Working Papers

Pension Design for Couples and Singles Job Market Paper

· with Annasofie M. Olesen

Abstract

Couples can share household expenses, which reduces their cost of living. Many social insurance programmes therefore pay lower rates per person to couples compared to singles. How does this change relationship choices? This paper exploits administrative data and institutional features of the Danish pension system to show that the separation risk increases by 18% in response to financial incentives in a state pension of this kind. This implies that the overall share of singles in the population increases by 2.3% for a 50,000 DKK (about 7,400 USD) difference in annual income by relationship status. We develop a limited-commitment life-cycle model with endogenous divorce and retirement to evaluate the fiscal and welfare consequences of this policy. We show that the mechanical cost of paying higher rates to singles makes up 54.4% of the total fiscal cost, with the remainder arising due to the behavioural responses in divorce and retirement timing. But while the policy redistributes from couples to singles, it also provides insurance against the risks of divorce and widowhood. For this reason, 40% of those starting as married prefer the status quo with the lower rate for couples to a revenue-neutral equalisation of pensions.

Divorce rate at state pension age (SPA) and early retirement age (VERP)
Figure 1. The divorce rate spikes at the state pension age (SPA), but not at the popular early retirement age (VERP).
Retirement after Divorce: A Structural Model of Pension Sharing and Insurance

· with Annasofie M. Olesen

Abstract

Pension-sharing rules at divorce vary widely across countries, but little is known about how alternative rules affect behaviour and welfare. These policies may provide valuable insurance to the pension-poorer spouse, particularly when divorce occurs late in life. We study their effects in Denmark, where pension wealth is not divided between divorcing spouses. Empirically, we show that divorce is associated with delayed retirement, particularly among women. We then estimate a dynamic structural model of household bargaining, divorce, and retirement to study how households insure themselves against late-life divorce and evaluate two alternative pension-sharing policies. Cash compensation for pension wealth differentials mainly smooths consumption around divorce and has little effect on retirement. In-kind pension sharing, by contrast, leads divorced women to retire earlier and reduces their reliance on public pensions. Ex ante, married households value the insurance provided by pension sharing, but the form of compensation matters: Only in-kind compensation generates net positive willingness to pay. Both policies generate small positive fiscal externalities.

Labour supply decisions of somewhat caring individuals

Abstract

How do couples make labour supply decisions when their degree of cooperation varies? This paper applies a model of two-agent decision-making that spans the full range from non-cooperation to perfect cooperation to household labour supply. The degree of cooperation is equivalent to how much partners care about each other. In- creasing caring improves individual and household utility, except when caring is very asymmetric. A rise in the lower earner’s wage decreases the higher earner’s utility by shifting the within-household distribution of consumption. Productivity in housework, by contrast, confers no power because its benefits are non-excludable. All else equal, the partner with the lower productivity in housework is better off. Labour supply elas- ticities rise with intra-household wage inequality, and more so the higher the degree of cooperation. Caring therefore matters: assuming perfect or no cooperation misses these patterns and misstates the efficiency gains from taxing labour income.