Job Search Effort during Unemployment: Insights from New Data and Theory

• Applicant: Dr. Joerg Heining
• Subject Area: Statistics and Econometrics ; Economic Policy, Applied Economics
• Term: from 2017 to 2019
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 329151064

This project aims to contribute to the intersection between Behavioral and Labor Economics by advancing the literature on job search. The project investigates the mechanisms that determine the disincentive effect of Unemployment Insurance (UI) benefits and in particular the high transition rate to employment at the end of UI-receipt (the so called spike at the exhaustion point).First, we collect detailed information on the efforts invested in job search of recipients of UI benefits which is a prerequisite for analyzing the mechanisms of the disincentive effect. We develop a new and highly innovative survey of unemployed workers collected via text messages (SMS) that will elicit longitudinal search effort information from job seekers. For this purpose, we have access to the universe of current UI recipients stored at the Institute for Employment Research (IAB), where we basically can observe and target individuals precisely based on their current eligibility duration and other characteristics in real time. For the survey, we will select individuals that have quasi-random differences in the duration of UI eligibility, using sharp eligibility increases at several experience thresholds. The collected data will be merged with administrative data of the IAB, allowing for a rich set of additional variables.Second, we contrast the predictions of different job search models using the collected survey data. This allows us to compare changes in job search within individuals. In particular, we test the predictions of a job search model with reference-dependent workers and contrast it with a standard job search model. Reference dependent unemployed are particularly sensitive in their job search to recent cuts in benefits. They increase their search effort when approaching the exhaustion point but get used to the lower level after UI-exhaustion and reduce their search effort again. The standard search model in contrast, predicts an increase up to the exhaustion point, but stays constant, afterwards. Empirically distinguishing the two is important since they have different policy implications.Third, by using the combined job search-admin data, we aim to advance the understanding of how job search assistance measures such as vacancy referrals by the case worker affect job search behavior. In addition, we investigate whether the effect of job search assistance on the search behavior varies between individuals close to, or far from the exhaustion point. Understanding the interplay of job search assistance and job search, can help to improve the targeting and therefore the efficiency of such assistance in practice and help to shed light on the question why some measures are effective and others not.

• DFG Programme: Research Grants
• International Connection: USA
• Cooperation Partners: Professor Stefano DellaVigna, Ph.D.; Professor Johannes Schmieder, Ph.D.