Research
Good Interviewers Matter with T. Barr, K. Gautier, and L. Munasinghe
Interviews present a puzzle: Research suggests they are of little value in screening hires, and yet they remain pervasive. Using a longitudinal dataset from a large call center company, we find that interviewers vary in quality along three dimensions of performance: making offers, convincing candidates to join the company, and screening for durable and productive hires. The variation in interviewer performance is larger for the first two dimensions than for screening. Furthermore, interviewers improve over tenure on the first two dimensions, while their ability to screen remains stagnant. Interviewers who are good at making offers are also good at convincing the candidate to accept them, and they do so without sacrificing quality of hire. We find that the company could realize cost savings of over 40 percent of its hiring and recruiting budget by reassigning interviews to their best interviewers. Thus, interviewer performance, especially in often-ignored dimensions, can have a major impact on search and matching of employers and employees.
This paper investigates the effect of information transmitted by referrals on search efficiency. We test several implications of our model using data from a call center company that contain rich information on applicants, employees, and referrers. The joint estimation of job offers, acceptances, turnover, and performance allows us to identify the contribution of referrals to search efficiency from the differences in expected performance between referred and non-referred applicants. The results show that referrals induce selection on unobservables at the job offer stage, which in turn drives referred employees to perform better, receive early promotions and stay longer on the job.
"Matching in Closed-Form: Equilibrium, Identification and Comparative Statics" with A. Galichon; Economic Theory (2016)
This paper provides closed-form formulas for a multidimensional two-sided matching problem with transferable utility and heterogeneity in tastes. When the matching surplus is quadratic, the marginal distributions of the characteristics are normal, and when the heterogeneity in tastes is of the continuous logit type, as in Choo and Siow (2006), we show that the optimal matching distribution is also jointly normal and can be computed in closed form from the model primitives. Conversely, the quadratic surplus function can be identified from the optimal matching distribution, also in closed-form. The analytical formulas make it computationally easy to solve problems with even a very large number of matches and allow for quantitative predictions about the evolution of the solution as the technology and the characteristics of the matching populations change.
"Incentives to Work or Incentives to Quit?" (submitted)
Pay incentives affect not only effort choice but also turnover and thereby the quality mix of the workforce. This paper investigates how considerations about the quality mix shape pay policy and profits within a structural model of effort choice, symmetric learning about match quality, and turnover. Using unique data from a US call center, I estimate the model in two steps adapting estimation methods for dynamic structural models to the analysis of employment dynamics. Then, I consider three classes of contracts: (1) compensation depends only on current output; (2) compensation depends also on past output; (3) compensation depends on all available information and may vary with tenure. The results indicate that experimentation to improve the quality mix is a primary concern that affects the optimal contracts in the three classes. Experimentation requires high turnover which is also associated with the destruction of accumulated specific human capital. The trade-of between experimentation and the accumulation of specific capital determines the characteristics of the optimal contract.
"Estimating the Effects of Incentives When Workers Learn About Their Match Quality" (submitted)
Employees often learn about their ability while working, and the resulting beliefs interact with pay incentives to shape employment outcomes. This paper investigates the issue within a model that incorporates learning about ability, effort, turnover, and variation in pay incentives. The empirical analysis is based on unique data from a US call center. The results show that turnover, due to selective attrition and experimentation, is a major channel through which incentives affect performance, profits, and individual welfare. Simulating the estimated model shows that neglecting the impact of beliefs and incentives on separation decisions makes estimates of the effect of incentives on effort twice as big as they should be.
"Dynamism and Economic Performance: A Preliminary Report" with E. Phelps
This paper is a step in the identification of the determinants of performance in developed economies. A basic finding is that differences in beliefs and attitudes concerning the workplace account for most of the differences in performance. Another finding is that there are four main dimensions, or categories, in which individuals differ in workplace beliefs and attitudes. We find that the most important of them for performance, as we might have suspected, is the one we labeled vitalism. This is followed by materialism (as measured by preference for consumption over leisure), social trust, and self-reliance. We also find that differences in institutional environment have little explanatory power once we control for beliefs and attitudes.
Research in Progress:
“Effect of Economic Beliefs and Attitudes on Economic Performance,” with E. Phelps (in preparation)
"Do Interviewers Affect Hiring and Employment Outcomes?" with T. Barr, K. Gautier, and L. Munasinghe (in preparation)
“Dynamism and Economic Performance: Preliminary Report” with E. Phelps; working paper
“Implicit Incentives of Referrals,” with L. Munasinghe and T. Barr
“Statistical Discrimination during the Hiring Process,” with L. Munasinghe and T. Barr
“The Perils of Estimating the Effects of Referrals,” with L. Munasinghe and T. Barr
“Matching with Supervisors, Co-workers, or Technology?” with L. Munasinghe