Marriage is bad news for women's careers. And, argues Jonathan Gershuny in the latest part of our series on work and the family, it is the husbands at home, rather than the boss at work, who is responsible

What follows is a familiar fairy story told in an unfamiliar way - Cinderella back to front - which has some striking implications for public and other sorts of policy. It starts with occupational status. Everyone knows that women marry up, men marry down. This is a "well established fact": men are more likely to be married to wives with lower or equal-status jobs, than they are to be married to higher status women; employed women tend to be married to husbands with equal or higher status jobs than lower. Just like the fairy story: the prince is, statistically speaking, likely to marry a scullery maid.

But research by our team at Essex University suggests that marriage causes women's careers to decline relative to their partners'. This is because they tend to take more career breaks - normally to have or look after children.

Job prestige is measured by the Hope-Goldthorpe scale. It is approximately normally distributed with a mean score of 46; the maximum score is 82 (this would be, for example, a judge or a surgeon) and the minimum score is 18 (a kitchen porter, general labourer).

É«ºÐÖ±²¥

Our team at the Essex Research Centre has used data from the British Household Panel Survey which monitors the lives of people in a sample of households (see panel below).

In the 1994/95 survey wives' mean HG score is 45, husbands' 50, giving an average difference of more than five points. Some 56 per cent of women are married to men more than two points higher on the scale, while only 31 per cent of men are married to women more than two points higher: allowing for a little hyperbole, the fairytale result. And essentially the same fairytale result comes out of any cross-sectional survey.

É«ºÐÖ±²¥

But a rather different picture emerges when we look at longitudinal evidence. We can combine information for the successive waves of the BHPS to build up a continuous picture of the whole of our respondents' working lives. And since the BHPS interview sample includes all adult members of BHPS households, and we have complete marital and cohabitation histories for all our sample members, we can match and compare the evolution of spouses' occupational status through the course of their marriages.

In figure one (right) we have a plot of the mean HG scores of husbands and wives who got married in the 1970s and remained together until at least 1993, starting from the year that each got married, for each of the first 15 years of marriage (the latest data were collected in 1994/5). Men show a regular upwards trend in their occupational prestige throughout this period; wives start off with, on average, a somewhat lower average occupational prestige than their husbands (three HG points) - but show an overall decline in occupational prestige over the period, so the spouse gap increases dramatically to about nine HG points after 15 years.

It emerges that husbands and wives when they get married have rather more similar occupational status than they have later in their marriages. It is not so much that spouses marry up or down - but rather that spouses' occupational standings diverge during the marriage. Part of the explanation of the "well established fact" is this: we get the status difference between the married partners because in cross-sectional survey evidence we are looking, on average, part of the way through the marriages, somewhere in between the initial rather narrow occupational status gap and the eventual much larger one.

Why do spouses' careers diverge in this spectacular fashion? The answer is provided by one of the core concepts of economic sociology: the process of the formation of economically salient "human capital". The more human capital the worker possesses, the better the job prospects; and human capital comes in part from family origins and education, and in part from recent and relevant work experience. The next pair of pictures divides marriages in the 1970s into two groups. It contrasts the occupational status trajectories of husbands and wives in couples which have adopted two different household work strategies: on one side are those couples where the wife continues to build up her human capital in a regular way during the marriage, by taking only a limited period (up to 23 months in the first 15 years of marriage) out of paid employment for family reasons; and on the other, are couples in which the wives' accumulation of salient experience in employment is more intermittent, they take longer (24 months or more) family breaks.

The "wife takes only short breaks" couples show more apparent variability in status year by year, since, as in the population as a whole, the size of this group in the BHPS is rather small (76 couples); even taking the couples married in the 1970s, only one in seven falls into this category. Taking the small numbers into account this picture tells us that there is little if any difference between the spouses' status in this group, and the husbands and wives have roughly similar upwards prestige trajectories.

It is among those couples where the wife takes long domestic-related work breaks that the divergence occurs. The initial spousal prestige gap is somewhat larger than in the previous picture, and, after five years or so, as the wives' accumulating domestic breaks take effect, their mean occupational prestige score takes a nose-dive.

Those women who take the lengthy domestic breaks have dramatically worse occupational status trajectories than those who take short ones. Those wives who take relatively high levels of responsibility for their household's domestic tasks by taking long family breaks, have the substantial declines in career attainment. It is those few wives who maintain their presence in the labour force (and thus manage to share domestic responsibilities less inequitably with their husbands) who also maintain their positive career trajectory.

In short, it is not that the prince marries the scullery maid. Rather, the prince marries a princess and turns her into a scullery maid.

É«ºÐÖ±²¥

É«ºÐÖ±²¥

The longitudinal data shows a within-marriage process in which the wife's occupational status is gradually lowered relative to the husband's as a result of the gendered distribution of domestic responsibilities.

This is an important finding. We know from other sources that women's earnings lag behind men's, even when we control appropriately for education and skill levels. It has until now been assumed that this reflects discriminatory practices in recruitment or promotion at work. As a result, in the developed world, we have adopted ever more stringent policies to outlaw sex discrimination at work. But the longitudinal BHPS evidence suggests that the origin of the discrimination may not be the workplace but the home. It may be that husbands at home, and not bosses at work, are responsible for the persistence of wives' low wages. And if this is the case, public policy may be misdirected.

And there is also a policy message at a rather more intimate level. Remember that it is in those households with a relatively large initial mean gap between the spouses' occupational statuses that the wives are somehow persuaded to take long domestic work breaks. Presumably the personal life-strategy implication is: if you are a scullery maid interested in occupational advancement, forget the prince.

And this is just one example of the sorts of results that are emerging from what is a new field of survey-based research. The BHPS, like the other panel surveys across Europe, contains a very wide range of longitudinal data, on employment, earnings, family structures, attitudes, health, housing and other things. These new longitudinal data sources allow us to move forward from the mere observation of statistical associations, which is ultimately all we can do with cross-sectional data, to the investigation of the real causes of social change, by looking at events and their consequences in the actual, historical sequences in which they really happen.

Jonathan Gershuny is professor in the research centre on micro-social change, Essex University.

MARRIAGES OF INCONVENIENCE

When a man gets a settled job does he then want to find a wife, or to start a family? Do people with poor job prospects have children to attract social benefits? When a wife gets a promotion does the husband take a part-time job? Or when the husband of a full-time employed wife fails to do his share of the housework, does she leave him or her job? If she leaves her job to care for the family, what job does she get when the children grow up, or if he runs off with his secretary?

New empirical, interdisciplinary social research agenda looks at these sorts of interactions among work, income and family life. The Essex Research Centre is investigating changes in household dynamics.

It collects and uses the British Household Panel Survey. This name is somewhat misleading. A panel study is one which follows its subjects over time. But the household itself does not really exist over time. People leave, set up on their own or co-reside with others who were not in the original household. Indeed it is precisely this fluctuation that makes households interesting and forms a good part of the subject of the research.

So the panel is in fact made up the individual members of the households in the first "wave" of interviews, plus their natural descendants. In September 1991 we established the initial random sample of 5,500 households, containing about 10,000 adults and 3,000 children. In each subsequent wave we follow all the wave one sample members and their descendants, into whatever households they are residing in, and then interview all the adults (aged 16+) and also children aged 11-15.

Year by year the sample remains representative - as it reproduces itself the same way as the population does.

É«ºÐÖ±²¥

The largest concentration of work on BHPS is at the research centre. The centre includes, in addition to the survey specialists, seven professors and 15 postdoctoral researchers. The interdisciplinary focus means that we have a strong representation of senior, interdisciplinarily inclined sociologists working on change in the United Kingdom class structure, household structures and social polarisation, on careers and household work strategies, and on new occupational classifications.