WORKSHOP ON MALE-FEMALE DIFFERENCES IN THE BRITISH LABOUR MARKET
October 5 1987

Despite the well-known increase in the participation by women in the British labour market there are still large differences between men's and women's work. Not only do women, on average, work fewer hours for pay and earn less per hour, they are largely concentrated in occupations not followed by many men. The main aim of the workshop, held by CEPR on October 5, was to explore the scope and structure of these differences.

The workshop was organized by Heather Joshi (Centre for Poulation Studies, London School of Hygiene, and Co-Director of CEPR's programme in Human Resources Since 1900), with financial assistance from the Employment Market Research Unit of the Department of Employment. There were four sessions at which seven papers were presented. The sessions were: 'How different is the allocation of men's and women's time ?'; 'Explanations of the male-female pay differences in two national cohort studies'; 'How badly paid are part-time jobs ?'; and 'Further issues in the estimation of earnings functions.' There were twenty-five participants with a variety of social science backgrounds being represented.

In the first session, Jonathan Gershuny (University of Bath) talked about time budget data and presented a paper, written in conjunction with John Robinson (University of Maryland), entitled 'Historical Changes in the Household Division of Labour.' Using survey data from six nations (U.K., U.S.A., Canada, Netherlands, Denmark and Norway) he examined changes in the allocation of men's and women's time between the household and the labour market (or alternatively between unpaid and paid work). Reacting to earlier research based on the American experience showing no significant decreases in the amount of time housewives devote to domestic work, he finds a clear downward trend for women and an upward trend for men through the 1970's and into the 1980's. Though women still spend much more than time men on domestic tasks - by over three hours per day, on average in Britain in 1984 - it appears that gender differentials in domestic responsibilities have declined over the past two decades. Furthermore, this covergence cannot be explained simply by the substitution of paid work for unpaid work (i.e. by the rise in female labour force participation). Two partial explanations for this trend were given. The first is the diffusion of "labour saving" household technology which has increased the efficiency at which household tasks can be performed. The second is a change in societal norms and values which have led to a reorientation of the division of labour within the household. Gershuny concluded his presentation with a discussion of the relevance of information about about the demands on people's unpaid time to questions of labour market, transport, education and child-care policies.

Peter Dolton (University of Bristol) noted that in order to use time budget analysis to its maximum potential and to test economic hypotheses, detailed breakdowns of time allocations for both spouses are required. Peter Sloane (University of Aberdeen) pointed to the difficulties associated with defining and measuring "unpaid" and "paid" both inside and outside the home. Peter Elias (University of Warick) stressed the need for longitudinal time budget data in which changes in time allocation can be more directly considered. Fiorella Padoa Schioppa (LUISS University, Rome) noted that dealing with bureaucracy was one category of unpaid work time that was increasing over time. In the second session, two papers, based on data collected in British cohort studies, were presented exploring male-female differentials in wages and occupational attainment. The first by Heather Joshi and Marie-Louise Newell -- 'Pay Differences Between Men and Women: Longitudinal Evidence from the 1946 Birth Cohort' -- was an exposition of research that appeared as CEPR Discussion Paper No. 156. The seco.d paper by Peter Elias (University of Warwick) was entitled: 'Occupational Attainment at Age 23: A Comparison of Men and Women'. For most of this c%ntury, male hourly wages, on average, exceeded those of women by about 67 percent. However, in the 1970's this gap narrowed and has levelled off at around 45 percent. Furthermore, women tend to be over-represented in occupations of lower status which are characterized by lower rates of pay. Previous researchers have concluded that the gender discrimation legislation of the 1970's, such as the Equal Pay Act (1970), the Sex Discrimation Act (1975) and the Employment Protection Act (1975), was largely responsible for narrowing of these differentials.

Focussing on male-female pay differences, Joshi and Newell examined data collected in the Medical Research Council's National Survey of Health and Development consisting of a sample of some 4000 men and women born in 1946. These individuals were interviewed at age 26 in 1972 and contacted again at age 32 in 1977 when equal pay legislation was in force. Their main conclusion is that the equal opportunities legislation of the 1970's was partially successful in reducing the male-female earnings gap resulting purely from discrminatory labour market practises. A sizeable gap, equivalent to about 30% of women's pay, remains otherwise unexplained by factors like differential education and employment experience. These differences varied by social class and appeared to open out in the late twenties. The authors argue that to insure equality, policies that explicitly recognise the unequal division of family responsibilities between men and women together with a more serious approach to occupational segregation must be followed .

Elias explored gender inequalities in the labour market by looking at the determinants of occupational attainment by men and women. His analysis was based on data collected in the National Child Development Study -- a sample of individuals born in 1958 who were interviewed at ages 7, 11, 16 and 23. He noted that the earlier processes important to occupational attainment, revealed by scores on math and reading tests, are very similar for males and females. He also finds that males have both lower school leaving ages and migration propensities. Using a prestige ranking of occupations evaluated at age 23, his major finding is that males have a 30 percentage point "advantage" over females even after a variety of factors relevant to occupational attainmant (e.g. marital status, educational attainment, social background, etc.) are controlled for. Peter Sloane (Aberdeen University) remarked that research that he is pursuing on full-time employees in the New Earnings Survey has shown a recent decline in female occupation segregation. Others doubted if this would apply to part-time employment.

John Ermisch (National Institute of Economic Research and CEPR) and Peter Dolton raised issues concerning the statistical methods used by Elias and they suggested alternative strategies. Sara Connolly (Nuffield College, Oxford) added that research in Oxford, also based on the National Child Development Study support the findings on the processes of occupational attainment described by Elias. Peter Sloane challenged the validity of the econometric convention, adopted by Joshi and Newell, of allocating discrimation to the unexplained component of the wage gap. Brian Main (St. Andrews University) strongly defended this approach. Mary Gregory (St. Hilda's College, Oxford) remarked that the effect of membership in union versus non-union occuptions should be taken into consideration in the analysis of male-female wage differences.

Two papers on full- and part-time female wages were read in the third session: "Hourly Earnings of Female Part-time versus Full- time Employees' by Brian Main (St. Andrews University); and "Women's Full-time and Part-time Wages: Further Issues Relating to Sample Selection' by John Ermisch (National Institute of Economic and Social Research and CEPR) and Robert E. Wright (London School of Hygiene). Both studies used data collected in 1980 Women and Employment Survey conducted by the Department of Employment. In 1980, the mean hourly wage for full-time workers was $1.90 compared to $1.60 for part-time workers. Building on this observation, both papers were concerned with decomposing this differential and estimating how human capital characteristics, such as cumulative employment experience and education, influence these wage rates. Despite different model specifications and estimating techniques, both papers reached similar conclusions. About 8 percentage points of the differential between full and part-time wage rates is left unexplained by differences in work experience and education between these two groups of workers. Part-time workers are remunerated less for both full and part-time work experience and education. Unlike Main, Ermisch and Wright find that part-time workers lose less from time spent out of the labour force. These findings suggest that not only is part-time employment concentrated in jobs requiring low levels of skill but that some part-time workers are underpaid relative to their qualifications.discriminated against in the labour market. Both papers stressed the importance of analysing the differences between full and part-time wages in order to further the understanding of male-female wage differences and gender inequalities.

Heather Joshi was concerned with the occupational dimension of the full-time/part-time distinction and stated that this distinction is likely very different across occupation groups. She presented new results from the NSHD data for 1978 that the unexplained pay gap between the wages of male and female full- timers was of the order of 24%, and between part-time females and full-time males 40%.Peter Elias suggested that the decision to work part-time is likely an outcome of familiy responsibilities and he felt that part-time workers are a marginalized work force.

The final set of papers dealt with issues relating to the empirical analysis of male-female differentials were raised. Heather Joshi and Robert E. Wright reported on research building on CEPR Discussion Paper No. 157 : (Family Responsibilities and Pay Differentials). The purpose of this research was evaluate the effects that martial status and children have on male and female wage rates, and to comment on the statistical problems that emerge when one attempts to estimate such effects. The data analysed were from the MRC National Survey of Health and Development. It was found that married men, relative to umarried men, were 10 percent better paided. However, the difference between fathers and men without children is slight. The situation is the direct opposite for women. Childless women, on average, earn 30 percent more than mothers and married women were paid less than unmarried women. Parenthood has little effect on the labour force participation rate of men but clearly lowers particaption rates for women. A statistical analysis revealed that about 15 percent of the pay gap between mothers and childless women can be attributed to factors associated with the bearing and rearing of children (e.g. loss of work experience during pregnancy). A series of econometric tests were applied in order to validate these findings and described in a presentation entitled 'Endogeneity of Employment Experience and Marital Status.' This involved allowing for family responsibilities and female employment experience having been jointly determined. It was found that these possible statistical problems are of no serious consequence and, therefore, these conclusions, based on a simple model, are unchanged.

Peter Dolton (University of Bristol) discussed technical issues that arise in the statistical measurement of gender discrimination about which he and Gerry Makepeace (University of Hull) have recently published. In one paper they advocate looking at the degree of gender differences across the whole distribution not just at the means which is customary. In the second paper, they discuss the interpretation of sample selection bias effects in wage equations and argue that these should be interpreted carefully even when they appear to have the "wrong" sign. They also tabled a paper that examined the occupational choice of a sample of male-female British university graduates.

The workshop concluded with a general discussion chaired by David Stanton (Department of Employment). A large part of this discussion concentrated on drawing out the policy implications of the papers presented and methodological issues were also raised. The "human capital" approach to the understanding of gender differentials, which was used as the theoretical framework in many of the presented papers, was criticised, by Shirley Dex (University of Keele), for its underlying assumption that occupational segregation is an outcome of choice. Technical issues relating to the empirical estimation of gender discrimation were also voiced.

It was argued that there are two main reasons for trying to improve women's earning opportunities -- efficiency and equity. It was stressed that women's potential contribution to the development of the economy is held down by child bearing and rearing responsibilities and limited access to the whole spectrum of occupations. The equity case is particularly strong when one considers single parent families. A rising proportion of British children are growing up on the low income provided by state benefits because the labour market offers few/better alternatives to single mothers.

Britain was unfavourably compared to other countries, such as Sweden and Denmark, where public child-care facilities, the structure of the tax system, and entitlement to parental leave for both sexes make it easier for mothers of young children to continue their careers. It was noted that in Italy a system of parental leave for fathers had recently been introduced, but this programme has not been a success in terms of participation rates. It was pointed out that British institutions are not well suited to couples who choose to share roles as care-givers and providers. The continued disadvantage faced by women relative to men in the labour market perpetuates this situtation. The British Government's opposition to the European Directive on Parental Leave epitomizes the obstacles still facing British women.