Over the last two decades, most OECD countries have reformed their social security in order to make early departures from the labor market increasingly difficult. Despite the fiscal gains that are expected from these reforms, it is likely that these gains from longer careers will be partly offset by increasing expenses on other social security programs. This article sheds light on this issue by ex-ploring the consequences of postponing access to an old-age unemployment program from age 58 to 60. The program provides laid-off workers with a combination of unemployment benefits and a monthly supplement paid by the employer until the full retirement age. Exploiting a rich set of administrative data, we study the effect of this reform on workers' employment and various social security benefits (i.e. unemployment, disability, early retirement and compensated working time reductions), using a triple difference method as identification strategy. Our results show that, for men, the reform had a positive effect on employment, with a small positive effect on a program called Time-Credit, i.e., a social security program that facilitates working time reductions at the end of the career. For women, we find no significant effect on employment but instead a large spillover effect on unemployment. We find that gender differences in job characteristics can help to explain this difference, since women are more likely to work in part-time, low-wage and blue-collar occupations than men, and no significant employment effects are found for these groups of workers.