I examine the effects of providing workers with relative performance information (RPI) on employers' promotion decisions and the impact of those decisions on worker performance. In my experimental setting, the job after promotion requires higher-level abilities than the current job. I find that workers increase their effort to improve their current job performance after a promotion opportunity is announced because they expect this to increase their chances of promotion, even though the new task requires higher-level abilities. Moreover, because employers anticipate that workers who have RPI will react negatively if they see that the best current job performer is not promoted, employers promote the best current job performer, rather than the worker best suited for the next job, more often when workers have RPI than when they do not. Consistent with the Peter Principle, I find that when workers have RPI, the promoted worker's performance is lower after promotion because the promoted worker lacks the ability to perform the new job well. Finally, in a supplemental experiment, I find that providing workers with feedback on their abilities to perform the next job, in addition to current job RPI, improves the effective sorting of workers, but it comes at the cost of reduced promotion-based incentives. My results suggest that, notwithstanding the benefits documented in previous studies, RPI also imposes potential costs that firms should take into account when deciding whether to provide workers with RPI. JEL Classifications: M41; M51; M52. Data Availability: Contact the author.

We experimentally examine how nonprofit pay levels affect nonprofit managers' misreporting and how different types of penalties deter such behavior. Absent any penalties for misreporting, we find similar levels of, but different mechanisms for, misreporting by managers who select into lower- versus higher-paying nonprofits. Specifically, lower-paying nonprofits attract mission-focused managers who misreport to advance their nonprofit's mission and justify doing so by their personal sacrifice of wealth, whereas higher-paying nonprofits attract a majority of self-focused managers who justify misreporting by focusing on their personal outcomes. Because managers who select into their nonprofit at different pay levels have different motivation and justification for misreporting, we predict and find that penalizing the nonprofit is more effective in deterring misreporting in lower-paying nonprofits, while penalizing the manager is more effective in higher-paying nonprofits. Our study contributes to theory on the behavioral drivers of nonprofit misreporting and informs practice of ways to deter it.

Professional development and training are essential for organizational growth and success, but resource constraints often force managers into a dilemma: Should they train their lower performing or higher performing employees? While lower performers with greater skill gaps can benefit more from training than higher performers, managers must also consider employees’ fairness concerns. Using an interactive experiment, we predict and find that how employees react to their manager's allocation of training resources depends on the prevailing fairness norms in the work environment. When egalitarian norms dominate, such as when employees have little control over their job tasks, lower performers take the receipt of resources for training for granted and react negatively when the manager allocates those resources to higher performers instead. In contrast, when meritocratic norms dominate, such as when employees have more control over their job tasks, higher performers react more negatively than lower performers to not receiving training resources. Additional experiments suggest that these fairness‐norm–dependent reactions are most pronounced under tournament incentives and may disappear under individual performance‐based pay. Overall, our study highlights the importance for managers to consider both employee productivity and the prevailing fairness norms of the work environment when allocating scarce training resources.

Managers often face the choice between promoting an internal employee and hiring an external candidate. Using an interactive experiment, we examine the drivers of managers’ promote/hire decisions and internal employees’ behavior before and after those decisions. Consistent with gift exchange theory, we find that employees exert costly effort to increase the chance of being promoted, and they raise their effort level as the promote/hire decision becomes imminent. Managers respond by promoting those who exert high effort, despite employees’ inferior ability compared to external candidates. Results suggest that managers view employees’ past effort as both a gift to reciprocate and a signal of their future effort. Moreover, we find that managers are more likely to promote internally rather than hire externally under a less precise performance measurement system, and this result is driven by managers who observe low employee output. Finally, we find that total effort is significantly higher when managers promote internally versus hire externally.

We investigate the consequences for non-promoted executives (NPEs) in CEO tournaments. We find that NPEs' total incentives decrease following the end of a tournament based on evidence of their reduced future promotion prospects and limited adjustments to their compensation. Consistent with the theory that NPEs leave in response to this loss in incentives, results indicate that turnover is higher for NPEs who: (1) are ex ante more competitive for promotion, (2) compete in open tournaments without an heir apparent versus closed tournaments with an heir apparent winner, and (3) compete in tournaments with an outsider versus insider winner. Departed NPEs' subsequent career outcomes suggest that the labor market assesses NPEs who leave after open tournaments more favorably than those who leave after closed tournaments and tournaments with an outsider winner. Overall, evidence suggests that promotion tournaments can weed out low-quality managers, but also cause the unintended turnover of high-quality managers.

To address the empirical phenomenon that organizations often reward time on the job as an end in itself, we design an experiment in which participants solve anagram puzzles, manipulating whether a compensation pool generated from the output of paired workers is allocated based on the individual inputs of relative time spent or on the individual outputs of puzzles solved. Relative to an output-based allocation, we find that an input-based allocation leads participants to spend more time on the task. However, when paired participants have widely different abilities, an input-based allocation also leads to less effort intensity, defined as puzzles solved per unit of time spent. We attribute these findings to fairness considerations, an interpretation we corroborate in a second experiment with purely individual incentives that finds the same effort duration advantage of input-based pay, but no offsetting disadvantage in effort intensity.

This study investigates how real activities management (RAM) by employees to meet internal targets set by managers affects their performance, and how managers’ awareness of employees’ opportunity to engage in RAM moderates this effect. Using two experiments, we predict and find that when managers are unaware of employees’ RAM opportunity, employees who face miscalibrated targets use RAM to restore their own incentives to work hard, resulting in increased employee performance. Counter to conventional wisdom, we also predict and find that making managers aware of employees’ RAM opportunity hurts employee performance. Such managers overestimate employees’ RAM and adjust targets aggressively such that they become more difficult to achieve, leading to lower employee motivation and less trusting employee-manager relations. Overall, our findings show that the opportunity to engage in RAM can have the indirect benefits of motivating higher employee performance, but increasing managers’ awareness of this has negative performance and relational effects. Data Availability: The data and research instrument are available from the authors upon request.