Recall and Unemployment

Résumé

Using data from the Survey of Income and Program Participation (SIPP), we document that recalls of former employees are very frequent and associated with dramatically different unemployment and post-unemployment outcomes, relative to those of separated workers who change employer. Specifically, over 25% of all workers who are separated from their jobs (including those who immediately leave the labor force), 40% of all separated workers who remain initially unemployed, and roughly half of separated workers who remain continuously unemployed until they regain employment within two years (completed spells), go back to work for their previous employer after the jobless spell. These shares vastly exceed that of separated workers who enter unemployment on a temporary layoff, despite the fact that a significant fraction of them end up either taking another job or leaving the labor force. The reason, and one of our main findings, is that of all the workers who are permanently separated and start looking for another job, who are the bulk of the unemployed, close to 20% eventually return to their last employer. Recalled workers had twice the tenure with their previous employers before being separated, spend just over half the time unemployed, experience a more favorable real wage change (if permanently separated) and switch occupation much less often after the jobless spell. Finally, of all workers who lose jobs but remain unemployed, only those who are eventually recalled show negative unemployment duration dependence; those who change employer leave unemployment at slower but roughly constant hazard over their spell, which is also more procyclical than the hazard rate of recall. To make sense of this evidence, we introduce a recall option and aggregate productivity shocks in the standard search-and-matching model of the labor market, `a la Mortensen and Pissarides (1994). While new matches require costly search and are mediated by a matching function, recalls are free and triggered both by aggregate shocks and by job-specific shocks that continue after separation. The recall option is lost when the unemployed worker accepts a new job.