In this essay we focus on loss of working capacity due to old-age, with special focus on gender. We analyze different ways to organize public pension systems and give the pros and cons of different features with respect to gender. The main objective of a pension system is to relieve poverty in old-age, but also to offer an insurance against an (unexpected) long life. Will a certain design favour women more than another design and in what respect? The first part of the paper analyses pension system from a theoretical point of view and examines to what extent the pension schemes in China, France, Ghana, Jordan, Mexico, Poland and Sweden have features that are advantageous to women. In the second part we focus on empirical evidence mainly from the Swedish pension reform.

We identify the following features as being important to prevent poverty among women and to secure income replacements. A pension system should:

• be public and mandatory in order to make unisex life tables possible;
• provide economic incentives for women to work in the formal sector;
• not punish women by giving them a lower rate of return on lifetime contributions than men;
• not punish child rearing;
• provide a minimum guarantee;
• index pensions;
• provide joint-and-survivor annuities, which provide continued benefit payments to a surviving spouse;
• allow accrued pension rights to be divided in the event of a divorce.

These features are important because women's labour market behaviour differ from men. In particular women work more part-time, have more interrupted careers and are lower paid. Furthermore women have longer life expectancy and are more likely to become widows. It is important to provide adequate pensions and income replacement but a pension system should not compensate for gender differences in the labour market, as that would reinforce traditional gender roles and preserve discrimination in the labour market.