Invisible Women: Exposing Data Bias in a World Designed for Men

social security benefits early to pay for maternity leave – and then delaying their retirement payments to offset the costs. It’s easy to see why the idea is attractive: it comes without a cost, at least to the government. But it is far from cost-free to women. The gender pay gap and the time women take off to care for children already results in lower social security benefits for women, a problem this policy will exacerbate.⁸⁵ And given women live longer and spend more of their later years in ill health they arguably need more money for retirement, not less.⁸⁶ As a result, the main impact of this policy would be to compound the problem of feminised old-age poverty.

US universities provide another example of how gender-blind leave policies can end up discriminating against women. US academics in the tenure-track system have seven years to receive tenure after getting their first academic job or they’re fired. This system is biased against women – especially women who want to have children, in part because the years between completing a PhD and receiving tenure (thirty to forty) coincide with the years these women are most likely to try for a baby.⁸⁷ The result? Married mothers with young children are 35% less likely than married fathers of young children to get tenure-track jobs,⁸⁸ and among tenured faculty 70% of men are married with children compared to 44% of women.⁸⁹

Universities have done little to address this – and even those that have tried, have often done so in gender-blind ways that may end up exacerbating the problem they were trying to solve.⁹⁰ In the 1990s and early 2000s, a number of US universities adopted what was intended as a family-friendly policy: parents would receive an extra year per child to earn tenure. But it isn’t gender-neutral ‘parents’ who need this extra year. It is specifically mothers. As the University of Michigan’s Alison Davis-Blake drily noted in the New York Times, ‘giving birth is not a gender-neutral event’.⁹¹ While women may be (variously) throwing up, going to the toilet every five minutes, changing nappies or plugged into their breast pump during this extra year, men get to dedicate more time to their research. So instead of giving a leg up to parents, this policy gave a leg up to men, and at women’s expense: an analysis of assistant professors hired at the top fifty US economics departments between 1985 and 2004 found that the policies ultimately led to a 22% decline in women’s chances of gaining tenure at their first job. Meanwhile men’s chances increased by 19%.⁹²

The analysis came in a working paper and the totality of its findings have been challenged⁹³ – but given what we already know about the disparity between mothers and fathers gaining tenure, and given what the data tells us about who actually does the care work (not to mention the gestating-and-giving-birth-and-breast feeding work) there seems little reason not to make such policies dependent on who is actually carrying the child, and/or who the main carer is. To date, this has not happened.

This is not to say that paternity leave is not important. It certainly is. Beyond the simple matter of fairness (fathers should have the right to be involved in their children’s lives), the data we have shows that properly paid paternity leave has a positive impact on female employment. At close to 80% by 2016, Sweden has the highest female employment figures in the EU.⁹⁴ It also has one of the highest levels of paternity-leave uptake in the world, with nine out of ten fathers taking an average of three to four months’ leave.⁹⁵ This compares with a more typical OECD level of one in five fathers taking any parental leave at all – falling to one in fifty in Australia, the Czech Republic and Poland.⁹⁶

This disparity is unsurprising: Sweden has one of the most generous (and, when it was introduced, innovative) paternity-leave policies in the world. Since 1995, Sweden has reserved a month of parental leave (paid at 90% of earnings) exclusively for fathers. This month cannot be transferred to the mother: the father must use this leave or the couple lose it from their overall leave allowance. In 2002, this increased to two months and in 2016 it was further increased to three months.⁹⁷

Prior to the introduction of the ‘use it or lose it’ leave for fathers, only about 6% of men in Sweden took paternity leave, despite the fact that it had been available for them since 1974. In other words, men didn’t take the leave on offer until forced to by the government. This pattern has been repeated in Iceland, where the introduction of a ‘daddy quota’ doubled the amount of leave taken by men, and in South Korea, where the number of men taking leave rose more than threefold following the introduction in 2007 of a father-specific entitlement.⁹⁸ Proving, however, that no good data goes unignored, in 2015 the UK government saw fit to introduce a shared parental leave policy with no allowance reserved exclusively for men. Predictably, the take-up has been ‘woefully low’, with just one in a hundred men requesting leave in the twelve months after it was introduced.⁹⁹

The introduction of a daddy quota has not been a marked success in Japan, but this is in no small part due to a design that doesn’t account for either the gender pay gap or women’s bodily reality. While fathers have two months reserved for them out of a possible fourteen months of shared leave, after the first six months of leave, pay decreases from two-thirds of the parent’s salary to just half. Given that women need to recover from pregnancy and giving birth, and may be breastfeeding, they are most likely to take leave first, leaving the higher earner (Japanese men earn on average 27% more than Japanese women) to take the biggest salary hit.¹⁰⁰ It is unsurprising, therefore, that only 2% of Japanese men take the months they are entitled to.¹⁰¹ Japan’s