At a CEPR lunchtime meeting on Thursday 27 June, Heather Joshi presented results of her recent research on the effects of divorce on pension rights in the UK. Heather Joshi is a Senior Research Fellow at the Centre for Population Studies at the London School of Hygiene and Tropical Medicine and a Research Fellow in the Human Resources programme of the Centre for Economic Policy Research. Her remarks were based on her CEPR Discussion Paper No. 550, `The Pension Consequences of Divorce', written jointly with Hugh Davies. Financial support for the research presented in this paper and for the meeting was provided by the Joseph Rowntree Foundation. The views expressed by Ms Joshi were her own, however, and not those of the Rowntree Foundation nor of CEPR, which takes no institutional policy positions.
Joshi noted that divorce will play an increasing role in determining income distribution for both demographic and legal reasons. Pension rights are usually a major asset of a marriage (together with a house); but under current legislation they can neither be cashed in nor assigned to anyone who is not a scheme member. Therefore they cannot be divided in a divorce settlement, although in practice they are sometimes taken into account in cases where there are sufficient other assets to offset them. Few of the growing numbers of divorced women expected to reach pensionable age over the next decades will have access to pensions linked to their ex-husbands' earnings; but divorcees account for a higher proportion of the female than the male population, and proportion of women over 60 who are divorced is expected to increase from 3% in 1985 to 13% in 2025.
Although most wives now work, they typically earn significantly less than their husbands, so the pension rights earned by divorced women seldom match the rights to share in their husbands' pensions and associated widow's entitlements they would have enjoyed had they remained married. Joshi reported simulations of lifetime earnings, derived from econometric analyses of the UK government's 1980 Women and Employment Survey, which suggest that the earnings to pensionable age of a mother of two who remains married may be little more than a quarter of those of her similarly qualified husband. This fraction varies with the number of children, the woman's earning power and the matching of the partners' qualifications; roughly half the gap is due to motherhood, which typically involves a break in employment followed by lower hours and pay. The earnings of childless wives and unmarried women also fall short of those of similarly qualified men. Unless they are caught in the `poverty trap' of the benefit system, divorced women tend to earn more than married women, but less than men.
Eventually in the next century the Basic State Pensions earned by wives in their own right are likely to approximate their husbands'. Provisions already exist which enable divorced women to substitute their ex-husbands' National Insurance contributions for their own to make up their Basic State Pension. The earnings gap between spouses is reflected more strongly in earnings-linked pensions of various types; and the contrast between different pension types is just as striking as that between men's and women's pensions.
Joshi reported that a pension under SERPS the UK government's State Earnings-Related Pension Scheme is worth much less than a good Final Salary pension generated by the same earnings with one employer. A `Money Purchase' pension with maximum contributions (whose value will depend on how successfully the contributions have been invested) lies between the two. Reaping the advantages of either private scheme requires greater contributions than are paid into SERPS. Home Responsibilities Protection, which is paid under SERPS for time spent caring for children regardless of whether the parent has contracted out of the scheme when in full-time work, does little to bridge this gap. Nor does the inclusion of part-time employment in private pension schemes make much difference to women's pensions, since such earnings are low and partly covered by SERPS. Couples who are both well-qualified and in final salary occupational schemes enjoy the best pension prospects; `career housewives' and divorced women of low earning power who have been out of employment a lot fare badly; and pension prospects are worse for divorced mothers with large families or longer marriages, or those who have been kept out of employment by the lack of child-care facilities or the structure of the benefit system.
Dividing pensions at divorce would require new provisions in both pension and tax law. Joshi considered a scheme to split the rights accumulated during the marriage only, by paying a transfer to the partner with the smaller entitlement usually the wife from the other's scheme when the beneficiary reaches pensionable age. This does not equalize the ex-partners' final pensions, but it narrows the gap. For a typical case with two middle-income earners in SERPS, a transfer of £4 per week from an ex-husband's scheme will only raise an ex-wife's earnings-linked pension to £18 below the £25 required to keep her out of means-tested assistance if she has no other income. If both partners are in Final Salary schemes, in contrast, a transfer of £15 will increase her pension to £83 per week. Since most such transfers are small, the gains to divorced women will not always guarantee their freedom from hardship, and the `losses' faced by husbands and their subsequent wives are limited. Joshi noted from her simulations that the pension rights earned by second wives on the basis of normal working patterns could partly accommodate the splitting of their husbands' pension rights and their associated widow entitlements with first wives.
Joshi also noted that the alternative proposal for compensation in divorce based on modifying partners' independent pension rights to take account of their pooling of resources to rear children would be virtually impossible to organize in practice. Joshi reported simulations indicating that the existing Home Responsibilities Protection can only moderate and not eliminate such losses. Much of the loss to the wives is incurred after the divorce, when they are bringing up the children alone. Such a loss will usually be smaller than the lifetime pension gap between partners; and only by coincidence will the difference between pensions earned during marriage equal the pension losses arising from responsibility for children. Effective pension coverage is best afforded to such mothers by pensions that are not linked to earnings (such as a better Basic State Pension).