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

by Caroline Criado Perez

  Whenever I raise the issue of the unpaid-work imbalance between men and women, I am invariably faced with the comment, ‘But, surely, it’s getting better? Surely men are gradually doing more of their share?’ And at an individual level, sure, there are men who are doing more. But at a population level? Well, no, not really, because it turns out that the proportion of unpaid work men do is remarkably sticky. An Australian study found that even in wealthier couples who pay for domestic help, the remaining unpaid work is still distributed at the same male to female ratio, with women still doing the majority of what’s left.10 And as women have increasingly joined the paid labour force men have not matched this shift with a comparative increase in their unpaid work: women have simply increased their total work time, with numerous studies over the past twenty years finding that women do the majority of unpaid work irrespective of the proportion of household income they bring in.11

  Even when men do increase their unpaid work, it isn’t by doing the routine housework12 that forms the majority of the workload,13 instead creaming off the more enjoyable activities like childcare. On average, 61% of housework is undertaken by women. In India, for example, five out of women’s six daily hours of unpaid labour are spent on housework, compared to men’s thirteen minutes.14 It’s also rare for men to take on the more personal, messy, emotionally draining aspects of elder care work. In the UK up to 70% of all unpaid dementia carers are women,15 and female carers are more likely to help with bathing, dressing, using the toilet and managing incontinence.16 Women are more than twice as likely as men to be providing intensive on-duty care for someone twenty-four hours a day, and to have been caring for someone with dementia for more than five years.17 Female carers also tend to receive less support than male carers so they end up feeling more isolated and being more likely to suffer from depression – in itself a risk factor for dementia.18

  Men, meanwhile, have carried on engaging in leisure pursuits – watching TV, playing sports, playing computer games. US men manage to find over an hour more spare time per day to rest than their female counterparts,19 while in the UK, the Office for National Statistics found that men enjoy five hours more leisure time per week than women.20 And an Australian study found that what little leisure time women do have is ‘more fractured and combined with other tasks’ than men’s.21

  The upshot is that around the world, with very few exceptions, women work longer hours than men. Sex-disaggregated data is not available for all countries, but for those where the data exists, the trend is clear. In Korea, women work for thirty-four minutes longer than men per day, in Portugal it’s ninety minutes, in China it’s forty-four minutes, and in South Africa it’s forty-eight minutes.22 The size of the gap varies from country to country (the World Bank estimates that in Uganda women work an average of fifteen hours every day to men’s average of nine hours), but the existence of a gap remains more or less constant.23

  A 2010 US study on the imbalance between the amount of unpaid work done by male and female scientists found that female scientists do 54% of the cooking, cleaning and laundry in their households, adding more than ten hours to their nearly sixty-hour work week, while men’s contribution (28%) adds only half that time.24 The women in their data set also did 54% of parenting labour in their households, while male scientists did 36%. In India, 66% of women’s work time is spent on unpaid labour, while only 12% of men’s work is unpaid. In Italy, 61% of women’s work is unpaid compared to 23% of men’s. In France, 57% of their work is unpaid compared to 38% of men’s.

  All this extra work is affecting women’s health. We have long known that women (in particular women under fifty-five) have worse outcomes than men following heart surgery. But it wasn’t until a Canadian study came out in 2016 that researchers were able to isolate women’s care burden as one of the factors behind this discrepancy. ‘We have noticed that women who have bypass surgery tend to go right back into their caregiver roles, while men were more likely to have someone to look after them,’ explained lead researcher Colleen Norris.25

  This observation may go some way to explaining why a Finnish study26 found that single women recovered better from heart attacks than married women – particularly when put alongside a University of Michigan study27 which found that husbands create an extra seven hours of housework a week for women. An Australian study similarly found that housework time is most equal by gender for single men and women; when women start to cohabit, ‘their housework time goes up while men’s goes down, regardless of their employment status’.28

  The Economist isn’t alone in forgetting about women’s unpaid workload in their discussions of ‘work’. When business magazines like Inc publish think-pieces telling us that ‘science’ tells us ‘you’ shouldn’t work more than forty hours a week,29 or when the Guardian informs us that ‘your job could be killing you’ if you work for more than thirty-nine hours per week, they aren’t talking to women, because for women there is no ‘if’.30 Women do work well over this amount. Regularly. And it is killing them.

  It starts with stress. In 2017 the UK’s Health and Safety Executive (HSE) released a report on stress in the workplace which revealed that, in every age range, women had higher rates of work-related stress, anxiety and depression than men.31 Overall, women were 53% more stressed than men, but the difference was particularly dramatic in the age range thirty-five to forty-four: for men the rate was 1,270 cases per 100,000 workers; for women it was nearly double that, at 2,250 cases per 100,000 workers.

  The HSE concluded that this disparity was a result of the sectors women work in (stress is more prevalent in public service industries, such as education, health and social care), as well as ‘cultural differences in attitudes and beliefs between males and females around the subject of stress’. These may well be part of the reason, but the HSE’s analysis is sporting a pretty dramatic gender data gap.

  Since 1930 the International Labour Organization (ILO) has stipulated that no one should exceed forty-eight hours a week at work, by which they meant paid work.32 Beyond this number of hours workers start incurring health costs. But there is a growing consensus that things may be a little bit more complicated than that.

  A 2011 analysis of data collected on British civil servants between 1997 and 2004 found that working more than fifty-five hours per week significantly increased women’s risk of developing depression and anxiety – but did not have a statistically significant impact on men.33 Even working forty-one to fifty-five hours seemed to increase the probability of mental health problems in women. This was in line with a 1999 Canadian study34 and a 2017 analysis35 of six years of data from the Household Income Labour Dynamics of Australia Survey, both of which found that women had to work far fewer paid hours than men before their mental health started to deteriorate.

  But it’s not only about mental health. Swedish studies have found that moderate overtime work increases women’s hospitalisation and mortality rate, but has a protective effect for men.36 A 2016 US paper on the impact of long work hours over a thirty-two-year period found a similar gender disparity.37 Working moderately long hours (forty-one to fifty hours per week) was ‘associated with less risk of contracting heart disease, chronic lung disease, or depression’ in men. By contrast, such hours for female workers led to consistent and ‘alarming increases’ in life-threatening diseases, including heart disease and cancer. Women’s risk of developing these diseases started to rise when they worked more than forty hours per week. If they worked for an average of sixty hours per week for over thirty years, their risk of developing one of these diseases tripled.

  So, what’s going on? Is this all proof that women are in fact the weaker sex?

  Not exactly. In fact, the Australian study found that although the average man could work substantially longer hours than the average woman before his mental health was negatively impacted, there was one group of workers for whom the gender gap was much narrower. These workers are called the ‘unencumbered’, that is, workers with little to no c
are responsibilities. For the unencumbered, both men’s and women’s work-hour thresholds were much closer to the forty-eight hours stipulated by the ILO. The problem is, women aren’t unencumbered. It’s just that the work they do is invisible.

  When Ryan Gosling thanked his partner Eva Mendes at the 2017 Golden Globes for her unpaid work, acknowledging that without it he would not be on stage accepting an award, he marked himself out as a rare man.38 Far more usual is the impressively unperceptive man Guardian columnist Hadley Freeman wrote about in 2018: ‘“I have kids and I work full-time,” one boss crossly told a friend of mine who asked to have Fridays off. “Yes, and your wife quit her job to look after the kids,” my friend couldn’t quite bring herself to reply.’39

  This man simply couldn’t see – or perhaps didn’t want to see – all the unpaid work that gets done around him. The unpaid work that enables him to have kids and easily work full-time in paid employment. It doesn’t occur to him that the reason he doesn’t need Fridays off is not that he’s better than his female co-worker, but rather that, unlike him, she doesn’t have a full-time wife at home.

  Of course most male bosses in heterosexual relationships won’t have a full-time wife at home, because most women can’t afford to quit work entirely. Instead, women accommodate their care responsibilities by going part-time. In the UK, 42% of women compared to 11% of men work part-time, and women make up 75% of part-time workers.40 And part-time work is paid less per hour than full-time work – in part because it’s rare that a high-level post is offered as a job-share or with flexible working hours. Women end up working in jobs below their skill level that offer them the flexibility they need41 – but not the pay they deserve.42

  In Scotland in 2016 the average hourly gender wage gap was 15% – but this average hid the substantial disparity between full-time and part-time work.43 For those in full-time work the hourly gap went down to 11%, but the hourly pay gap between men working full-time and women working part-time was 32%. In 2017, median hourly pay for full-time employees across the UK was £14 per hour,44 compared with £9.12 for part-time employees.45

  Some call women’s segregation into low-paid work a choice. But it’s a funny kind of choice when there is no realistic option other than the children not being cared for and the housework not getting done. In any case, fifty year’s worth of US census data46 has proven that when women join an industry in high numbers, that industry attracts lower pay and loses ‘prestige’,47 suggesting that low-paid work chooses women rather than the other way around.

  This choice-that-isn’t-a-choice is making women poor. A recent Organisation for Economic Co-operation and Development (OECD) study found that the gender pay gap in hourly wages is substantially higher in countries where women spend a large amount of time on unpaid care compared to men.48 In the UK, women make up 61% of those earning below the living wage,49 and the Institute for Fiscal Studies has found that the gender pay gap widens over the twelve years after a child is born to 33%, as women’s careers – and wages – stagnate.50 The US pay gap between mothers and married fathers is three times higher than the pay gap between men and women without children.51

  Over time, these pay gaps add up. In Germany a woman who has given birth to one child can expect to earn up to $285,000 less by the time she’s forty-five than a woman who has worked fulltime without interruption.52 Data from France, Germany, Sweden and Turkey shows that even after accounting for social transfers that some countries employ to recognise the contribution women make through their unpaid care work, women earn between 31% and 75% less than men over their lifetimes.53

  This all leaves women facing extreme poverty in their old age, in part because they simply can’t afford to save for it. But it’s also because when governments are designing pension schemes, they aren’t accounting for women’s lower lifetime earnings. This isn’t exactly a data gap, because the data does mostly exist. But collecting the data is useless unless governments use it. And they don’t.

  Largely on the advice of international financial institutions such as the World Bank, the last two decades have seen an increasing global shift from social insurance to (often privately managed) individual capital account schemes.54 The payments a pensioner receives are directly based on their past contributions and the number of years during which the person is expected to collect benefits. This means women are penalised for the following: having to take time out for unpaid care work; early retirement (still a legal requirement in certain countries and professions); and for living longer.

  Other policies simply benefit men more than women. These include Australia’s recent tax concessions for pension funds (men are likely to have a higher pension pot),55 and the UK’s recent shift to auto-enrolment. As with many pensions around the world, this policy makes the standard error of forgetting to compensate women for the time they have to take out of the paid labour force to attend to their unpaid care load. As a result, women ‘miss out on vital contributions to their pension’.56 More unforgivable is the British system’s failure to account for the fact that women are more likely to have several part-time jobs in order to combine their paid and unpaid workloads.57 In order to qualify for the auto-enrolment pension, a worker must earn at least £10,000 a year. But while many women do earn past this threshold, they earn it from multiple employers – and combined earnings are not counted towards the threshold. This means that ‘32% or 2.7 million employed women will not earn enough to benefit from auto-enrolment compared to 14% of employed men’.58

  A counterpoint is provided by Brazil, Bolivia and Botswana, which have achieved close to universal pension coverage and smaller gender gaps ‘thanks to the introduction of widely available non-contributory pensions’.59 Women in Bolivia are credited with one year of pension contributions per child, up to a maximum of three children. As a side benefit (and a more long-term solution to the problem of feminised poverty), pension credits for the main carer have also been found to encourage men to take on more of the unpaid care load.60 Which raises the question: is women’s unpaid work under valued because we don’t see it – or is it invisible because we don’t value it?

  Alongside addressing male bias in pensions, governments must address feminised poverty in old age by introducing policies that enable women to stay in paid work. That starts – but certainly doesn’t end – with properly paid maternity leave.

  EU countries with comprehensive support for working parents have the highest rates of female employment.61 Numerous studies world wide have shown that maternity leave has a positive impact on women’s participation in the paid labour market.62 This impact is seen not only in the raw numbers of women employed, but also in the number of hours they work and the income they earn. It has been shown to be particularly beneficial for low-income women.63

  There is a caveat, however: not all maternity-leave policies are made equal. The length of time and the amount of money on offer matter. If women aren’t given enough time off, there is a risk they will leave the paid labour force entirely,64 or transition to part-time work.65 When Google noticed that they were losing women who had just given birth at twice the rate of other employees, they increased their maternity leave from three months at partial pay to five months at full pay. The attrition rate dropped 50%.66

  With the exception of the US, all industrialised countries guarantee workers paid maternity leave,67 but most countries aren’t hitting the sweet spot either in pay or the length of leave allowed. And they certainly aren’t hitting them both together. A recent Australian analysis found that the optimum length of paid maternity leave for ensuring women’s continued participation in paid labour was between seven months to a year,68 and there is no country in the world that offers properly paid leave for that length of time.

  Twelve countries in the OECD offer full replacement wages, but none of these countries offers more than twenty weeks, with the average being fifteen weeks. Portugal, for instance, one of the countries that offers 100% replacement wages, offers only six weeks of l
eave. Australia, by contrast offers eighteen weeks of maternity leave – but at 42% of earnings. Ireland offers twenty-six weeks – but at only 34% of earnings. For women in these countries the full length of time they’re technically allowed to take off can be, as a result, academic.

  British politicians like to boast (particularly in the run-up to the EU referendum) that the UK offers a ‘more generous’ maternity leave than the fourteen weeks mandated by the EU’s 1992 Pregnant Workers Directive.69 This is technically true, but it doesn’t mean that women in the UK get a good deal in comparison to their European counterparts. The average length of paid maternity leave across the EU is twenty-two weeks.70 This figure hides substantial regional variation in both pay and length. Croatia offers thirty weeks at full pay, compared to the UK’s offering of thirty-nine weeks at an average of 30% pay. In fact a 2017 analysis placed the UK twenty-second out of twenty-four European countries on the length of ‘decently paid maternity leave’ it offered its female workforce (1.4 months).

  And now that Britain is leaving the EU, the country is likely to fall even further below its European neighbours. Since 2008, the EU has been trying to extend its maternity-leave ruling to twenty weeks on full pay.71 This proposal was stuck in stalemate for years, and finally abandoned in 2015 thanks in no small part to the UK and its business lobby, which campaigned strenuously against it.72 Without the UK, the women of the EU will be free to benefit from this more progressive leave allowance. Meanwhile Martin Calla-nan (now a Brexit minister) made a speech to the European Parliament in 2012 in which he included the Pregnant Workers Directive in his list of the ‘barriers to actually employing people’ which ‘we could scrap’.73