Even in countries with high average job security, workers with low tenure typically enjoy very limited job protection. This study analyzes the impact of such a feature on job duration. It uses a 1999 British reform that increased job security for workers with 1–2 years of tenure. The firing hazard for these workers decreased by 26% relative to the hazard for workers with 2–4 years of tenure. The firing hazard for workers with 0–1 year of tenure also decreased by 19%, which is consistent with better recruitment practices and hence improved match quality.

On the leading job board CareerBuilder.com, high-wage job postings unexpectedly attract fewer applicants, and this is the case even within a detailed occupation. Viewed through the lens of our directed search model, this negative relationship is indicative of substantial applicant heterogeneity within an occupation. Empirically, we find that job title heterogeneity is key: within a job title, jobs with 10% higher wages do attract 7.7% more applicants. Furthermore, our findings are consistent with a higher return to worker quality for hires in “manager” and “senior” job titles. Overall, our findings demonstrate the power of words in the matching process.

How does unemployment insurance (UI) affect unemployed workers’ search behavior? Search models predict that until benefit exhaustion, UI depresses job search effort and increases reservation wages. Over an unemployment spell, search effort should increase up to benefit exhaustion and stay high thereafter. Meanwhile, reservation wages should decrease up to benefit exhaustion and stay low thereafter. To test these predictions, we link administrative registers to data on job search behavior from a major online job search platform in France. We follow over 400,000 workers, as long as they remain unemployed. We analyze the changes in search behavior around benefits exhaustion and take two steps to isolate the individual response to unemployment benefits. First, our longitudinal data allows us to correct for changes in sample composition over the spell. Second, we exploit data on workers eligible for 12–24 months of UI as well as workers ineligible for UI, to control for behavior changes over the unemployment spell that are independent of UI. Our results confirm the predictions of search models. We find that search effort (the number of job applications) increases by at least 50% during the year preceding benefits exhaustion and remains high thereafter. The target monthly wage decreases by at least 2.4% during the year preceding benefits exhaustion and remains low thereafter. In addition, we provide evidence for duration dependence: workers decrease the wage they target by 1.5% over each year of unemployment, irrespective of their UI status.

This paper shows that more highly concentrated labour markets experience more positive employment effects of the minimum wage. In the most concentrated labour markets, employment rises following a minimum wage increase. The paper establishes its main findings by studying the effects of local minimum wage increases on a key low-wage retail sector, and using data on labour market concentration that covers the entirety of the U.S. with fine spatial variation at the occupation level. The results carry over to the fast-food sector and the entire low-wage labour market and are robust to using proxies of labour market concentration available for a broader range of industries, such as the number of establishments and population density. A model of oligopsonistic competition can explain these effects: there is more room to increase wages in high-concentration areas where wages tend to be further below marginal productivity. These findings provide evidence supporting monopsonistic wage setting as an explanation for the near-zero minimum wage employment effect documented in prior work.