Is there such thing as a gender wage gap? Ben Southwood is Head of Macro Policy at the Adam Smith Institute Introduction There is a persistent ‘gap’ between the average annual, weekly, and hourly wages earned by men and women in developed countries. Within the EU, it ranges from close to 25% in Estonia to more like 3% in Malta; within the OECD from over 30% in South Korea to around 5% in New Zealand. Many campaigners worry that this gap is down to discrimination. Even if it’s not down to direct discrimination when employers hire or set pay, activists worry that it may be down not to preferences for different sorts of work, but social pressure or discrimination at another level. Talented women may be stuck in lower-ranked roles, earning the same in absolute terms, but less relative to their talent. Women might be directed away from certain careers and into others by their parents or social groups treating girls and boys differently, or by schools, where teachers could potentially lead girls into humanities and boys into sciences. Typically women take on more childcare work when couples have children, and in many countries women do more housework. If these ‘gaps’ are down to socio-cultural conditioning or pressure we might have concerns there. According to some left-leaning activists, any differences between men and women’s career trajectories, even on average, must be down to ‘structural’ or ‘socially-constructed’ factors, since they believe men and women are biologically very similar, at least in all career-relevant ways. On one view these socially constructed differences aren’t necessarily bad for women, if women rate their lives as equally (or more satisfying) and happy as men’s. But there is also some evidence of biological differences between men and women. Whether or not differences come from biology or society I will assume that what matters are people’s preferences, happiness and satisfaction with life. The situation, as suggested above, appears quite different across developed countries. Indeed, general perceived gender discrimination also varies almost as widely as the gender wage gap. But, in sketching a broad picture of how I think the landscape is best understood, I will take research from advanced countries to be sufficiently general such that research for, say, Sweden helps inform us about the situation in, say, Switzerland. And data from the UK will be taken to be informative about relations in the USA. The UK situation According to the 2014 Annual Survey of Hours and Earnings, in the UK, women working full-time earned a mean of 9.4% less than men per hour while women working part-time earned 5.5% more per hour. Since full-time workers earn more, women earned 19.1% less than men on average (again per hour). This was the lowest figure on record—back in 1997 women earned closer to 30% less than men. This aggregate figure obscures some very interesting variation. For example, among those working 30 or more hours a week, women aged 22-29 earned 1.1% more than men aged 22-29, and also very slightly more than men aged 30-39. The aggregate gap only appears for women aged 40 or more. Combined with the overall decline, this could either suggest: first that men and women’s labour market experiences are getting more similar over time and newer cohorts suffer less in the way of discrimination; or second that something happens later in life to open up a gap. I shall lean more towards the second explanation (though both have a role to play) because a body of persuasive evidence suggests that liberal labour market policies and reduced overall societal gender discrimination lead to more difference between the genders not less. Women use their freedom and increased wealth to express gendered characteristics. Women earn more than men before age 40 because they earn more and better qualifications. They earn less later because they prefer to work less and more flexibly, often, but not always, because of childcare preferences or responsibilities. Discrimination The simplest form of what I call the ‘discrimination hypothesis’ is that men and women who do the exact same job earn different amounts. Undoubtedly there are cases where this has been true—including some reality TV programmes—but we have powerful reasons to expect that the practice of paying men and women different amounts for the exact same job wouldn’t be persistent or widespread. Basic economics tells us that discrimination is not a sustainable strategy in a moderately competitive market—even if a large proportion of firms are run by sexists. All we need is open markets and a small fraction of non-sexist people, or just people whose greed overpowers their sexism. This, shown early on by University of Chicago Nobel Prize-winning economist Gary Becker, is because that firms that pay women less than their productivity are losing out on profit their less prejudiced or more greedy competitors can freely grab. If a woman can do £20 of work an hour and is being paid only £10, then any firm can try and big her away from her current occupation by offering her up to £19.50 or £19 or whatever leaves them their minimum profit margin. And the evidence seems to bear this out in practice. According to Erica L Groshen’s 1991 paper The Structure of the Female/Male Wage Differential: Is It Who You Are, What You Do, or Where You Work? even back then women and men (in the USA) were earning exactly the same for doing the same job. At the very least, we are not seeing employers miss out on very obvious profit opportunities. They are not allowing their competitors to bid away, for example, their female clerks by offering wages ever closer to the male wage for the same job—or that process has already occurred. It’s still possible that men are getting more aggressively promoted than similar women. But the data comes out against that too. George Levi-Gayle, Limor Golan and Robert A Miller, in their 2012 paper Gender Differences in Executive Compensation and Job Mobility find that the reason (American) women make up fewer high level executives is mainly because they trade down to lighter work or leave the workforce entirely. If you just look at women who ‘stay’ in the workforce, they earn more than men and are more likely to get promoted: Fewer women than men become executive managers. They earn less over their careers, hold more junior positions, and exit the occupation at a faster rate. We compiled a large panel data set on executives and formed a career hierarchy to analyze mobility and compensation. We find, controlling for executive rank and background, that women earn higher compensation than men, experience more income uncertainty, and are promoted more quickly. Among survivors, being female increases the chance of becoming chief executive officer. The unconditional gender pay gap and job-rank differences are primarily attributable to female executives exiting the occupation at higher rates than men. And overall this should not be surprising, given the data on the overall wage gap mentioned above: it doesn’t exist until later ages, when women take time out to work less or spend more time looking after their children. Fertility The link between fertility and the gender wage gap is very clear, and the mechanism is obvious. Women take time out of the workforce to have their children, and they often take time out to raise them—instead of, for example, sending them to childminders, nurseries, and so on. When their kids go to school they often prefer to work part-time and flexibly, to allow them to take their children to school, help with homework, make dinner and so on. These are not legal requirements or imposed by firms, and are down either to women’s preferences or social pressure. A wide range of studies provide strong evidence for this conclusion from various angles. A 2014 paper - Fertility Effects on Female Labor Supply: IV Evidence from IVF Treatments by Petter Lundborg, Erik Plug and Astrid Würtz Rasmussen — studied Danish women who applied for IVF treatment. Their results, displayed in the charts above, show steadily inclining wages for women until they have a child or get IVF treatment. If they are successful and have a child, their wages drop off a cliff as they take time off. When they start growing again, it is at a lower trend, consistent with them having lost skills or built up skills less, or their being unwilling to work long hours. This drop looks a lot like the drop from non-IVF fertility, suggesting that it truly comes from fertility. Another 2014 paper — The Gender Pay Gap Across Countries: A Human Capital Approach by Solomon Polachek and Jun Xiang - looks around the world rather than at changes within one population. They find that fertility rates are one of the main drivers of the wage gap, along with top tax rates and the average age gap between men and women at first marriage (which is possibly a proxy for general social discrimination against women). This paper explicitly concentrates on labour market institutions that are related to female lifetime work that affect the gender wage gap across countries. Using ISSP (International Social Survey Programme), LIS (Luxembourg Income Study) and OECD wage data for 35 countries covering 1970-2002, we show that the gender pay gap is positively associated with the fertility rate, positively associated with the husband-wife age gap at first marriage, and positively related to the top marginal tax rate, all factors which negatively affect women’s lifetime labour force participation. A third 2014 paper, again by Solomon Polachek but this time working with Xu Zhang and Xing Zhou and entitled A Biological Basis for the Gender Wage Gap: Fecundity and Age and Educational Hypogamy, reinforces this by looking at the effect of China’s one-child policy on the gender wage gap. They find that the implementation of the policy, which drove fertility in China down by 1.2-1.4 births per woman, led to hugely narrowed educational gaps, a narrower division of labour in the home, and a diminishing pay gap twixt husband and wife. Women themselves, according to a report out this year from the ILO, rate inadequate non-discrimination laws as the least important issue needed to help women advance in business. By contrast they rate family responsibilities and societal roles of women as the most important two issues. There’s a lot more evidence out there, although fertility is far from the only cause. Even women who don’t have kids tend to prefer shorter working days and shorter working weeks; less competition; and jobs whose benefits are mainly non-pecuniary. Preferences The interesting question is mainly whether the gender wage gap is down to women’s preferences or down to firm or societal pressure. One way of deciding this question is looking to see if there is firm discrimination or societal pressure. Another is looking at women’s preferences. A late 2014 paper, Life Paths and Accomplishments of Mathematically Precocious Males and Females Four Decades Later, from authors David Lubinski, Camilla P Benbow, and Harrison J Kell does just that. Two cohorts of intellectually talented 13-year-olds were identified in the 1970s (1972–1974 and 1976–1978) as being in the top 1% of mathematical reasoning ability (1,037 males, 613 females). About four decades later, data on their careers, accomplishments, psychological well-being, families, and life preferences and priorities were collected. Their accomplishments far exceeded base-rate expectations: across the two cohorts, 4.1% had earned tenure at a major research university, 2.3% were top executives at ‘name brand’ or Fortune 500 companies, and 2.4% were attorneys at major firms or organizations; participants had published 85 books and 7,572 refereed articles, secured 681 patents, and amassed $358 million in grants. For both males and females, mathematical precocity early in life predicts later creative contributions and leadership in critical occupational roles. On average, males had incomes much greater than their spouses’, whereas females had incomes slightly lower than their spouses’. Salient sex differences that paralleled the differential career outcomes of the male and female participants were found in lifestyle preferences and priorities and in time allocation. The results are very striking. Women favour short working weeks, clean working conditions, respect and flexibility. Men want merit pay, the ability to take risks, a well above average salary and to work with things as well as or as opposed to people. These differences in preferences surely explain a large amount of the differences in career outcomes. Women simply desire different career trajectories to men. These different career trajectories (especially shorter and more flexible work weeks) lead to different pay. This difference in pay isn’t necessarily fair or unfair—it simply reflects lower skills built up due to lower and less sustained time on the job. In fact, a 2014 paper from Claudia Goldin ingeniously explains exactly why these different preferences result in a gender wage gap. She finds that there is no or next-to-no gap in any industry where there are constant returns to work. Any sector where your tenth hour in a day produces just as much value as your first has no gender wage gap. By contrast, those sectors, for example law, where one must build up experience and capital, have increasing returns to hours. A sixty-hour week produces more than double a thirty-hour week. Women’s preference for working less, combined with the inherent technology in a field, leads to a gender wage gap. Men and women are having different sorts of careers—but are they differentially satisfied? Perhaps unsurprisingly, the fact that careers are in line with preferences tends to lead to greater job satisfaction. Although, interestingly, there is a wide and growing literature on the ‘paradox of female happiness’—women are tending to report lower happiness the more they focus on have a traditionally-male career, and the less they focus on traditionally-female activities. For example, Marianne Bertrand’s 2014 paper Career, Family, and the Well-Being of College-Educated Women found: The biggest premium to life satisfaction is associated with having a family. While there is also a life satisfaction premium associated with having a career, women do not seem able to ‘double up’ on these premiums. A qualitatively similar picture emerges from the emotional well-being data. Among college-educated women with family, those with a career spend a larger share of their day unhappy, sad, stressed and tired. Betsey Stevenson and Justin Wolfers found in a 2009 paper, that: By many objective measures the lives of women in the United States have improved over the past 35 years, yet we show that measures of subjective well-being indicate that women’s happiness has declined both absolutely and relative to men. The paradox of women’s declining relative well-being is found across various datasets, measures of subjective well-being, and is pervasive across demographic groups and industrialized countries. Relative declines in female happiness have eroded a gender gap in happiness in which women in the 1970s typically reported higher subjective well-being than did men. These declines have continued and a new gender gap is emerging—one with higher subjective well-being for men. A paper from Brookings scholars Carol Graham and Soumya Chattopadhyay from 2012 backs this finding up: that women tend to report lower satisfaction, well-being and happiness as they move more into traditionally male areas. Gender equality A further reason to question the discrimination narrative comes from a wide range of studies that find that countries with more gender equality in general see bigger differences in characteristics between the sexes. To survey a few of the major findings: More gender egalitarian countries often have more sex-segregated occupations (Charles 2011) People choose more sex-typed subjects for academic study in more developed, egalitarian countries (Charles & Bradley 2009) The more gender equal a country, the bigger the gap in mental rotational task scores (a key measure of visuospatial skills) between girls and boys (Lippa, Collaer & Peters 2010) Since the US has become more gender egalitarian—comparing the cohort of youths surveyed in 1979 to 1997—gender differences in competitiveness have come to account for more of the gender wage gap (McGee, McGee & Pan 2014) Women differ across the world in personality, but most in the countries where women have more rights and freedoms (Schmitt, Realo, Voracek & Alli 2008) Conclusion The studies from the 80s and 90s are increasingly irrelevant. Women have responded to women’s liberation by in many ways increasing their expression of their gender with their new freedoms. This is not a bad thing. As we’ve seen, it does not lead to less happiness for women, or less satisfaction with their career or jobs. Markets do well at sending people where their skills are useful and balancing out their own preferences about work. Women prefer shorter weeks and flexible work. This, under current technology, makes them less productive workers. Firms, caring only about worker productivity, pay women less because they produce less. Where women produce the same, they pay them the same. As we’ve seen, women in their 20s and 30s earn more than men, on average in the UK. Women who don’t leave the work force make more than men and are more likely to be promoted to chief executive. Gender discrimination in the workforce is yesterday’s problem. There is an interesting philosophical question as to whether the preferences that generate these differences come from social construction of gender identity—as many feminists argue—or from biological sex differences. But even if society is to blame, it’s not clear that the wage ‘gap’ or discrimination model are at all useful in understanding sex and gender in the labour market. And if society is to blame it’s far from clear that there is anything we can do about this difference (even if we want to when it appears to promote satisfaction and wellbeing). There may be such thing as a gender wage ‘gap’, but it doesn’t seem like we should care. Gender discrimination in the workforce is yesterday’s problem