Written by: Alicja Sielska

The problem of gender inequality in the labor market is still important and debated not only in the scientific world but in politics and the media. Proof can be found in the awarding of this year’s Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel to Claudia Goldin “for having advanced our understanding of women’s labour market outcomes.”

The wage gap has for many years been a symbol of women’s inferior position in the labor market and of gender discrimination. In 2014, then president Barack Obama (2014) stated in a speech, “Now, here’s the challenge: Today, the average full-time working woman earns just 77 cents for every dollar a man earns; for African American women, Latinas, it’s even less. And in 2014, that’s an embarrassment. It is wrong.” The issue of wage equality is highlighted by many international institutions. In the World Bank’s (2018) view, “The world is essentially leaving $160 trillion on the table when we neglect inequality in earnings over the lifetime between men and women.”

But is inequality in pay actually the result of discrimination? Too often the wage gap is presented as the difference between the average wages of men and women, or the unadjusted gap. Such a presentation is misleading because it does not compare identical workers. By analogy, we pay more per kilogram of organic apples than chemically fertilized ones, which are considered less healthy. The production process is different, and, as a result, so is the product. Similarly, we should not be surprised that the work of a male worker who graduated from high school, compared to a woman with a university degree, is subject to a different and, on average, lower valuation.

Various statistical methods, including the Oaxaca-Blinder decomposition (Oaxaca 1973, Blinder 1973), attempt to compare the wages of similar employees. Thanks to this method we can identify what part of the unadjusted (or raw) wage gap arises because female and male workers differ in age, place of residence, union membership, veteran status, marital status, education, length of working life, profession, and job position held, among other characteristics—the explained component. After subtracting the explained component from the raw gap, we are left with the unexplained component, which is often taken to indicate the scale of discrimination against women in the labor market. But is that correct?

Consider Blinder’s pioneering paper studying, among other things, the wage gap between white men and white women using data from the Michigan Survey Research Center in 1967. He calculated that white men earned on average 45.8 percent more than white women, 15.7 percentage points of which was the explained component, consisting of the above factors. In other words, if women and men were identical, the gender wage gap would be 15.7 percentage points lower. Among the observed variables, the most important were length of working life (women’s wages did not increase over their working lives), education (men had more), and conditions in the local labor market conditions (men were less sensitive) (Blinder 1973). However, around two-thirds of the wage gap remained unexplained.

Does that unexplained gap indicate the scale of wage discrimination against women?

No.

First, the size of the explained component depends on what factors the researcher considers, and this affects the size of the unexplained component.

Second, discrimination can be hidden in the explained component. For example, if we include vertical occupational segregation (the glass ceiling) among the factors that differentiate pay, then we can say that women on average earn less because they are more likely to hold lower positions, but that may be precisely because of discrimination against women by employers who do not want to promote them.

Third, there may be hidden subcomponents in the unexplained component that also determine salary but are most often not considered because they are difficult to measure. Examples include women’s greater aversion to competition or negotiation, job availability, and stereotype threat.

In the 1970s, Oaxaca (1973) aptly noted, “If it were possible to control for virtually all sources of variation in wages, one could pretty well eliminate labour market discrimination as a significant factor in determining wage differentials by sex (or race)” (p. 699). We, however, still make the mistake of treating the gender wage gap as a definite sign of large-scale discrimination against women in the labor market. Such discrimination may exist, but its scale is not accurately conveyed to us by the wage gap even in its adjusted form.

• Blinder A. S. (1973). Wage Discrimination: Reduced Form and Structural Estimates. Journal of Human Resources, 8(4), pp. 436–55. 
• Oaxaca R. (1973). Male-Female Wage Differentials in Urban Labour Markets. International Economic Review, 14(3), pp. 693–709. 
• Obama B. (2014). Remarks by the President on Equal Pay for Equal Work. The White House, https://obamawhitehouse.archives.gov/the-press-office/2014/04/08/remarks-president-equal-pay-equal-work [Accessed: 03.11.2023]. 
• World Bank. (2020). Girls’ education, https://www.worldbank.org/en/topic/girlseducation [Accessed: 03.11.2023].

Explaining the Gender Wage Gap
By Alicja Sielska is available now. A sample chapter is available on Elgaronline.