Sickness and injury leave in France: moral hazard or strain?

From 1997 to 2001, the total payment to compensate for sickness and injury leaves increased dramatically in France. Since this change coincided with a decrease in unemployment rate,three hypothesizes should be proposed as possible explanations consistently with the literature: moral hazard (workers fear less to loose their job, therefore use sickness leave more confidently); strain (workers work longer hours or under more stringent rules); labor-force composition effect (less healthy individuals are incorporated into the labor force). We investigate the first two strands of explanation using a household survey (ESPS) enriched with claims data from compulsory health insurance funds on sickness leaves (EPAS). We model separately number of leaves per individual (cumulative logit) and duration of leaves (random-effect model). According to our findings, in France, the individual propensity to take sickness leave is mainly influenced by strain in the workplace and by a labor-force composition effect. Conditional duration of spells is not well explained at the individual level: the only significant factor is usual weekly work duration. Influence of moral hazard is not clearly ascertained: it has few impact on occurrences of leave and no impact on duration.