Jointleisure before and after retirement :A double regression discontinuity approach

In the literature on partners' joint retirement decisions one of the explanations for joint retirement is externalities in leisure. Exploiting the law on early retirement age in France, we use a regression discontinuity approach to identify the causal effect of retirement on hours of leisure, separate and together, of individuals in a couple. We use a sample of couples drawn from a French Time Use Survey for the analysis. We find that the own retirement probability increases significantly at age 60 for both partners, which supports our identification strategy. However, the own probability to retire does not increase when the partner reaches age 60. These findings are corroborated using a larger sample of over 20 000 couples drawn from French Labor Force Surveys. We conclude that upon partners’ retirement leisure hours spent separately from the partner increase by at least as much as leisure hours together. The increase in hours of joint leisure is not larger than that in house work hours either. We find that hours of leisure together only increase significantly when the younger partner (presumably the last to retire) retires. When the older partner (presumably the first to retire) retires, hours of joint leisure do not increase. These conclusions are robust to using four alternative definitions of leisure together as well as to narrowing the size of the sample on the two sides of the age discontinuity. They are also well in line with recent literature on retirement strategies of individuals in a couple, casting doubts on whether partners always retire together.