We show that the increasingly popular use of subjective assessments of employee “potential”contributes to gender gaps in promotion and pay. Using data on management-track employees from a large retail chain, we find that women receive substantially lower potential ratings despite receiving higher job performance ratings. Differences in potential ratings account for 30-50% of the gender promotion gap. Women’s lower potential ratings do not appear to be based on accurate forecasts of future performance: women outperform male colleagues with the same potential ratings in terms of their future performance ratings, both on average and conditional on promotion. Yet, even in these cases, women’s subsequent potential ratings remain low, suggesting that firms persistently underestimate the potential of female employees.

Using administrative records from a large national US retailer, we find managers learn to discriminate ``on the job'' as they gain experience hiring workers of different races. First, we find that idiosyncratically positive and negative previous experiences with black hires seed the race of future hires, consistent with managers updating their beliefs about the productivity of these workers based on their hiring history. Second, the degree of updating is larger for black than white workers, consistent with managers having weaker priors. Third, because positive experiences increase black hiring and negative ones decrease it, negative biases are slower to self-correct than positive ones. Fourth, these dynamics, combined with the fact that black workers are already in a minority, yield particularly large and protracted declines in black hiring following a manager's negative experiences. Our results suggest that managers develop biased beliefs from endogenous learning about racial groups, which systematically disadvantages minority workers.

We examine whether the reasons that employers provide for rejecting job candidates affect their likelihood of applying for future positions, and differential responses by gender. Through a randomized controlled field experiment among job candidates rejected for positions by a staffing company, we find that relative to men,  women are less likely to apply for future positions after being rejected. Furthermore, we find that this gap is nearly eliminated by informing applicants that they were rejected for ``fit'' rather than ``quality'' or by providing no reason for the job rejection. We present survey evidence that workers view the quality message as demeaning and the no-reason message as ambiguous. Our findings lend support for hypotheses that women have relative tastes for non-competitive and transparent application procedures, and that gender disparity in job search persistence may be reduced by framing rejection in terms of fit.

TBA.

TBA.

TBA.

Using data on 36,949 newly-hired commission-based salespeople at a major U.S. retailer, we find that white, black and Hispanic managers within the same store are more likely to hire workers of their own race. This may be caused by managers' intrinsic taste for hiring same-race applicants (taste-based discrimination), or by their greater ability to screen same-race applicants (screening discrimination). We derive the testable implications of these models for the mean and variance of productivity and show that white and Hispanic hiring is consistent with the screening hypothesis. We find little evidence of complementarities between supervising (rather than hiring) managers or employee discrimination in this setting.

This study proposes a unified, dynamic framework based on Turnover Event Theory (TET) to evaluate the effects of dismissals, layoff announcements, and voluntary turnover on subsequent work unit voluntary turnover. Applying our approach to 1,620 retail stores over 22 months, we show that modeling exit events as a dynamic and interdependent system adds to our ability to predict subsequent human capital outflow. Dismissals had the weakest total effect on subsequent voluntary turnover, layoff announcements had the strongest and most immediate effects, and voluntary turnovers had moderate but lasting effects. We also find that these three exit reasons each exhibit a distinct pattern of subsequent turnover intensity and longevity. Based on characteristics of work units in our setting, our results correspond to a cumulative worker-level average effect of 0.17 quits following dismissals, 0.23 quits following quits, and 2.2 quits following a layoff. We find that these “multiplier” effects are concentrated among workers of similar performance: high performer exits beget high performer quits, just as low performer exits beget low performer quits. Our findings suggest analyses offering individual-level estimates of turnover will generally underestimate the broader, work-unit level consequences of individual exit events on similar workers.

Startup ventures often require experienced employees, such as professional managers, in order to grow. However, risk considerations may discourage such employees – who often have stable, high-paying jobs – from joining new ventures. We hypothesize (a) based on the portfolio effect stemming from being in a dual-career household, having a spouse whose career is prioritized within the couple mitigates risk as a barrier to joining startups; and (b) due to gender norms, this effect is pronounced for men compared to women. Using survey data from married professional managers, we find support for these hypotheses. Our findings demonstrate how dual career households can affect the pool of potential talent for startup ventures, which has implications for understanding dual careers, startup venture growth, and regional entrepreneurial ecosystems.

Does monitoring workers improve or impair their productivity? Existing studies offer conflicting predictions. Using personnel and operational data from an Indian garment manufacturing plant, we examine how an RFID monitoring intervention on three of the plant's twelve production lines affected productivity. We find that the effect of monitoring varied by the complexity of the work performed. Using variation in work complexity both across and within production lines, we find that monitoring significantly increased productivity for simple work but significantly decreased productivity for complex work. We contribute to research on monitoring and productivity by demonstrating how key job characteristics that make work meaningful, such as complexity, can moderate the effect of monitoring on productivity by affecting workers' intrinsic motivation. Results also suggest that not only does the Hawthorne effect exist, but its direction can be positive, negative, or neutral depending on work complexity.

In three experiments, we examine the effects of employer reputation in an online labor market (Amazon Mechanical Turk) in which employers may decline to pay workers while keeping their work product. These three experiments test the value of the employer reputation system for workers, employers, and the market. Specifically, in audit study of employers by a blinded worker, we find that working only for good employers yields 40% higher wages. Second, in an experiment that varies reputation, we find that good-reputation employers attract work of the same quality but at twice the rate as bad-reputation employers. Lastly, we exploit the natural experiment of instances when the reputation system servers are down, and find that the reputation system is serving the market by attracting work to small, good employers who appear to rely on the system to attract workers, and apparently away from the largest and best-known among good employers. This is the first clean, field evidence that employer reputation serves as a collateral against opportunism in the absence of contract enforcement in online markets.

The best worker isn't always the best candidate for manager. In these cases, do firms promote the best potential manager or the best worker in their current job? Using microdata on the performance of sales workers at 214 firms, we find evidence consistent with the Peter Principle: firms prioritize current job performance when making promotion decisions at the expense of other observable characteristics that better predict managerial quality. We estimate that the costs of managerial mismatch are substantial, suggesting that firms make inefficient promotion decisions or the incentive benefits of emphasizing current performance is also high.

We develop and test a theoretically informed and generalizable empirical framework for evaluating the performance gap between internally and externally hired workers. First, human capital theory predicts that internal hires will be immediately more productive than external hires. Second, contextual learning predicts that internal hires will be more productive with time. Finally, theories of commitment, which are rarely applied to this literature, predict that internal advancement enhances retention among high performers (“positive retention”). Applying a general empirical framework for quantifying the relative contributions of these mechanisms to a retailer with 109,063 commissioned salespeople and their 12,931 managers, we find that the gap in our setting is primarily driven by positive retention: High performers and internal hires are less likely to quit, and crucially, high-performing internal hires are especially unlikely to quit. When high-performing internal hires do quit, they tend to cite reasons unrelated to work rather than advancement opportunities. By typically examining performance and retention in isolation, researchers and organizations may be underestimating the importance of internal advancement as a means of retaining of high performers. 

Shared bonus pools, in which a worker’s bonus depends both on a worker’s share of the pool (which serves as the incentive) and on the size of the pool (which is largely outside of the worker’s control), are a common method for distributing bonus pay. Using variation in the size of the bonus pool generated by a manufacturing plant’s gainsharing plan, which varies incentives for quality and worker engagement, we evaluate the conditions under which incentives distributed from bonus pools have incentive effects. Overall, results are cautionary: the evidence suggests gainsharing’s benefits operate outside of the incentive channel, and incentives can backfire if they are too small or too diluted by group performance metrics. Lastly, we illustrate how random variation in the size of bonus pools offers researchers a powerful, readily available, and underused tool for studying how workers respond to the availability and strength of incentives.

This paper explores intermediary agency problems, specifically the use and misuse of authority and incentives in organizational hierarchies. I offer a principal-agent-supervisor model inspired by sales settings. Through it, I propose organizations delegate authority over rank-and-file workers to immediate managers, even though the performance of rank-and-file workers is known to the firm, because managers' private information allows them to distinguish ability from luck. The model yields the result that managers on the cusp of a quota have a unique personal incentive to retain and provide downward quota adjustments to poor performing subordinates, allowing me to identify of the interests of the manager as separate from the firm. I parametrically estimate the model using detailed person-transaction-level microdata from 244 firms that subscribe to a “cloud”-based service for automating transaction processing and compensation. I estimate 13-15% of quota adjustments and retentions among poor performers are explained by the managers' unique personal interest in meeting a quota. I draw upon organizational research to suggest how performance analytics software may promote organizational efficiency.

I present a model of couples' job search whereby men segregate into occupations that are geographically clustered (such as nuclear engineers) and women sort into occupations that are geographically dispersed (such as jobs in health, business support, or education) in advance of marriage and in anticipation of future co-location problems. Using the Decennial Census, I calculate a measure of geographic clustering-the share of members in that occupation that would need to relocate for that occupation to have the same employment-to-population ratio in all US metropolitan areas. Results confirm men segregate into geographically-clustered occupations, and that these occupations involve more-frequent early career relocations for both sexes. I also find that the minority of the men and women who depart from this equilibrium experience later marriage, higher divorce, and lower earnings. Results are consistent the theory that marriage and mobility expectations foment a self-fulfilling pattern of occupational segregation with individual departures deterred by earnings and marriage penalties.

I present a theory of couples’ job search whereby women sort into lower-paying geographically dispersed occupations due to expectations of future spouses’ geographically clustered occupations and (thereby) inability to relocate for work. Results confirm men segregate into clustered occupations, and that these occupations involve more frequent early career relocations for both sexes. I also find that the minority of the men and women who depart from this equilibrium experience delayed marriage, higher divorce, and lower late-career earnings. Results corroborate the theory’s implication that marriage and mobility expectations foment a self-fulfilling pattern of occupational segregation with individual departures deterred by earnings and marriage penalties.

Using a unique data set of hospitals’ direct financial support to nursing schools and faculty in 2008 and 2010, this paper finds that firms may pay for technologically-general skill training that is made de facto-specific by search frictions. Because this support is clearly general, clearly paid by the firm, and information asymmetries appear minimal, it overcomes the typical dilemmas of empirical research in firm-sponsored general education. Interviews and existing studies suggest hospitals extract rents on incumbent nurses, and an oligopsony model is proposed whereby hospitals sponsor training when they employ a sufficient share of local nurses. Consistent with the model, surveys conducted in 2008 and 2010 find that hospitals employing a greater share of its MSA’s registered nurses are indeed more likely to support local schools and faculty financially, net of size and other institutional controls. Implications for human capital theory, monopsony, and nursing manpower are discussed.

The authors use experiences with interest arbitration for police and firefighters under New York State’s Taylor Law from 1974 to 2007 to examine the central debates about the effects of this form of arbitration on collective bargaining. They draw on old and new data to compare experience with interest arbitration in the first three years after it was adopted with experiences from 1995 to 2007. They find that no strikes have occurred under arbitration, rates of dependence on arbitration declined considerably, the effectiveness of mediation prior to and during arbitration remained high, the tripartite arbitration structure continued to foster discussion of options for resolution among members of the arbitration panels, and wage increases awarded under arbitration matched those negotiated voluntarily by the parties. Econometric estimates of the effects of interest arbitration on wage changes in a national sample suggest wage increases between 1990 and 2000 in states with arbitration did not differ significantly from those in states with non-binding mediation and factfinding or states without a collective bargaining statute. The length of time required to complete the arbitration process increased substantially and several critical employment relations issues facing the parties have not been addressed within the arbitration system. The authors suggest these findings should be considered by both critics and supporters of proposals to include a role for interest arbitration in national labor policy.