For a start, the data are frequently misunderstood and misrepresented. Variations in hourly wages or bonuses between men and women are often interpreted – wrongly – as evidence of different pay for the same work. This sort of discrimination would, of course, be illegal. It would presumably be uneconomic too; if women were indeed willing to do the same work for less money, they would surely be over-represented in the highest paying jobs.
This isn’t just what critics are saying. The ONS has made a similar point when describing the national data. In its words, ‘the gender pay gap figures … do not show differences in rates of pay for comparable jobs, as they are affected by factors such as the proportion of men and women working part-time or in different occupations’.
Some suggest that this does not matter. For example, Richard Murphy has argued that the gender pay gap data are not supposed to address equal pay for equal work, but are ‘about precisely why women are not offered the same job opportunities as men, and so are paid less as a result’. Elsewhere, Ben Chu has suggested that ‘when it comes to various dimensions of equality in the workplace, some information is better than no information’.
I’d take issue with both. One problem is that the data do almost always seem to be interpreted as evidence of unfairness in pay. The misleading ‘#PayMeToo’ hashtag is simply an extreme example of this. But even a decent FT article on the gender pay gap was headlined ‘how women are short-changed in the UK’, and reported that ‘almost 90 per cent of women still work for companies that pay them less than male colleagues’. ‘Short-changed’? ‘Colleagues’? It is hard to escape the conclusion that women are earning less for the same work.
Unhelpfully, the government is encouraging firms to report aggregate data in the form of ‘women earn x% less than men’. This reinforces the impression of discrimination. As the ONS again says, ‘there is not one single measure that adequately deals with the complex issue of gender pay differences’, so why pretend that these reports can do this?
What’s more, even if correctly interpreted, the raw data are simply too crude. My colleague Kate Andrews has already pointed out many of the flaws. For example, firms are not required to distinguish between full-time and part-time employees. This is important, because hourly rates for full-time employment tend to be higher than for part-time. A substantial ‘gender pay gap’ may therefore simply reflect different patterns of working, with women more likely to work part time.
Of course, some would see this as evidence of discrimination. Richard Murphy, like many campaigners, seems to take it for granted that any difference in outcomes is a result of differences in opportunities, rather than choices freely made. But we simply cannot tell that from the data. Indeed, evidence that fails to fit this narrative is typically ignored. ONS data show that that women working part-time are actually paid more per hour, on average, than men.
The ONS has also done work suggesting that only 36% of the full-time gender pay gap can be explained by a particular set of factors that included occupation. Some have interpreted this to mean that the other 64% is a measure of discrimination. But this would be wrong, as the ONS itself says. Crucially, the factors quantified by the ONS did not include family structures, caring responsibilities, education and career breaks. These are factors that may also vary by gender but are not necessarily due to discrimination.
This is my response to Ben Chu’s point too. Normally, I’m also in favour of more information rather than less, especially where this may expose problems and encourage debate. Gender discrimination may well be one factor that explains why there are fewer women in higher-paid jobs. Nonetheless, the gender pay gap data are a poor way of measuring this.
That might not matter so much, of course, if the additional data did not come with additional costs. But they may actually make things worse. For example, they may encourage outsourcing of lower-paid jobs which happen to be taken by women (to avoid inclusion in a firm’s own return). They might deter the recruitment of women for initially lower-paid jobs, such as technical apprenticeships, that might be the gateway to higher pay in future.
Gender pay gap reporting may also lead to positive discrimination that unfairly penalises men and creates resentment towards women, including those who have advanced purely on merit. The relentless media coverage suggesting that the deck is stacked against working women may discourage others. And the administrative burden on employers is magnified by the additional work required to justify gender pay gaps that have entirely legitimate explanations.
Ideally, then, I’d argue that the gender pay gap reporting should be rolled back. Dropping them completely may not be realistic. But the government could reduce the frequency of reporting, currently annually, perhaps to once every five years. This would cut the compliance burden and discourage short-term measures to manipulate the data. In the meantime, the government should refocus on tackling the disadvantages that some women undoubtedly face – such as the availability of affordable childcare – rather than demonising employers.
In summary, the gender pay gap data are so widely misinterpreted that they add little value, and may even create new problems. They might have the potential to become a valuable tool in tackling the discrimination that still exists. For now, though, they seem to be generating more heat than light.