Collaborative Research: Reference-Dependent Job Search

University Of California-Berkeley, Berkeley CA

Investigators

Abstract

With sustained long-term unemployment as one of the key challenges in the US labor market, it is critical to consider alternative approaches to understand unemployment. The project proposes that behavioral economics can contribute to our understanding of unemployment, and in particular, focuses on reference dependence in job search. Reference-dependence suggests avenues for a potential redesign of unemployment insurance systems, including the switch to multiple-step unemployment systems. The PIs examine evidence in this regard from a natural experiment on an existing unemployment system, and from an upcoming survey of unemployed workers. Based on this analysis, they will be able to elaborate policy recommendations on possible reforms of the Unemployment Insurance system. The intellectual merits of the project are threefold. First, it incorporates one of the best-established behavioral phenomena, reference dependence, in a job search model and shows that reference dependence leads to a potentially novel understanding of job search patterns. Reference-dependent unemployed workers are particularly sensitive in their job search to recent cuts in benefits. Hence, they display a unique response to the path of unemployment benefits, and in particular to switches from a 1-step system to a 2- or 3-step UI system. Second, the project tests the predictions of the reference-dependent job search model using a reform in Hungary which changed the delivery of unemployment insurance benefits from a single-step system to a two-step system, holding approximately constant the overall generosity. This reform is an almost ideal testing ground for the model and provides a setting to derive predictions about the welfare implications of altering the path of UI benefits. Third, the project develops a new survey of unemployed workers that will elicit longitudinal search effort information from job seekers. This panel data set will allow for testing a unique prediction of the reference-dependent model: that search intensity increases leading up to benefit expiration, and then decreases again. The project builds on and expands new insights from the field of Behavioral Economics, which emphasizes the importance of reference points (such as past benefit levels or the income of a comparison group) to advance the understanding of the job search behavior of individuals who are receiving UI benefits. The researchers intend this project to be a step in building sustained interest in Behavioral Labor Economics as an active research area, both in their own future research and in attracting interest by other researchers and policy-makers.