Countries differ widely in whether pension wealth is divided at divorce or remains separate property. This paper provides the first evidence on how such policies shape divorce behavior, labor supply, and late-life outcomes. Using rich administrative data, we document that divorce is linked to delayed retirement for women in systems where pensions are not shared. We then develop a dynamic structural model of household bargaining, divorce, and retirement to evaluate counterfactual pension-splitting rules. Relative to a no-compensation benchmark, introducing pension division at divorce substantially lowers retirement ages among divorced women, while effects for divorced men and married couples are small or negligible. For an equivalent transfer, labor supply responses are markedly stronger for women than for men. When public pension benefits are means-tested against other income, overall state pension expenditures rise slightly, as behavioral labor supply responses more than offset the mechanical reduction in pension wealth inequality generated by sharing.

This paper studies savings behavior and its determinants in old age. A large empirical literature documents low drawdown of wealth late in life. I confirm this result in the context of the Netherlands, where there are extensive public insurance schemes covering both health and care expenditures. To investigate the determinants of this observed behavior, I provide suggestive evidence that bequest motives cannot fully account for this empirical finding. The results have implications for the retirement consumption smoothing benefits of private saving in working life, and is important in considering how to insure retirees against risks such as long-term care late in life.

This paper studies to what extent marital and cohabitation decisions respond to financial incentives in a state pension. As is common in many benefit programs around the world, the Danish state pension pays more to individuals living alone to make up for the higher cost of living for singles. In response, at the age cutoff for eligibility, the likelihood of living alone jumps up by 0.2\% pt. due to both an increase in separation and divorce risk for existing couples and a fall in the likelihood of moving in with a new partner. By focusing on a subsample of individuals and couples already retired, it can be ruled out that this is due to retirement itself, which commonly takes place under an early retirement program which is not means-tested.

This paper provides a non-cooperative model of domestic violence that models the decision to enter a marriage and the decision to abuse one’s partner. The effect of policy interventions on aggregate abuse rates is decomposed into a marriage-rate effect and an abuse-rate effect, which are analysed separately. The model predicts that lower divorce costs increase marriage rates while decreasing abuse rates. Improvements in women’s outside options relative to the surplus from marriage can have ambiguous effects on abuse, with the potential to either reduce or exacerbate it. In contrast, policies that widen the utility gap between single and married men are predicted to lower abuse rates. By explicitly distinguishing between the marriage-rate and the abuse-rate effect, this framework highlights how policies aimed at reducing domestic violence may have adverse effects.