Women's Earnings
Who Cares?

How much greater would a typical British mother's cash earnings be over her lifetime if she remained childless? How much does her responsibility for childrearing reduce her time in paid work and the rate at which such time is paid? This paper provides a detailed description of a method which can be used to estimate the cash opportunity cost of having children in contemporary Britain.

Such an estimate is relevant to a number of issues: women's hidden contribution to the reproduction of society, their relatively weak economic position, their pension rights, and equitable financial arrangements on divorce. This paper, however, focuses on the economic analysis of fertility behaviour.

Earnings forgone by the mother are not the only cost of childbearing: there are also direct costs and the cost of time diverted from unpaid activities by both parents. However, the income which mothers sacrifice may well be a key element in any economic decision about whether and when to have children. Little childcare time is diverted from the labour market by men or purchased from it to substitute for mothers' time.
The simulations of forgone earnings make use of data from a 1980 survey of a cross-section of British women, the Women and Employment Survey (WES). In this paper the WES data are used to estimate the relationship between hourly pay and previous work experience and current hours of work. The data are first used to generate profiles of labour force participation and hours worked for representative women with different numbers of children. The data are also used to estimate a relationship in which hourly rates of pay depend upon previous work experience and current hours of work. Income for women with different numbers of children can then be estimated by combining the hourly wage rate equation with the participation and hours worked profiles.

The analysis of participation patterns indicates that the presence of dependent children causes interruptions in labour force membership and reduces hours worked in subsequent employment. Analysis of the hourly pay of women employed in 1980 suggests that the rate at which mothers are paid is reduced by interruptions in past work experience and by current employment on only a part-time basis.

Earnings forgone as a result of motherhood are then calculated as the difference between the income of someone whose employment history is unaffected by childrearing (but who is otherwise a representative individual) and the simulated earnings of a typical mother. The difficulties of estimating an average rather than a representative figure are discussed.
The technique used in this paper indicates that a typical British mother, in bringing up two children, forgoes earnings of £122,000 over her lifetime. This reflects an eight-year absence from employment after earning £6,000 per annum at age 24, followed by twelve more years spent in part-time employment than are spent by her child-free counterpart, mostly while her children are still at school. The income forgone can be attributed, in roughly equal proportions, to three factors: staying out of employment for eight years, lower hours of work thereafter, and lower hourly rates of pay.

These estimates of forgone earnings appear to be roughly double the direct cost of having children. They do not increase proportionately with the number of children in a family, but are sensitive to the time intervals between births. The opportunity costs of having the first child are around three times the direct expenditure on the child. The opportunity costs of subsequent births are approximately equal to the direct costs, and they increase with the interval between births. This suggests that the mother's labour market prospects should have a larger influence on decisions about starting a family than on decisions to have more children.

Some of these findings are in marked contrast to those reported by Calhoun and Espenshade for US mothers. The WES data indicate that British women are less likely than their US counterparts to return to paid work before their child reaches school age, but more likely to participate when they have no children. The present simulations therefore attribute much more lost employment to motherhood than does the US study. Calhoun and Espenshade calculate that opportunity costs are insensitive to birth spacing and roughly proportional to family size. They estimate these costs as $25,000 (in 1981 prices) per child for white families, which is less than one third the US estimate of direct expenditure on children. Both studies, however, find that the opportunity costs are insensitive, at least if undiscounted for their pattern over time, to the timing of the first birth.

The contrasts between the two sets of estimates may partly reflect differences in the use of paid childcare in the two countries, which give rise to the different patterns in employment histories. The different methods used in the two studies also exaggerate the differences between Britain and the United States. For example, Calhoun and Espenshade use a multi- state labour force life-table. Such a method involves assumptions which are not consistent with British life history data: UK women typically make fewer changes of employment status than such a model implies.

Our estimates can form the basis for considering the costs and benefits of alternative forms of childcare by quantifying the long-term benefits of measures which make it easier for mothers of young children to maintain continuity in their paid careers.