Duke 2008 | ASHE | Equity and Efficiency in Health and Healthcare

In this article, we focus, within a European framework, on the important role of health in the decision of males and females from 50 to 64 about whether or not they will take part to the labour market. We use a latent variable model (Bound (1991) and Campolieti (2002)) to estimate the effect of disability status on the labour force participation of older men in Europe; with data from a new European database SHARE. Labour force participation depends on a group of socio-demographic variables and unobservable latent disability. In a preliminary step, we estimate an equation of participation by directly introducing the self-reported disability, but the ?true' disability status is unobserved. Thus in a second step following Bound methodology, we use estimations of self-reported disability. In order to take into account the self-reported biases, we instrument theses measures with information on health conditions (health indicators relating to the diseases, deficiencies, Body Mass Index,...). . We use a specific method to take into account the multinational character of our database. With the estimation of labour force participation equation with direct introduction of health, we find: (1) Reporting directly 'restricted in one's activity' reduces the probability of labour force participation from between 16 to 11 points for females and between 24 and 18 for men. (2) The health status variables, all appear to have a negative effect, significant or almost, on labour force participation. (3) The diseases with the greatest effect are: Stroke (- 20 points), hip fracture (- 20 points), parkinson (- 16 points), etc.. for men; and Stroke (- 15 points), hip fracture (- 15 points), parkinson (- 10 points), etc.. for females. (4) The BMI marginal effect is weak for both gender. We also find the classical effects of socio-demographic variables on employment participation: (1)The level of education has a positive effect on employment participation, while age has the opposite effect. (2) Moreover, to live with a partner who works has a positive impact on the LFP (reference is to live alone); there is an opposite impact for people who lives with a partner unemployed. (3) The household size has a negative impact on the LFP for females and positive for men.For the estimation of disability equation, we show that: (1) Most diseases have a positive effect on the declaration of disability (except for cholesterol for men and cholesterol, blood pressure, stomach, parkinson, cataracts for women). (2) BMI has an effect on self-reported activity limitation. (3) Family structure and the employment status of partner have no effect. (4) The level of education has a negative effect on disability status. (5) The Euro-d score has a positive effect. And (6) the grip result has a negative effect on disability status.The main result of estimation of labour force participation equation in two stages are: (1) The estimated coefficients of the proxy values are significantly negative. The value associated with this proxy is higher in absolute value than the value estimated in the estimations of labour force participation equation with direct introduction of health status and disability. This suggests that using the health self-reported measure leads to a downward bias in the impact of disability status on labour force participation.