The aim of the present work is to study the impact of unemployment on coronary heart diseases (CHD) in Italy on a sample of male manual workers in the private sector.
We investigate the association between CHD and different unemployment experiences (ever unemployed; short, mid and long cumulative unemployment), exploiting a large Italian administrative database on careers and health. We designed a 12-years pre-treatment period, on which to balance individuals’ characteristics; a 7-years treatment period to measure the unemployment occurrence; a 5-years follow up to observe CHD. The workers characteristics and the probability of receiving the treatment are balanced by means of Propensity Score Matching. Standard diagnostics on the balancing assumption are discussed and satisfied, while the robustness to violations of the unconfoundedness assumption is evaluated by a simulation-based sensitivity analysis.
We find a significant increase of CHD probability for workers who experience more than three years of unemployment (RR=1.91, p<0.1), and for those who exit unemployment starting a self-employment activity (RR=1.70, p<0.1). Using different selections of the study population, a clear pattern emerges: the healthier and more labour market attached are workers during pre-treatment, the greater is the negative impact of long term unemployment on health (RR=2.79, p<0.01).
The very large representative sample (n=69,937) and the deep longitudinal dimension of the data (1985-2008) allowed us to minimize the risks of health selection and unemployment misclassification. Moreover, the adopted definition of unemployment corrected some undercoverage and misclassification issues that affect studies based on a purely administrative definition and that treat unemployment as a unique career event disregarding the duration of the experience.