LAUSR

BACKGROUND Job loss is a well-established social determinant of health. Recent research has taken an 'institutional turn', asking whether unemployment support could buffer the health consequences of job loss. Here, we exploit a quasi-natural experiment based on the Fornero reforms in Italy, which increased wage replacement rates from 60% to 75% on 1 January 2013. METHODS We employed difference-in-difference models using longitudinal data covering 202 incidents of job loss from the EU-Survey on Income and Living Conditions to quantify the impact of job loss on changes in self-reported health prior to and after the Fornero reforms (2011-14). RESULTS Job loss pre-Fornero was associated with health declines -0.342 [95% confidence interval (CI): -0.588 to -0.096] but did not significantly influence health post-Fornero 0.031 (95% CI: -0.101 to 0.164). The difference-in-difference estimate was 0.373 (95% CI: 0.107-0.639), or a -0.51 standard deviation in self-reported health, consistent with the buffering hypothesis. To put the magnitude of this estimate in perspective, the incidence of a chronic illness, such as diabetes, results in a similar magnitude decline in self-reported health. CONCLUSIONS Our analysis contributes to a growing body of evidence that the impact of job loss on health depends critically on the strength of social protection systems and, in some cases, could be eliminated completely.