Pré-publication, Document de travail: This paper examines the causal impact of later retirement on doctor visits among the French elderly. This question is of interest since spillover effects may arise if later retirement increases healthcare expenditure. I exploit the 1993 French pension reform in a two-stage least square to deal with the endogeneity of retirement. This reform leads to a progressive increase in claiming age, cohort by cohort from 1934 to 1943. I use a two-part model to disentangle between extensive and intensive margin. I use the administrative data HYGIE to observe both healthcare consumption between 2005 and 2015 and past careers. I find that an increase in retirement by four months decreases significantly the probability to have at least one doctor visit per year by 0.815 percentage point and decreases the number of doctor visits by 1.14% between ages 67 and 75. This effect is driven by the consumption of generalist doctor visits, and tends to be stronger for the first ages of consumption observed.