with Aaron Sojourner and Akhmed Umyarov
[Download] [Replication Tools]
In two experiments, we examine the effects of employer reputation in Amazon Mechanical Turk, an online labor market in which employers may decline to pay workers while keeping the work product. In the first experiment, a research assistant who is blinded to reputation performs tasks posted by employers with good, bad, or no online reputations. Results confirm the value of reputation; due to shorter tasks and rarer nonpayment, effective wages among good reputation employers are about 40 percent greater than those for neutral- or bad-reputation employers. In the second experiment, we create multiple employer identities endowed with different exogenously introduced reputations. We find that employers with good reputations attract workers at nearly twice the rate as those with bad reputations with no discernible difference in quality. We interpret these results through the lens of an equilibrium search model in which the threat of a bad reputation deters employers from the abuse of authority even in the absence of contractual protections of workers. The results demonstrate the value of employer reputation systems for workers and employers, and thus for labor market efficiency.
with Sima Sajjadiani
This paper examines whether plantwide incentives for quality actually have incentive effects, a puzzle because high level incentives are common but potentially vulnerable to freeriding and other obstacles. We exploit a natural experiment embedded in a common but economically peculiar feature of self-funded incentives distributed under a gainsharing plan: the plant's variable costs create exogenous variation in the eligibility and magnitude of incentives. Although incentives are salient and potentially large, and despite sufficient statistical power to detect differences quality-related item returns within 1% (on an 8.5% overall rate), we find evidence that item returns stays the same in weeks there is an incentive for quality, and may actually rise when incentives are small. We interpret this as evidence that the strict incentive effects of such plantwide programs are ineffective or possibly counterproductive, at least by their intended mechanism. Rather, the effectiveness of gainsharing plans may rest on programs typically adopted simultaneously or through a sense of shared ownership that is imparted regardless of whether employees are eligible for incentives.
with Frank Levy and Raimundo Esteva
Most published estimates of the economic return to college rest on a series of best-case assumptions that often overstate returns and, most importantly, obscure differences in return across different institutions. We simulate the economic return to college under more realistic assumptions using U.S. Census data combined with administrative data from the more selective University of California system and the less selective California State University system. Specifically, we adjust for delayed graduations, the probability of dropping out, progressive taxes on earned income, and risk aversion. We perform a bounding exercise for ability bias. These each reduce expected returns to a Bachelor’s degree. Contrary to prior “best case” estimates, and under reasonable bounds for the ability bias, we find that the return to a college degree in 2010 could be less than the interest on unsubsidized Stafford loans. Returns are particularly modest for young men at the less-selective CSU system, largely due to high dropout rates, delayed graduation, and a lower effect on labor force participation compared to women. Our analysis begins to bridge the gap between standard estimates of the economic return to college and the institutional performance metrics reported in the Obama Administration’s College Scorecard.
Benson, Alan, Danielle Li, and Kelly
Shue. 2014. “When Good Tournaments Make Bad Matches: Evidence of the Peter
Principle in Sales.” In Progress.