Unemployment and Health: A Meta-Analysis

Abstract

This paper is a meta-analysis on the relationship between unemployment and health. Our meta-dataset is made up of 327 study results coming from 65 articles published in peer-reviewed journals between 1990 and 2021. We find that publication bias is important, but only for those study results obtained through difference-in-differences or instrumental variables estimators. The average effect of unemployment on health is negative, but small in terms of partial correlation coefficient. We investigate if findings are heterogeneous among several research dimensions. We find that unemployment is mostly effective on the psychological domains of health and that short- and long-term unemployment spells equally affect health. Dealing with endogeneity issues is important and, when this is done, the unemployment effects on health are closer to be nil.

Keywords: unemployment, health, meta-analysis, meta-regression, publication bias

JEL Classification: C52, I10, I12, J64