It looks like progress. Slow, to be sure, but progress nonetheless. Twenty-five years ago, women’s hourly earnings were on average about 24 per cent less than men’s. That gap has fallen to 19 per cent. If you look at total average earnings, which takes account of the fact that women are less likely to work and that if they do work they work fewer hours on average, then the gap in total earnings has fallen from well over 50 per cent to 40 per cent. These are still big gaps, but progress seems tangible. Both the employment and wage gaps have narrowed.

Yet even this feeble progress has stalled. The hourly wage gap has barely shifted in the past 15 years. Moreover, virtually all the progress there has been is related to the huge strides that women have made in catching up with, and then overtaking, men in terms of their educational attainment.

Those pay and employment gaps remain despite the fact that women are now considerably more likely than men to be graduates. The lack of progress on hourly wages since the mid-2000s has occurred despite those educational gains. If women hadn’t become better-educated than men, we might well be looking at pay and employment gaps barely changed in the past quarter of a century.

That is one of the deeply disappointing conclusions of work being published today by my colleagues the Institute for Fiscal Studies, with co-authors at the London School of Economics, as part of our review of inequalities led by Angus Deaton, the Nobel prize-winner.

The study also reveals that there has been a big social change in where the worst of the inequality lies when it comes to hourly wages. It used to be that gaps in hourly wages were most pronounced for the least well-educated women, those whose highest qualifications were GCSEs or below. With the help of the national living wage, these more poorly paid women have closed the gap on their male equivalents. It may not have been intended as such, but probably the most successful policy in helping to bridge at least one gap between men and women has been the introduction of, and continued increases to, the minimum wage.

The least well-educated women remain much less likely to be in work at all. Among those in work, though, it is now the highest-educated women who suffer the biggest hourly wage penalty. Women are missing from the top of the earnings distribution. Nine out of ten at the very top (the top 0.1 per cent) are men. A woman 90 per cent of the way up the female earnings distribution earns about the same as a man 75 per cent of the way up the male distribution.

Why are these gaps so persistent? The Equal Pay Act is 50 years old. There has been an awareness of these inequalities for a long time, often considerable political and media focus and genuinely substantial public investment in childcare provision.

Most of the gap is to do with children. Before the birth of a child, women are just as likely as men to be in work and earn only a modest amount less, on average. That changes when children come along. Many give up work altogether, many others start working part-time or move to work closer to home. That causes an immediate drop in earnings relative to men, but also a slow burn as they fall further and further behind in terms of hourly wages. That slow burn has become more important as more and more women have started off in graduate jobs with high earnings potential but have lost out on the earnings growth that tends to come in those sorts of jobs in their thirties and forties. There is much less in the way of pay progression at this stage in life for both men and women with lower levels of education.

All this doesn’t happen simply because women tend to earn less before childbirth and hence couples make the rational economic decision for the woman to focus on childcare and for the man to focus on earning. Even where the woman in a couple earns more than the man before giving birth, she is much more likely to be the one to give up work or to move to part-time work. Social attitudes and norms shape individual preferences and these are sticky over time.

Attitudes are closely related to outcomes. One fascinating cross-country study showed a very close correlation between people’s views on the appropriate role of women and the earnings penalty to having children. The more people felt that women with young children should stay at home, the less mothers earned.

Where there is hope for progress is that there appears to be a feedback loop from policy to attitudes. In countries where the policy environment is most supportive of maternal work, such as those in Scandinavia, the idea that mothers should stay at home to care for their young children has little support. And attitudes change with experience. West German mothers in closest contact with mothers from East Germany, for whom the rates of maternal employment were far higher than in the West, adjusted their work patterns to almost completely mirror those of their new peers.

After all this time, it is easy to start to believe that these gender differences are fixed. They needn’t be. An accumulation of policies consistently supporting a more equal sharing of responsibilities between parents, or big policy reforms challenging gender roles, may help to build up a change in attitudes that leads to permanent change in norms.

Irrespective of any weight we might put on gender inequalities, there are huge economic costs associated with the status quo. Even expensive policies have the potential to pay for themselves if they successfully ensure that the talents of both women and men are put to their most productive uses, whether in the labour market or at home.

This article was first published in The Times and is reproduced here with kind permission.