditions, there is an effectively equivalent derandomized persuasion mechanism, and these two mechanisms have the same set of equilibria. Overall, this paper provides a rationale for the fact that persuasion mechanisms are often deterministic in practice.
... (2011) and Marchegiani et al. (2016), in all periods, subjects were confronted with a tedious and focus-demanding task, which consists of counting the number of ones in tables composed by 50 randomly placed ones and zeros. This real-effort task has various advantages: (1) no prior (economic) knowledge is needed from the subjects, (2) nearly no learning is possible from the subjects throughout the game periods, (3) subjects' performance is measurable without difficulty, (4) the boringness ensured a positive cost-of-effort from the subjects, and (5) the pointlessness ensured that no subject could derive any ben-efit (e.g. ...
... The number of tables they correctly counted was used to design a feasible contractual effort in subsequent parts of the experiment, but subjects were not informed about this. As in Marchegiani et al. (2016) and Cosaert et al. (2019), this phase permitted to normalize the cost-of-effort for the task across players by scaling the individual performance targets, assigned in Phase II, to the subjects' actual abilities. ...
... As mentioned by Marchegiani et al. (2016), the (acceptable) target should be achieved by exerting a high, but not too high, effort on the task 8 . The ideal target, however, has voluntarily been determined to be more difficult to achieve. ...
This thesis aims at contributing to make the energy renovation market long-lasting and self-sustaining. To achieve this, our objective is to quantify the risk of not achieving energy performance after renovation. First, we analyze households’ psychological factors that should be considered to improve energy consumption prediction models. Drawing on the Je rénove BBC program, we highlight four cognitive biases that negatively impact the difference between actual and predicted energy consumption. We then study the most appropriate contract structures improving the flow and quality of renovation projects, encouraging craftsmen to work better. On one hand, we determine optimal contracts for an Agent who has to perform two tasks and underestimates the impact of one of them on the building's performance. On the other hand, we test individual-based and group-based incentives on the ability of several real Agents (craftsmen) to coordinate, according to their initial training (DORéMI, …).
... Some theories in related fields, like the economics of law and crime, have investigated the differential deterrence effects of judicial systems that commit Type I errors (i.e., convicting innocent people) or Type II errors (i.e., failing to convict those who are guilty) (Rizzolli & Stanca, 2012). Gamba and Stanca (2016) modeled a bidding contest and varied inclusion and exclusion errors; Marchegiani, Reggiani, and Rizzolli (2016) analyzed whether failing to reward a deserving agent is more detrimental than rewarding an undeserving agent. ...
... Specifically, we amend traditional (agency) perspectives with social identity arguments (Ashforth & Mael, 1989;Tajfel & Turner, 1986). Whereas agency theory predicts equally negative effects of both types of false rewards (Gamba & Stanca, 2016;Marchegiani et al., 2016), we anticipate that false positives are tangible artifacts and symbols that conflict with collective, behaviorally relevant belief systems and norms. ...
... Performance appraisals may be subjective or based on objective output measures. With subjective appraisals, raters often exhibit severity, leniency, or centrality biases (Landy & Farr, 1983;Marchegiani et al., 2016;Murphy & Cleveland, 1995), and ratings tend to be unreliable (Viswesvaran, Ones, & Schmidt, 1996). Objective measures, such as productivity or sales volume (Rynes et al., 2005), can remedy such concerns but are seldom available, or they might be available only at aggregate, instead of individual, levels. ...
This study investigates the effects of rewards in a research and development (R&D) setting in which employees’ inventive efforts lead to patented inventions. Pay for performance (PFP) for inventions is associated with two challenges: Low-quality inventions may be rewarded (false positives), and high-quality inventions may be overlooked (false negatives). Building on previous findings regarding the motivational and informational effects of rewards, we use social identity theory to predict that different types of inventors react differently to such false positive and false negative information. Specifically, we hypothesize that PFP that produces false positives has detrimental effects on corporate inventors with a taste for science, who are motivated by scientific prestige, reputation, and intellectual curiosity. The empirical results from survey data related to
3,995 inventor–patent pairs show that, for this particular group of inventors, false positives are associated with reduced effort in research activities and fewer interactions with peers in the R&D department. In addition, these effects are stronger when firms have many patents and thus provide less noisy information to corporate inventors.
... The experiment consists of a real-effort game played repeatedly. Following Abeler et al. (2011) and Marchegiani et al. (2016), in all periods, subjects are confronted with a tedious and focus-demanding task, which consists of counting the number of ones in tables composed by 50 randomly placed ones and zeros. This task does not require any prior knowledge, and performance is easily measurable. ...
... The number of correct tables is used to design a feasible contractual effort in subsequent parts of the experiment, but subjects are not informed about this. As in Marchegiani et al. (2016) and Cosaert et al. (2019), this phase permits to normalize the costof-effort for the task across players by scaling the individual performance targets, assigned in the repeated real-effort game, to the subjects' actual abilities. ...
... Underlining such drawbacks are studies showing that when employment contracts are framed as a combination of bonus and penalty incentives, they appear as less attractive, which can be accredited to Loss Aversion (Brink & Rankin, 2013). On the contrary, experimental research using the principal-agent model extended to consider Loss Aversion and reference-dependent preferences has found that failing to reward a deserving agent is significantly more destructive than rewarding undeserving agents, which could indicate that incentive contracts modelled after Loss Aversion would elicit a greater increase in performance than regular gain contracts (Marchegiani et al., 2016). Specifically, in regards to performance, experimental research has found that variable pay framed as a loss is associated with an increase in both work effort and performance and with a decrease in organizational deviant behavior (Merriman & Deckop, 2007). ...
... In this paper, we tested whether the effect of Loss Aversion is specific to previously tested domains of behavior or if Loss Aversion could be utilized in order to increase performance (cognitive and mechanical) as a part of an employee compensation scheme, in an experimental task (Ariely et al., 2009;Etchart-Vincent & l'Haridon, 2011;Kahneman & Tversky, 1979;Levy, 1994). Simultaneously, and to our knowledge for the first time, we tested whether the magnitude of such a compensation scheme would have a positive or negative effect on the level of (cognitive/mechanical) performance in such a task, as theorized by the Yerkes-Dodson Law (Calabrese, 2008;Teigen, 1994;Yerkes & Dodson, 1908 (Brink & Rankin, 2013;Fryer Jr. et al., 2012;Hochman et al., 2014;Hossain & List, 2012;Imas et al., 2017;Marchegiani et al., 2016). Moreover, it contributes to the existing research on the Yerkes-Dodson Law and the relationship between varying magnitudes of an incentive and its impact on performance (Broadhurst, 1957;Calabrese, 2008;Corbett;Gino, 2016;Hockey, 1997;Piefke & Glienke, 2017;Teigen, 1994;Yerkes & Dodson, 1908). ...
Monetary bonus schemes are one of the most well-used forms of employee compensation in the modern business world. Yet, such schemes are primarily constructed as gains to incentivize an increase in work effort and performance. Using insights from behavioral economics, we construct a novel extrinsic compensation system modelled after Loss Aversion and the Yerkes-Dodson Law. We test this system using two experimental designs that measure performance based on cognitive- and mechanical efforts, respectively. In study 1 we find no difference in cognitive performance between 4 different bonus schemes. In a pre-registered study 2 we again find no difference in mechanical performance between 4 different conditions. Contrary to previous research, our findings suggest no significant effect of bonus schemes modelled after Loss Aversion and Yerkes-Dodson Law, whether administered alone or in combination. We discuss limitations and implications for compensation design and research.
... By contrast, leniency bias affects it positively, which could be explained by an increased intrinsic motivation resulting from perceived fairness. Contrary to Bol (2011), Marchegiani, Reggiani andRizzolli (2016) find in a real effort laboratory experiment that leniency is detrimental to performance. But rewarding undeserving agents has a less detrimental effect than failing to reward deserving agents (the "severity bias"), contrary to what agency theory predicts. ...
... Baseline/ Respect treatments in 3 market conditions Flat pay (+) Praise used as a coordination device to initiate relational contracts; increases effort Marchegiani,Reggiani, Rizzolli (2016) Counting occurrence of 1s in tables -84 ...
... The firing threats, which happen only in rare instances and do not occur every time the agent performs poorly, enable the principal to overlook the agent's bad performance. Second, Marchegiani et al. (2016) study the effects of leniency in performance appraisal on employees' effort in a laboratory experiment and find that leniency bias, as compared with severity bias, is more beneficial for effort provision of loss-averse employees. This experimental evidence is consistent with my finding that the optimal stochastic contract under loss aversion is characterized by leniency. ...
I examine whether stochastic contracts benefit the principal under moral hazard and loss aversion. Incorporating the agent's expectation‐based loss aversion and allowing for stochastic contracts, I find that stochastic contracts reduce the principal's cost as compared with deterministic contracts. The optimal stochastic contract pays a high wage not only when good signals are realized but also with a positive probability after the realization of bad signals. The findings have an important implication for designing contracts for loss‐averse agents: the principal should insure the agent against wage uncertainty by employing stochastic contracts that increase the probability of a high wage.
... These findings suggest that the impact of loss aversion depends on the incentive scheme, and performance can be improved, or at least the negative impact of loss aversion reduced, if incentives are properly adjusted. For example, Marchegiani, Reggiani and Rizzolli (2016) show that it helps for managers to be lenient when workers are loss averse; and Dodonova and Khoroshilov (2006) show that compensation packages with a large share of stock options are optimal for loss averse managers. In this paper, we further contribute to understanding how loss aversion interacts with incentives, and how it can be harnessed to improve performance. ...
We study the optimal allocation of prizes in rank-order tournaments with loss averse agents. Prize sharing becomes increasingly optimal with loss aversion because more equitable prizes reduce the marginal psychological cost of anticipated losses. Furthermore, loss aversion can boost effort if prizes are sufficiently equitable, but otherwise effort declines with loss aversion. Overall, these results give credence to more equitable allocations of competitive rewards. A win-win scenario is where optimal prizes are equitable even under loss neutrality, in which case the principal benefits from agents' loss aversion.
... These early discussions anchored loss aversion within fundamental management accounting judgments and decision-making contexts such as budgeting (Lee 1994) and the recovery of sunk costs (Schaubroeck and Davis 1994). By 2016, there was a peak in the annual number of publications (7), with several studies investigating the principal-agent relationship, specifically focusing on the impact of agents' loss aversion on business performance when control mechanisms are inadequate (Grolleau et al. 2016;Marchegiani et al. 2016). Moreover, Figure 2 demonstrates a similar trend in the average citations per year. ...
To date, the studies on managerial loss aversion have produced contradictory findings, making it impossible to: (i) identify the ultimate impact of managerial loss aversion on the value that organisations create for themselves and for their stakeholders, and (ii) mitigate the effect of managerial loss aversion to improve corporate value creation. With the aim of filling this gap, the authors of this paper first performed a Systematic Literature Review (SLR), resulting in 65 relevant papers. The 65 papers were then analysed through a Thematic Analysis (TA), which was aimed at isolating and revising the single effects of managerial loss aversion on the corporate value creation process. Once it became clear when and how managerial loss aversion leads to negative impacts on corporate value creation (such as suboptimal investments in corporate social responsibility, short-term-oriented budget expenditures, illegal corporate conduct in favourable contexts, and low demand for audit quality), a novel theoretical framework was built. This framework proposes some preliminary approaches to mitigate these detrimental effects. In particular, future empirical research may operationalise potential debiasing strategies, derived from critical analysis of the literature, to reduce managerial loss aversion in different business settings, thereby improving corporate value creation.
... The relative success of the candidate-keeping her abilities constant-declined by approximately 20% (in relative terms) when the quality of the neighborhood and housing conditions decreased.Non-parametric testing, for each pairwise comparison of the treatments, statistically validates this descriptive evidence. Exploiting the within-subject design of our experimental setting, McNemar permutation tests(Westfall et al., 2010;Marchegiani et al.,Table 1refines the non-parametric analysis employing a regression framework that allows us to easily control for the randomized ancillary within-call variations we implemented in the individual applications to disguise them. ...
Through a correspondence study, this paper investigates whether employers discriminate job applicants based on their living conditions. Exploiting the natural setting provided by a Rapid Re-housing Program, we sent 1,347 job applications for low-qualified front-desk jobs in Brno, Czech Republic. The resumes exogenously differed in only one main aspect represented by the address of the applicants, signaling both the quality of the neighborhood and the quality of the housing conditions in which they were living. We found that while the higher quality of the district has a strong effect in increasing the hiring chances (+20%) the actual improvement of the living conditions standards, per se, does not generate any significant positive effect.
... By contrast, leniency bias affects it positively, which could be explained by an increased intrinsic motivation resulting from perceived fairness. Contrary to Bol (2011), Marchegiani et al. (2016 find in a real-effort laboratory experiment that leniency is detrimental to performance. But rewarding undeserving agents has a less detrimental effect than failing to reward deserving agents (the "severity bias"), contrary to what agency theory predicts. ...
... The preference to provide a higher and lower rating as the presence of incentives knowledge is due to the intention of users to minimize the negative effect they may cause to the drivers. Marchegiani et al. (2016) supported this prediction by stating that failing to reward a high-performer is more detrimental than failing to rewarding low-performer agents, which motivates evaluators to be more lenient. Thus, the hypothesis is proposed as follows: ...
Online transportation as a developing industry in Indonesia applies bonus-or-terminate incentives based on both objective and subjective performance evaluation. Regarding bias problems often found in the subjective evaluation, this paper aims to examine factors that influence bias in performance evaluation of the online transportation drivers. The data were collected by an online survey to users of online transportation in Indonesia using convenience sampling. Multiple regression analysis was utilized to analyze 163 data. The result shows that users of online transportation services tend to generate a biased rating, which is leniency bias. Altruism and knowledge of incentives scheme as the users’ internal factors significantly affect the biased evaluation. On the other hand, the external factors of the users, i.e., the frequency of usage and evaluation timing, do not significantly affect the biased evaluation. However, the other external factor, i.e., travel distance and duration, is found to affect the leniency bias positively. This paper concludes that more information related to the evaluation object and the inherent characteristic of an individual as the effect of collectivist national culture may lead to the generation of biased performance evaluation by the evaluator to help evaluated party avoiding penalty/termination in the competitive working environment.
... More recently, Bol (2011) analyzes compensation data from a financial service provider in the Netherlands and shows that leniency bias is higher when the employeemanager relationship is stronger. In an experiment, Marchegiani et al. (2014) show that failing to reward a deserving agent may be detrimental to productivity and this may provide a rationale for leniency bias. For a thorough survey of the empirical relevance of this phenomenon, especially when evaluations matter for pay setting, see Prendergast (1999). ...
In this paper we investigate the task the supervisor should be optimally charged with in an agency model in which the principal faces corruption concerns. We highlight a fundamental tradeoff between monitoring the agent's effort choice and auditing it ex-post. Monitoring proves more effective in tackling corruption, since the supervisor sends the report before the profit realization. By taking advantage of the supervisor's uncertainty about the state of nature, the principal can design a compensation scheme which prevents all forms of corruption at a lower cost. Conversely, auditing allows the principal to save the cost of supervision when the profit realization already suffices to set the compensation due to the agent. We show that the choice between monitoring and auditing crucially depends on the supervisor's ability to falsify information and the cost of hiring a supervisor.
This study explores biases in performance appraisal systems within Indonesian government institutions and proposes a tailored framework to address these challenges. Identified biases include confirmation, similarity, central tendency, anchoring, status quo, and leniency/severity biases, exacerbated by political factors, bureaucratic culture, and hierarchical structures unique to Indonesia. These biases often result in appraisals reflecting political maneuvering rather than true employee performance. Using qualitative interviews and thematic analysis, the study uncovers the root causes of these biases and evaluates their impact on fairness and effectiveness. The proposed framework integrates Equity Theory, Cognitive Bias Theory, and goal-setting principles, offering solutions such as structured appraisal methods, quarterly reviews, enhanced evaluator training, and behaviorally anchored rating scales (BARS). These recommendations address the systemic issues of bias, promote transparency, and foster a fair and motivating work environment. This research uniquely contributes to the literature by focusing on Indonesia’s public sector, where cultural and political factors significantly influence performance appraisals. It provides actionable strategies for policymakers and practitioners to create appraisal systems that are fair, effective, and aligned with the organizational goals of government institutions, ultimately enhancing employee engagement and organizational performance.
We study biases and the informativeness of subjective performance evaluations in an online experiment, testing the implications of a standard formal framework of rational subjective evaluations. In the experiment, subjects in the role of workers perform a real effort task. Subjects in the role of supervisors observe samples of the workers’ output and assess their performance. We conduct six experimental treatments varying (i) whether workers’ pay depends on the performance evaluation, (ii) whether supervisors are paid for the accuracy of their evaluations, and (iii) the precision of the information available to supervisors. Moreover, we use the exogenous assignment of supervisors to workers to investigate the association between supervisors’ social preferences and their rating quality. In line with the model of optimal evaluations, we find that ratings are more lenient and less informative when they determine bonus payments. Rewards for accuracy reduce leniency and can enhance informativeness. When supervisors have access to more detailed performance information, their ratings vary more with the performance signal and become more informative. Contrary to expectations, we do not find that more prosocial supervisors are systematically more lenient when their ratings affect workers’ payoffs. Instead, they are more diligent in their rating behavior, resulting in more accurate and informative performance evaluations.
This paper was accepted by Marie Claire Villeval, behavioral economics and decision analysis.
Funding: Funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany’s Excellence Strategy [Grant EXC 2126/1-390838866].
Supplemental Material: The online supplement and data files are available at https://doi.org/10.1287/mnsc.2022.02267 .
The development of the Internet preceded significant changes in the environment that have drastically affected behaviour of individuals. In order to achieve competitive advantage in the current hyper-competitive market, the products and services are becoming increasingly customized. Marketing as a strategic function in a company has become one of the imperatives of modern business, and the development of communication technology has given marketing activities a new framework. With the evolution of social networks, a brand new concept of marketing, e-marketing, is rapidly developing. Social networks have changed the way we communicate and individuals have become active content creators on social networks. This fact used to be frightening because companies would lose control over their own communication, but today companies encourage users to actively participate in social media communication. In this new marketing concept, "influencers" have a much more significant market impact on individuals. The focus of this paper is the impact of “influencers” on consumer behaviour. The paper analyses how consumers perceive “Influencers”, and whether they affect their decisions. Therefore, the aim of the paper is to show the results of the “Influencer’s” impact on consumer behaviour and to determine their impact on consumers’ buying decisions.
Tests of labor supply models often rely on wages. However, wage variation alone generally cannot disentangle the classical time separable model and its extensions: reference dependent preferences (income targeting) and time nonseparable preferences (disutility spillovers; timing-specific preferences). We set up a novel laboratory experiment in which individuals choose their working time. We vary, independently, wages, historical income paths, and cumulative past work. We also vary the timing of experimental sessions. Statistical tests and stochastic revealed preference methods cannot reject the classical model in favor of income targeting or disutility spillovers, but the data suggest that labor supply varies by time-of-the-day.
Through a correspondence study, this paper investigates whether employers discriminate job applicants based on their living conditions. Exploiting the natural setting provided by a Rapid Re-housing Program, we sent 1347 job applications for low-qualified front-desk jobs in Brno, Czech Republic. The resumes exogenously differed in only one main aspect represented by the address of the applicants, signaling both the quality of the neighborhood and the quality of the housing conditions in which they were living. We found that while the higher quality of the district has a strong effect in increasing the hiring chances (+20%) the actual improvement of the living conditions standards, per se , does not generate any significant positive effect.
This study provides three real-effort experiments on how a forced distribution rating system (FDRS) influences team collaboration. In the first and the second experiment, we examine the performance implications of an FDRS in a card sequencing task (1) when working alone and (2) when working in a team. In the third experiment, we test how an FDRS affects knowledge sharing within teams. Our findings show that an FDRS increases the speed of completing the card sequencing task when working alone and decreases the speed of completing the card sequencing task when working in a team. Beyond that, we find that an FDRS also significantly decreases knowledge sharing within teams. As the FDRS was perceived as unfair in collaborative settings but not when working alone, we provide evidence on the role of perceived justice concerning the effects of an FDRS and shed light on the psychological and economic consequences of introducing an FDRS in environments where team collaboration is essential for success.
The reputation information provided by market feedback systems tends to be compressed in the sense that reliable and unreliable sellers have similar feedback scores. The experiment presented here features a market in which what a buyer receives is a noisy signal of what was actually sent. We focus on the influence the noise has on endogenously given feedback. The attributional uncertainty creates room for leniency in feedback giving. We find that buyer leniency reduces the informativeness of the feedback system and, in combination with uncertainty, diminishes seller trustworthiness. With a noisy signal, buyers pay about the same prices but get significantly less.
This paper was accepted by Uri Gneezy, behavioral economics.
We investigate a multi-agent moral-hazard model where agents have expectation-based reference-dependent preferences a la Koszegi and Rabin (2006, 2007). We show that even when each agent's probability of success in a project is independent, team incentives can be optimal. Because the agents are loss averse, they have first-order risk aversion to wage uncertainty. This causes the agents to work harder when their own failure is stochastically compensated through other agents' performance. In the optimal contract, both high- and low-performance agents are equally rewarded if most agents accomplish their projects; otherwise only high-performance agents are rewarded.
On the surface, it is not readily apparent how some performance appraisal research issues inform performance appraisal practice. Because performance appraisal is an applied topic, it is useful to periodically consider the current state of performance research and its relation to performance appraisal practice. This review examines the performance appraisal literature published in both academic and practitioner outlets between 1985 and 1990, briefly discusses the current state of performance appraisal practice, highlights the juxtaposition of research and practice, and suggests directions for further research.
Recent research suggests that loss framed contracts are an effective instrument for principals to maximize the effort of their agents. Framing effects arise from defining thresholds that vary the salience of losses and gains while preserving payoff equivalence of the underlying contract. While under Prospect Theory a loss frame should lead to more effort we show that contract thresholds also exert a suggestive effect on performance that can trump the impact of loss aversion. Loss framing therefore carries a risk. As agents focus their effort choice on the expressed thresholds, poorly selected thresholds reduce effort and the principal might prefer offering a contract that does not impose a threshold at all. On the other hand, imposing demanding thresholds may push effort beyond levels predicted by Prospect Theory.
Cultural comparisons enjoy increasing popularity in economics. Since cultural comparison must abandon random allocation to treatments, it is unclear whether differences found between countries can be attributed to country characteristics or are merely driven by differences in subject pools. In experiments in two Chinese cities and at two campuses in Ethiopia, we show that within-country differences are negligible. Differences between the two countries, on the other hand, are large.
We analyze a task-assignment model in which a principal assigns a task to one of two agents depending on future states. If the agents have concave utility, the principal assigns the task to them contingent on the state. We show that if the agents are loss averse, a state-independent assignment - assigning the task to a single agent in all states - can be optimal even when the principal can write a contingent contract at no cost. (c) 2013 The Authors. Published by Elsevier B.V. All rights reserved.
Organizational justice has the potential to create powerful benefits for organizations and employees alike.
These include greater trust and commitment, improved job performance, more helpful citizenship behaviors,
improved customer satisfaction, and diminished conflict. We demonstrate the management of
organizational justice with some suggestions for building fairness into widely used managerial activities.
These include hiring, performance appraisal, reward systems, conflict management, and downsizing.
The correlates of distributive, procedural, and interactional justice were examined using 190 studies samples, totaling 64,757 participants. We found the distinction between the three justice types to be merited. While organizational practices and outcomes were related to the three justice types, demographic characteris-tics of the perceiver were, in large part, unrelated to perceived justice. Job performance and counterproductive work behaviors, considered to be outcomes of perceived justice, were mainly related to procedural justice, whereas organizational citizenship behavior was similarly predicted by distributive and procedural justice. Most satisfaction measures were similarly related to all justice types. Although organizational commitment and trust were mainly related to procedural justice, they were also substan-tially related to the other types of justice. Findings from labora-tory and field studies are not always in agreement. Future research agendas are discussed. ᭧ 2001 Academic Press The study of work-place justice has been proliferating in recent years. Whereas early studies on justice were conducted in the early 1960s (Adams, 1963, 1965), the majority of studies on justice in organizations were published We thank Dr. Dan Ilgen and three anonymous reviewers for their helpful and constructive comments. Address correspondence and reprint requests to Yochi Cohen-Charash, Institute
We review the experimental evidence on risk aversion in controlled laboratory settings. We review the strengths and weaknesses of alternative elicitation procedures, the strengths and weaknesses of alternative estimation procedures, and finally the effect of controlling for risk attitudes on inferences in experiments.
We analyze optimal contracts in a hierarchy consisting of a principal, a supervisor and an agent. The supervisor is either neutral or altruistic towards the agent, but his preferences are private information. In a model with two supervisor types, we find that the optimal contract may be very simple, paying the supervisor a flat wage independent of his type and his evaluation of the agent's effort. Such a contract induces the neutral type of supervisor to report the agent's performance truthfully, while the altruistic type reports favorably independent of performance. Accordingly, overstated performance (leniency bias) may be the outcome of an optimal contract under informational asymmetries.
Making use of a subjective performance appraisal system, it is a well-established fact that many supervisors tend to assess the employees too good (leniency bias) and that the appraisals hardly vary across employees of a certain supervisor (centrality bias). We explain these two biases in a simple theoretical model and discuss determinants of the size of the biases.
The tendency of supervisors to escalate their commitment of a previously expressed opinion by biasing performance ratings was examined in the context of a "real" organization. The hypotheses of the study were (a) that supervisors who participate in a hiring or promotion decision and agree with the eventual decision would positively bias subsequent performance appraisal ratings for that employee, and (b) that supervisors who participate in the original decision but disagree with the decision would bias subsequent performance appraisal ratings in a negative direction. Cases in which the supervisor had not participated in the hiring or promotion decision were used as a control condition. The study was conducted in a large public-sector organization with a sample of 354 clerical employees. Data provide strong support for both hypotheses, demonstrating both positive and negative escalation biases. The implications of these findings for research on escalation and for organizational policy are discussed. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Tested the possibility that liking may influence rating accuracy by operating as an integral dimension in 66 undergraduate raters. Ss evaluated vignettes of 4 professors; liking was manipulated with 40 trait terms (e.g., amusing, greedy, bashful) that engendered different liking levels, but had little implication for professor performance. Results indicate a significant effect on rating accuracy, suggesting that liking is an integral dimension that is difficult to separate from performance dimensions. Results support the potential importance of affect in appraisal. (40 ref) (PsycINFO Database Record (c) 2012 APA, all rights reserved)
The most ubiquitous method of performance appraisal is rating. Ratings, however, have been shown to be prone to various types of systematic and random error. Studies relating to performance rating are reviewed under the following headings: roles, context, vehicle, process, and results. In general, cognitive characteristics of raters seem to hold the most promise for increased understanding of the rating process. A process model of performance rating is derived from the literature. Research in the areas of implicit personality theory and variance partitioning is combined with the process model to suggest a unified approach to understanding performance judgments in applied settings. (6 p ref) (PsycINFO Database Record (c) 2012 APA, all rights reserved)
A review of the literature revealed that superiors were often thought to be poor sources of performance feedback for their subordinates. A study was undertaken to discover if and when delay and upward distortion of feedback occurred. It was expected that feedback to moderately low performers would be delayed longer and distorted upward more than would feedback to moderately high performers. Further, superiors of moderately low performers were expected to anticipate a less pleasant reaction to feedback by superiors and to believe that their subordinates liked them less than superiors of moderately high performers. A 2–2 design was used with the factors feedback vs no feedback and moderately high vs moderately low subordinate performance. 168 college students served as superiors of a subordinate who was a confederate. Ss monitored and rated their subordinate's performance, then those in the feedback condition gave feedback at a time of their own choosing. All hypotheses were supported except the delay hypothesis. (29 ref) (PsycINFO Database Record (c) 2012 APA, all rights reserved)
The analysis of data from experiments in economics routinely involves testing multiple null hypotheses simultaneously. These different null hypotheses arise naturally in this setting for at least three different reasons: when there are multiple outcomes of interest and it is desired to determine on which of these outcomes a treatment has an effect; when the effect of a treatment may be heterogeneous in that it varies across subgroups defined by observed characteristics and it is desired to determine for which of these subgroups a treatment has an effect; and finally when there are multiple treatments of interest and it is desired to determine which treatments have an effect relative to either the control or relative to each of the other treatments. In this paper, we provide a bootstrap-based procedure for testing these null hypotheses simultaneously using experimental data in which simple random sampling is used to assign treatment status to units. Using the general results in Romano and Wolf (Ann Stat 38:598–633, 2010), we show under weak assumptions that our procedure (1) asymptotically controls the familywise error rate—the probability of one or more false rejections—and (2) is asymptotically balanced in that the marginal probability of rejecting any true null hypothesis is approximately equal in large samples. Importantly, by incorporating information about dependence ignored in classical multiple testing procedures, such as the Bonferroni and Holm corrections, our procedure has much greater ability to detect truly false null hypotheses. In the presence of multiple treatments, we additionally show how to exploit logical restrictions across null hypotheses to further improve power. We illustrate our methodology by revisiting the study by Karlan and List (Am Econ Rev 97(5):1774–1793, 2007) of why people give to charitable causes.
Recent research suggests that loss-framed contracts are an effective instrument for principals to maximize the effort of their agents. Framing effects arise from defining quantity and quality thresholds that vary the salience of losses and gains even while preserving the payoff equivalence of the underlying contract. While plausible interpretations of prospect theory’s loss-aversion insight suggest that a loss frame would lead to more effort, we show that contract thresholds also exert a norm-framing effect on performance that can trump the impact of loss aversion. Loss framing therefore carries a risk: poorly selected thresholds may reduce effort. Principals may prefer to avoid this risk by offering contracts that impose no threshold at all.
Rater errors are systematic biases affecting the distribution of assigned ratings or the construct validity of ratings, typically in the context of performance rating or grading. Rater errors that influence the distribution of ratings include leniency and severity errors and central tendency errors. Leniency and severity errors are the tendency to assign predominantly high or low ratings on a scale; central tendency error derives from a propensity to avoid assigning extreme values, which leaves most ratings near the middle of the scale. Such errors can reduce utility of ratings for decision making. For example, if raises or bonuses are contingent upon ratings and if almost all employees are rated near the top, it becomes impossible to apply contingency. In research, it is feared that these errors can restrict the range of measures, thereby reducing effect sizes.
In this paper we investigate the task the supervisor should be optimally charged with in an agency model in which the principal faces corruption concerns. We highlight a fundamental tradeoff between monitoring the agent's effort choice and auditing it ex-post. Monitoring proves more effective in tackling corruption, since the supervisor sends the report before the profit realization. By taking advantage of the supervisor's uncertainty about the state of nature, the principal can design a compensation scheme which prevents all forms of corruption at a lower cost. Conversely, auditing allows the principal to save the cost of supervision when the profit realization already suffices to set the compensation due to the agent. We show that the choice between monitoring and auditing crucially depends on the supervisor's ability to falsify information and the cost of hiring a supervisor.
The paper performs an in-depth comparison of four incentivised risk elicitation tasks. We show by means of a simulation exercise that part of the often observed heterogeneity of estimates across tasks is due to task-specific measurement error induced by the mere mechanics of the tasks. We run a replication experiment in a homogeneous subject pool using a between subjects one-shot design. Results shows that the task estimates vary over and above what can be explained by the simulations. We investigate the possibility the tasks elicit different types of preferences, rather than simply provide a different measure of the same preferences. In particular, the availability of a riskless alternative plays a prominent role helping to explain part of the differences in the estimated preferences.
Economics has much to do with incentives--not least, incentives to work hard, to produce quality products, to study, to invest, and to save. Although Adam Smith amply confirmed this more than two hundred years ago in his analysis of sharecropping contracts, only in recent decades has a theory begun to emerge to place the topic at the heart of economic thinking. In this book, Jean-Jacques Laffont and David Martimort present the most thorough yet accessible introduction to incentives theory to date. Central to this theory is a simple question as pivotal to modern-day management as it is to economics research: What makes people act in a particular way in an economic or business situation? In seeking an answer, the authors provide the methodological tools to design institutions that can ensure good incentives for economic agents. This book focuses on the principal-agent model, the "simple" situation where a principal, or company, delegates a task to a single agent through a contract--the essence of management and contract theory. How does the owner or manager of a firm align the objectives of its various members to maximize profits? Following a brief historical overview showing how the problem of incentives has come to the fore in the past two centuries, the authors devote the bulk of their work to exploring principal-agent models and various extensions thereof in light of three types of information problems: adverse selection, moral hazard, and non-verifiability. Offering an unprecedented look at a subject vital to industrial organization, labor economics, and behavioral economics, this book is set to become the definitive resource for students, researchers, and others who might find themselves pondering what contracts, and the incentives they embody, are really all about.
We discuss the cognitive and the psy- chophysical determinants of choice in risky and risk- less contexts. The psychophysics of value induce risk aversion in the domain of gains and risk seeking in the domain of losses. The psychophysics of chance induce overweighting of sure things and of improbable events, relative to events of moderate probability. De- cision problems can be described or framed in multiple ways that give rise to different preferences, contrary to the invariance criterion of rational choice. The pro- cess of mental accounting, in which people organize the outcomes of transactions, explains some anomalies of consumer behavior. In particular, the acceptability of an option can depend on whether a negative outcome is evaluated as a cost or as an uncompensated loss. The relation between decision values and experience values is discussed. Making decisions is like speaking prose—people do it all the time, knowingly or unknowingly. It is hardly surprising, then, that the topic of decision making is shared by many disciplines, from mathematics and statistics, through economics and political science, to sociology and psychology. The study of decisions ad- dresses both normative and descriptive questions. The normative analysis is concerned with the nature of rationality and the logic of decision making. The de- scriptive analysis, in contrast, is concerned with peo- ple's beliefs and preferences as they are, not as they should be. The tension between normative and de- scriptive considerations characterizes much of the study of judgment and choice. Analyses of decision making commonly distin- guish risky and riskless choices. The paradigmatic example of decision under risk is the acceptability of a gamble that yields monetary outcomes with specified probabilities. A typical riskless decision concerns the acceptability of a transaction in which a good or a service is exchanged for money or labor. In the first part of this article we present an analysis of the cog- nitive and psychophysical factors that determine the value of risky prospects. In the second part we extend this analysis to transactions and trades. Risky Choice Risky choices, such as whether or not to take an umbrella and whether or not to go to war, are made without advance knowledge of their consequences. Because the consequences of such actions depend on uncertain events such as the weather or the opponent's resolve, the choice of an act may be construed as the acceptance of a gamble that can yield various out- comes with different probabilities. It is therefore nat- ural that the study of decision making under risk has focused on choices between simple gambles with monetary outcomes and specified probabilities, in the hope that these simple problems will reveal basic at- titudes toward risk and value. We shall sketch an approach to risky choice that
A prospect theory model combining loss aversion and diminishing sensitivity predicts that the link between incentives framing and effort is ambiguous: small penalties yield higher effort, but isomorphic contracts with large penalties decrease effort. We conduct two experiments (a framed field and a conventional lab experiment) in which economically equivalent contracts are framed as menus of either (i) bonuses, (ii) penalties, or (iii) bonuses and penalties. The experimental results confirm the main intuition of the model as subjects performed best when bonuses and penalties are combined. A follow-up lottery experiment confirms that both loss aversion and diminishing sensitivity influenced the performance.
This study examines the determinants and performance effects of centrality bias and leniency bias. The results show that managers respond to their own incentives and preferences when subjectively evaluating performance. Specifically, information-gathering costs and strong employee-manager relationships positively affect centrality bias and leniency bias. The findings also indicate that performance evaluation biases affect not only current performance ratings, but also future employee incentives. Inconsistent with predictions based on the agency perspective, the results show that managers' performance evaluation biases are not necessarily detrimental to compensation contracting. Although centrality bias negatively affects performance improvement, the evidence does not reveal a significant negative relation between leniency bias and performance. Rather, leniency bias is positively associated with future performance, which is consistent with the behavioral argument that bias can improve perceived fairness and, in turn, employee motivation.
This paper discusses aspects of recruiting subjects for economic laboratory experiments, and shows how the Online Recruitment System for Economic Experiments can help. The software package provides experimenters with a free, convenient, and very powerful tool to organize their experiments and sessions.
The role of social and situational influences in the performance-rating process has received relatively little research attention yet merits increased attention. Although there has been acknowledgment of the role of social and situational factors in shaping rater cognition and evaluation, research has typically proceeded in a piecemeal fashion, isolating a single variable at a time. Such an approach fails to recognize that performance rating is a process with multiple social and situational facets that need to be considered simultaneously. In the present study, we tested a model of the performance-rating process, employing several social and situational variables that have been infrequently investigated and typically not in conjunction with one another. Results indicated support for the overall model and specific links within it. Implications of the results for performance-rating research are discussed. There is perhaps not a more important human resources system in organizations than performance evaluation. Supervisors' ratings of subordinates' performance represent critical decisions that are key influences on a variety of subsequent human resources actions and outcomes. Indeed, this pivotal role of performance evaluation has promoted systematic efforts to develop a more informed understanding of the performance-rating process. Landy and Farr (1980) issued a call for research investigating the cognitive processes underlying performance appraisal decisions. Although the process focus has generated considerable research concerning various components of performance-rating decisions, more comprehensive investigations incorporating several of those components has been lacking. Furthermore, process-oriented research has been limited by its reliance on laboratory studies (DeNisi & Williams, 1988). Whereas the cognitive processes involved in performance-rating decisions can be well illuminated in laboratory studies, the "quiet" nature of laboratory studies often does not match the "noisy" context in which performance-rating decisions are actually embedded (Lord & Maher, 1989).
Many scholars have emphasized the importance of subjective performance evaluations in employment relationships to provide
employees with effort incentives. This article examines when delegating subjective appraisals to managers is optimal. Delegation
facilitates more accurate performance evaluations but can also lead to collusion. To prevent collusive behavior, I show that
firms need to reduce effort incentives for employees and leave managers with economic rents through higher compensations.
Despite these costs of preventing collusion, I find that delegation can be optimal.
We examine how subjective performance evaluations are influenced by
the level and controllability of an accompanying measure of a separate performance
dimension. In our experiment, supervisors evaluate the office administration performance
of a hypothetical subordinate. We find that supervisors’ subjective evaluations
are directionally influenced by an accompanying objective measure of sales performance,
even after excluding participants who perceive informativeness across
measures. Consistent with concerns for fairness and motivation, we also find an
asymmetric uncontrollability effect—supervisors’ evaluations are higher when an
uncontrollable factor decreases the subordinate’s sales (i.e., they compensate for bad
luck), but are not lower when the uncontrollable factor increases the subordinate’s sales
(i.e., they do not punish for good luck). This evidence suggests that supervisors use
discretion provided to evaluate performance on one task to adjust for perceived
deficiencies in the evaluation of performance on other tasks. Our study integrates
theories of cognitive bias and motivation, highlighting the need to consider the potentially
interactive effects of different performance measures in multi-task settings.
Economists and psychologists have developed a variety of experimental methodologies to elicit and assess individual risk attitudes. Choosing which to utilize, however, is largely dependent on the question one wants to answer, as well as the characteristics of the sample population. The goal of this paper is to present a series of prevailing methods for eliciting risk preferences and outline the advantages and disadvantages of each. We do not attempt to give a comprehensive account of all the methods or nuances of measuring risk, but rather to outline some advantages and disadvantages of different methods.
We perform a comparative analysis of five incentivized tasks used to elicit risk preferences. Theoretically, we compare the elicitation methods in terms of completeness of the range of the estimates as well as their precision, the likelihood of triggering loss aversion, and problems arising when multiple choices are required. Using original data from a homogeneous population, we experimentally investigate the distribution of estimated risk preferences, whether they differ by gender, and the complexity of the tasks. We do so using both non-parametric tests and a structural model estimated with maximum likelihood. We find that the estimated risk aversion parameters vary greatly across tasks and that gender differences appear only when the task is more likely to trigger loss aversion.
We study an infinitely repeated principal-agent model with subjective evaluations. We compare the surplus in efficiency-wage equilibria and in bonus-payments equilibria. The agent receives a constant wage and is motivated by the threat of dismissal in efficiency-wage equilibria. The agent receives a bonus and quits the relationship after disagreements between his self-evaluation and the principal's performance appraisal in bonus-payments equilibria. We construct a class of equilibria with bonus payments that approach efficiency as patience increases. In contrast, payoffs from efficiency-wage equilibria are bounded away from the Pareto-payoff frontier for any discount factor. (JEL D82, J33, J41)
Empirical work testing for a trade-off between risk and incentives has had, at best, mixed success. This article provides two simple reasons, associated with subjectivity of performance appraisals, why we might not expect to see any negative relationship. Both reasons relate to empirically observed problems associated with monitoring: (i) supervisors sometimes bias their evaluations based on their personal feelings toward their subordinates, and (ii) supervisors will sometimes offer evaluations that reduce their costs. These aspects of monitoring are ignored in the standard model and can reverse the usual negative trade-off between risk and incentives.
A taxonomy is presented that categorizes theories of organizational justice with respect to two independent dimensions: a reactive-proactive dimension and a process-content dimension. Various theories within each of the four resulting categories are identified. The implications of the taxonomy are discussed with respect to clarifying theoretical interrelationships, tracking research trends, and identifying needed areas of research.
z-Tree (Zurich Toolbox for Ready-made Economic Experiments) is a software for developing and conducting economic experiments. The software is stable and allows programming almost any kind of experiments in a short time. In this article, I present the guiding principles behind the software design, its features, and its limitations
This study investigated the predictability of rater leniency from scores on an instrument designed to measure rater discomfort with performance appraisal situations. The 20-item Performance Appraisal Discomfort Scale (PADS) was administered twice to a sample of 178 undergraduate business students who rated performance of group members on three group projects under conditions designed to emulate features of actual appraisal situations. Results supported the notion that rater leniency is stable and predictable from PADS scores. Also, principal-component extracted and varimax rotated factors of the PADS were interpreted as corresponding to four distinct situational demands placed on raters in the course of performance appraisal. The results of this limited demonstration of the validity of PADS suggest that future application of the PADS or some like instrument may be useful for enhancing the validity of appraisal ratings and feedback utility.
This paper reviews literature of psychometric properties of self-appraisals of work performance. It summarizes evidence of leniency, variability, halo, bias, and construct validity. Comparisons with appraisals by supervisors, peers, and subordinates suggest that self-appraisals tend to show more leniency, less variability, and less discriminant validity. Different factor structures have been found among self, supervisor and peer-ratings. On the other hand, self-appraisals showed less halo. Self-appraisals were significantly correlated with other sources in some studies and failed to correlate in many others. Existing data do not allow any conclusion whether the quality of self-appraisals is a function of scale format, amount or rater training, type of judgment, or purpose of appraisal. The effects of the observed psychometric qualities of self-appraisals on various applications are discussed. Problems may exist when they are used for administrative decision making, diagnosis of training needs, applied criterion measurement, measurement of constructs in basic research, or for selection purposes.
Reviews of self–supervisor, self–peer, and peer–supervisor ratings have generally concluded that there is at best a modest correlation between different rating sources. Nevertheless, there has been much inconsistency across studies. Accordingly, a meta-analysis was conducted. The results indicated a relatively high correlation between peer and supervisor ratings (ρ= .62) but only a moderate correlation between self-supervisor (ρ= .35) and self-peer ratings (ρ= .36). While rating format (dimensional versus global) and rating scale (trait versus behavioral) had little impact as moderators, job type (managerial/professional versus blue-collar/service) did seem to moderate self-peer and self-supervisor ratings.
We experimentally study individual reactions to subjective performance evaluations in a laboratory experiment with agents working on and principals benefiting from a real effort task. We show that agents tend to sanction principals for feedback below the agents' self-evaluation even if the principals' feedback does not influence the agents' earning. Hence, feedback which falls short of the agents' self-evaluation can be inter-preted as an unkind act that triggers a negatively reciprocal response not only if the feedback determines an agent's earnings but also if it lacks monetary consequences. We propose a principal-agent model that accommodates this kind of payoff independent reciprocity and analyze the impact of the agent's sensitivity to negative feedback, the conflicts created in response to negative feedback, and the quality of subjective evalu-ations on behavior, optimal contracts, and welfare. JEL:D01; D02; D82; D86; J41.
This paper experimentally tests the predictions of a principal-agent model in which the agent has biased beliefs about his ability. Overconfident workers are found to earn lower wages than underconfident ones because they overestimate their expected payoff, and principals adjust their offers accordingly. Moreover, the profit-maximizing contract distorts effort by varying incentives according to self-confidence, although only the most successful principals use this strategy. These findings have implications for the labor market; in particular, self-confidence is often correlated with gender, implying that principals would prefer to hire men over women simply because they are more overconfident.