7. april 2021

Unemployment Insurance and Job Search Behavior

How does unemployment insurance (UI) affect unemployed workers’ search behavior? Daphné Skandalis set out to answer that question and her work has just been published in Quarterly Journal of Economics.

Daphné Skandalis (and her co-author Ioana Marinescu ) analyze the search behavior of unemployed workers around benefits exhaustion and find that on the one hand, longer unemployment insurance delays the increase in job applications at exhaustion; on the other hand, it allows job seekers to target better quality jobs. They also show that time spent unemployed has a negative effect on target wages, independent of UI benefits duration: during each year spent unemployed, workers decrease their target wage by 1.5%.

With spell and time of spell fixed effects, Daphné and Ioana estimate that unemployed workers increase their search effort by at least 51% in the year prior to benefits exhaustion and keep it high thereafter. Meanwhile, they decrease their monthly target wage by 2.4% to 3.5% in the year prior to benefits exhaustion and keep it low after.

The results are consistent with the predictions of search models. As predicted by the models, Daphné and Ioana find evidence that workers with more generous UI are more selective in their search. This is not necessarily reflected in their reemployment wages, however, as they also experience a decrease in their job opportunities over time due to duration dependence.

Furthermore, the results suggest that the increase in the reemployment rate before benefits exhaustion is caused by a change in individual search behavior, and that the decline afterward is due to a combination of individual behavior, dynamic selection, and duration dependence.

The results contradict the view that the spike in the reemployment rate at benefit exhaustion can be explained away by search-free mechanisms, such as storable job offers. Our findings highlight the importance of search frictions and thus provide a justification for the insurance role of unemployment benefits.

To test these predictions, Daphné and Ioana link administrative registers to data on job search behavior from a major online job search platform in France. They follow over 400,000 workers, as long as they remain unemployed. They analyze the changes in search behavior around benefits exhaustion and take two steps to isolate the individual response to unemployment benefits.

Intrigued and wanting to know more about the study? Then read the full article in Quarterly Journal of Economics.