5

This study attempts to analyse the causal links between the social participation (social capital) and health of elderly people in Europe. It is based on the three rounds of the Survey of Health, Ageing and Retirement in Europe (SHARE) among individuals aged 50 and over in 11 countries. In each of the first two rounds (2004 and 2006), one dichotomous variable identifies participation in social activities (associations, clubs, political parties, etc.) and seven dichotomous variables provide information about the physical and mental health of the respondents. A recursive bivariate probit model is used to estimate the impact of social participation in 2004 on health in 2006, and, reciprocally, of health in 2004 on social participation in 2006. As well as the usual control variables, the retrospective data from the third round of the survey of life histories (SHARELIFE) inform about the initial conditions of the sample. The results suggest a reciprocal causal effect: social participation benefits health and vice versa. However, health seems to have a bigger impact on social participation than the reverse. Consequently, elderly individuals in good health are more likely to preserve their health through the beneficial effect of social capital, whereas those in worse health are less likely to participate in social activities and, without their beneficial effect, are more likely to experience a faster deterioration in their health. In conclusion, despite its beneficial individual effects, social participation is a potential vector for aggravating health inequality among elderly people.