Economics serving society

Spousal Retirement: a Regression Discontinuity Study

Short link to this article:

Elena Stancanelli

JPEG - 59.1 kb

Ageing populations and growing budgetary pressures have pushed most OECD countries to introduce reforms aimed at extending working life, particularly by increasing the legal age of retirement. France was the pioneer in this measure. As in the majority of other countries, retirement policies are conceived for individuals, though 80 per cent of older people (aged more than 60 years) live with a partner and very likely plan their retirement together. In the United States as in France, three quarters of couples are dual-income couples, and most married women work full time. The employment rate of individuals aged over 60 years in France is, however, almost zero – one of the lowest in the OECD. It is therefore imperative, from a public policy point of view, to understand the retirement strategies of married workers.

In this article, Elena Stancanelli analyses a French government reform introduced in the summer of 1992, which - by increasing the required number of trimesters worked that are required before someone can receive the full pension - encouraged workers to put off their retirement date. She estimates the effect of this reform on couples’ decisions about their employment. She uses data from the INSEE employment enquiry, over several years, covering a sample of 50,000 seniors couples. The first conclusion concerns the “expected” effect of the reform: that it immediately reduced the probability of the two partners retiring at 60 (1) but also slightly increased the probability of being unemployed for the man and of declaring herself “at home” for the woman (2). The second conclusion is the revelation of a slight decrease in the probability that each partner would retire when his or her partner was affected by the 1992 reform. Unsurprisingly, the reform encouraged individuals to postpone their retirement slightly: leisure complementarity among couples (the desire to pass their leisure time together) is one of the principal reasons to choose to retire “together” according to a great many economic studies. Stancanelli then considers the legal retirement age: if the husband is more than 60 years old, his probability of retiring rises by around 30 percentage points and this has the effect of increasing the probability that the woman will also retire (by 2 to 5 percentage points), no matter what age the woman. However, these crossed effects of the retirement of a partner on the retirement of the other are, in the end, very low. On average, in France, husbands are two years older than their wives, and existing legal modalities “prevent” dual-income couples from retiring together; the husband often finds himself at home for many years while his wife continues to work enough in order to earn the full pension rate. Thus, this article brings to light the “negative” role that retirement age policies can play in impeding the possibility of seniors retiring together.
(1) Down around two points for men and four points for women.
(2) One and three percentage points respectively

Original title of the article : “Spousal Retirement: a Regression Discontinuity Study”
Published in : AGENTA Working Paper
Available at :