The effect of health on the labour force outcome among working age individuals in the UK

This thesis contains three empirical chapters on the effect of health on some key labour-force outcomes of working age individuals in the UK. In the first paper chapter two, the effect of health shocks on exit from employment among working age individuals using data from British Household Panel Survey was estimated. Factor analysis was used to model health as an unobservable concept with two correlated dimensions (mental and physical). Past mental and physical health status as well as mental and physical health deteriorations had significant effects on exit from employment. In the second paper, I addressed the effect of changes observed in different components of income after acute health shocks were experienced among working age individuals. Identification arrived from exploiting uncertainty in the timing of an acute health shock, defined by the incidence of cancer, stroke, or heart attack. Results after coarsened exact matching, showed that health shocks significantly reduced labour income and increased welfare income, with younger male workers experiencing the greatest reduction in their net income and no significant increase in welfare income. The impact of diabetes on exit from employment decisions of individuals in England was investigated in chapter four. Using data from English Longitudinal Survey of Aging, I utilised a recursive bivariate probit approach to test for the potential endogeneity of diabetes in employment outcomes. Parental history of diabetes was used as genetic instrumental variables. Results did not suggest that diabetes is endogenous. Investigation was advanced by employing a discrete time hazard model on the sample of male and female individuals aged 50 years or older in the first wave of ELSA, who were also in paid work. Results illustrated that being diagnosed with diabetes is associated with an increased hazard of leaving employment in estimated sample. Adverse effects on employment probability are higher among insulin or oral medication users.