| Summary: | <p> The first chapter of this dissertation studies a labor market where workers search for both more productive and more secure employment. In this environment, an unemployment spell begets future unemployment spells and the hazard rate into unemployment declines with tenure. In a laissez-faire economy, workers overvalue job security relative to productivity and unemployment benefits can increase welfare. I estimate the framework on German Social Security data and use it to study quantitatively the consequences of job loss. The model explains the large and highly persistent response in wages and employment known as the “unemployment scar”. The key driver of the long term losses is the original loss of job security and its interaction with the evolution of human capital.</p><p> The second part of this dissertation (joint with Laura Pilossoph) studies the Consequences of statistical discrimination against the long-term unemployed. Recent evidence suggests that employers statistically discriminate against the long-term unemployed (Kroft, Lange, and Notowidigdo (2013)). To map this evidence into its effect on the job-finding rate and long-term unemployment, we propose a search model of the labor market where dynamic selection on unobservables endogenously generates such discrimination. We show quantitatively and analytically that when discrimination is rationalized as statistical, it has only small effects on the job finding rate and contributes little to long-term unemployment. Firms discriminate to avoid costly interviews with workers they would not hire ex-post. But since they would not hire them ex-post, discrimination at the interview stage has no real effect on their counterfactual job-finding rates. This provides an important qualification to recent evidence on employer discrimination from field experiments.</p>
|
|---|
