Atypical employment and health : a meta-analysis

In this meta-analysis we provide new quantitative evidence on the relationship between the characteristics of working contracts and worker's health. We examine 52 studies covering 26 countries in the time period 1984 - 2010 with a combined sample size of 192. We apply a random effects model using odds ratios and their 95\% confidence intervals as measures for the effect size. We distinguish between six types of employment contracts with decreasing security levels (fixed-term, temporary, casual, on-call, daily, no formal contract) and classify the health outcomes into five subgroups (sickness absence, occupational injuries, health-related behavior, mental health and physical health). Furthermore, we control for selected dimensions of the socioeconomic environment of the studies, e.g. the unemployment rate and GDP growth rate. Summary findings show a higher risk of occupational injuries for atypical employees compared to the reference group. Atypical employment increases complaints about mental and physical health and has a negative impact on health-related behavior. Sickness absence works in the opposite direction and permanent employees are more likely to be absent from work. The heterogeneity of the effect sizes between different contracts of atypical employment is low. Effect sizes are country specific and depend on the health outcome indicators. The macroeconomic surrounding - unemployment rate and GDP growth rate - don't cause variation in study results. The 'healthy worker effect' may lead to an overestimation of the impact of workers' atypical employment contract on the health status. More research work which explicitly focuses on the problems of endogeneity, reverse causality and the selection bias is necessary. Furthermore, additional control groups and the employment biography of workers have to be taken into account.