The effect of retirement on social relationships: new evidence from SHARE

We analyze the causal effect of retirement on the size, composition and intensity of social relationships using data from the Survey of Health, Ageing and Retirement in Europe for 11 European countries. Our empirical strategy exploits the different retirement eligibility ages as instruments for the endogenous individuals’ retirement decisions and controls for time invariant individual characteristics. We show that retirement changes the composition of the individual’s social network, increasing the share of family members, and decreasing the share of colleagues and friends, while there is no effect on the network’s absolute size. Changes in the social network’s composition are associated with a higher overall satisfaction and more intense relationships. We argue that retirement induces a substitution between weak (friends or colleagues) and strong ties (family), along with an increase in the intensity of the surviving ties. Interestingly this substitution has a gender dimension: females mainly reduce the share of friends while males that of colleagues.