to allow the construction of a system of equations to estimate the relationship between chronic diseases and labour supply.
We focus on the participation decision rather than hours worked although we recognise that ill health associated with chronic disease is also likely to affect the number of hours worked. Labour force participation L is modelled as:
L=L(H, X, )
Where H is the health status of the individual, X are exogenous observable household
characteristics that affect productivity, and labour outcomes.
are unobserved household characteristics that affect
Ill health will reduce the probability of employment for several reasons. As summarized in (Disney, Emmerson et al. 2006), first poorer health may raise the current disutility of work. Second poorer health reduces productivity, and through demand, the return from work and consequently wages. Third poor health may entitle the individual to non wage income such as disability benefits. Last poorer health may also lower life expectancy raising the present value of current wealth and induce earlier retirement. In this kind of structural model much attention has been paid in the literature to the issue of measuring the true underlying health of individuals. A common measure of health status is overall self reported health. However while at any point in time there may be a correlation between self reported health status and labour outcomes this may well be due to bias in measurement as individuals who report no employment misrepresent their health status. The most obvious solution to this would be to use a more objective measure of health. The problem however is to find adequate proxies for the health conditions that affect activity. If we restrict the issue to one disease such as diabetes we assume that other illnesses either do not affect the activity decision or are uncorrelated with diabetes. Moreover a single equation model of labour force participation and chronic diseases assumes that the diseases are exogenous determinants of labour outcomes. There are a number of reasons to believe that this might not be the case. For example there may be a number of unobserved common risk factors across diseases and the decision to supply labour for example associated with peer groups or family background. In this paper we construct a structural model of a number of chronic diseases and labour supply that reduces the potential for this source of endogeneity bias as well as that associated with potential simultaneous across diseases. The latter might arise for example in the case of diabetes and heart disease each of which might simultaneously influence labour force participation but where diabetes might independently be a risk factor for cardiovascular disease. An important potential advantage of this approach is that it allows us to estimate the indirect effect of a range of potential social and behavioural influences on employment that operate through chronic illness without making the assumption that chronic diseases are independent except for observed common risk factors and that each disease is an exogenous determinant of labour supply behaviour. For example we can examine the indirect effect of obesity and smoking behaviour on employment through the effect on the risk of each chronic disease.
Chronic disease and labour force participation in Australia: an endogenous multivariate probit analysis of clinical prevalence data