Gender differences in subject choice leads to gender pay gap immediately after graduation

October 5^th is the deadline for big companies to report their gender pay gaps. In 2019 – before the pandemic disrupted data collection – women were paid 16% less per hour than men on average. The gap in average annual earnings was even larger, at 37%, since women are much more likely to work part-time.

There is a lot of focus on the role of unequal childcare responsibilities in understanding the gender pay gap. The wedge between women’s and men’s pay increases sharply after women have their first child. Women often choose jobs closer to home, limiting their options for well-paid jobs, and there are pay penalties for part-time work and time spent caring.

However, even before they have had their first child, women are paid less than similarly-qualified men. At age 25, the average male graduate earns 5% more per year than the average female graduate. This is despite the fact that women are more likely to get a first class degree or a 2:1 than men. By age 30 – before most graduates start having children – the gender pay gap in annual earnings stands at 25%.

What explains the gender pay gap in graduates’ early careers? A plausible explanation is that men and women choose to study very different subjects at university, with men being more likely to study subjects which lead to highly paid jobs.

The financial return to getting a degree – how much more a graduate earns compared to an otherwise similar non-graduate – varies enormously across subjects. Previous IFS research estimates that studying economics at university boosts women’s pay by 75% by age 30; this is more than ten times the return to studying creative arts (7.2%). However, women make up nearly two-thirds of creative arts graduates but less than a third of economics graduates.

In general, women are overrepresented in degree subjects with low financial returns (Figure 1). There are some exceptions – for example, medicine and law both have average or slightly above average shares of female students and very high returns.

Figure 1. Estimated returns to degrees and the share of female students, by subject

Source: J. Britton, L. Dearden, L. van der Erve and B. Waltmann, ‘The impact of undergraduate degrees on lifetime earnings’, Institute for Fiscal Studies, 2020, https://www.ifs.org.uk/publications/14729

Notes: Female shares were obtained by mapping gender shares in principal subjects for full-time students studying for standard undergraduate degrees from published HESA data to CAH2 subject groups. Returns were estimated using the Department for Education’s LEO dataset, which combines school, university and tax records for all students who have sat GCSEs since 2002. The estimation sample consists of students from the 2002 GCSE cohort who achieved at at least 5 A*-C at GCSE and have a Key Stage 5 record. Age 30 earnings data was taken from 2016/17 tax records. Returns were estimated using a regression model that controls for attainment at Key Stage 2, 4 and 5; Key Stage 5 subject; and other background characteristics such as socio-economic group, region, and ethnicity. The ranking of degrees is very similar using estimated returns for men. Size of dots reflects number of graduates by subject.

Differences in degree subject choices explain most of the gender pay gap soon after graduation (Figure 2). Of the 5% gap in annual earnings at age 25, 2.9 percentage points (55%) can be accounted for by university subjects, with A-Level subject choices making up a further 0.2 percentage points (5%). Subject choice continues to contribute between 3 and 5 percentage points to the gender pay gap over graduates’ early careers. But over this period, other factors lead to a widening of the gender pay gap, so that by age 30, subject choice explains only a fifth of the total gender pay gap. Other factors that come into play could include motherhood, gender differences in attitudes towards risk, recognition for group work, hours worked, the propensity to bargain over wages and ask for promotions, and discrimination.

Figure 2. Decomposition of the graduate gender earnings gap

Source: J. Britton, L. Dearden and B. Waltmann, ‘The returns to undergraduate degrees by socio-economic group and ethnicity‘, Institute for Fiscal Studies, 2021, https://www.ifs.org.uk/publications/15383

Notes: Percentage gap in mean earnings between men and women who attended university from the Department for Education’s LEO dataset, which combines school, university and tax records for all students who have sat GCSEs since 2002. The sample includes students from the 2002-2004 GCSE cohorts. Earnings data at ages 25-30 was taken from 2013/14 to 2016/17 tax records. Mature and part-time students were dropped from the sample, as were people with earnings of less than £1,000 in the tax year. The earnings gap was decomposed using a two-fold Kitagawa-Oaxaca-Blinder decomposition (Kitagawa, 1955; Oaxaca, 1973; Blinder, 1973). Following Neumark (1988), the non-discriminatory reference coefficient vector was estimated using a pooled regression over both men and women (including an indicator variable for gender). Results were broadly robust to using the coefficients for either men or women as reference coefficients instead.

Money isn’t – and shouldn’t be – the only factor used to decide degree subjects. But we should be concerned if information on the returns to different subjects isn’t easily available to young people, and if the large differences in subject choice (arts for girls, economics for boys) are driven as much by gender stereotypes as by true preferences. When it comes to a subject like economics, which delivers the very highest financial return for female (and male) graduates, there is an additional concern that many students cannot access the subject at all because it is not offered in their school. 

When the company data are published and there is debate on what firms should do, it is important to remember that the seeds of the gender pay gap are sown early. More needs to be done to educate and inform young people about subject choices at A level and university, particularly in a system like the UK where subject choices narrow at an early stage and where decisions taken early can have long-lasting effects.