The Incentives Platform team at Lyft has developed a platform for applying new methodologies at the intersection of causal inference, machine learning, and reinforcement learning to problems at scale. We utilize heterogeneous treatment effect algorithms to predict how different users (riders, drivers) will respond to a specific treatment (coupon, incentive, message, etc.). We then can apply various optimization algorithms to choose which users get which treatment while using bandit methodologies to balance an explore/exploit trade-off. This platform dramatically increases the degree to which we can customize the user experience and hit business goals while reducing the operational load of doing so.

Much of the new "gig economy" relies on reputation systems to reduce problems of asymmetric information. In most cases, these reputation systems function well by soliciting unbiased feedback from buyers and sellers. However, certain features of online labor markets create incentives for employers to misreport worker performance. This paper tests whether employers learn about worker productivity from public, subjective, performance reviews using data from a large online labor market. Starting with a simple model of employer learning in the presence of potentially biased reviews, I derive testable hypotheses about the relationship between public information and wages, worker attrition, and contract renewals. I find that these public reviews provide substantial information to the market and that other firms use them to learn about the productivity of workers. I also find evidence that these reviews affect how long workers stay in the labor market. Finally, using data on applications, I provide evidence of a mechanism for honest reviews. I show that workers punish firms that leave negative reviews by refusing to work for them again. Together, this body of evidence suggests that reputation systems in online labor markets provide significant information to both workers and firms and help reduce problems of asymmetric information.JEL Codes: D82 D83 J31 J49 * I thank Peter Kuhn for invaluable help with this project. I also thank Doug Steigerwald, Dick Startz, and seminar participants at the ZEW Workshop on Atypical Employment, the Trans-Pacific Labor Seminar, the WEAI Annual Conference, Berkeley, and UCSB for helpful comments and suggestions. I gratefully acknowledge financial support from the NET Institute (www.NETinst.org).The past 15 years have seen the rise of the "gig economy," with "tens of millions of Americans involved in some form of freelancing, contracting, temping or outsourcing" (Scheiber, 2015). Katz and Krueger (2016) find "that the percentage of workers [in the U.S.] engaged in alternative work arrangements-defined as temporary help agency workers, on-call workers, contract workers, and independent contractors or freelancersrose from 10.1 percent in February 2005 to 15.8 percent in late 2015". This increase in freelancers and independent contractors has been facilitated by the development of technological platforms that can bring together individuals (e.g. eBay, Uber, and Airbnb) and allows for more flexible working arrangements for both workers and firms.These platforms allow people on opposite sides of the world to transact and have the potential to greatly increase welfare by reducing transaction costs. One of the main requirements for the success of the new platforms is reputation:"successful online marketplaces have scaled because they have created well-designed reputation systems that allow users to identify trusted community members to interact with" (Stewart, 2014). Without a reputation system in place, it would be absurd to send money to someone on eBay or get in the car of an Uber driver. This p...