Question

It appears that, in a particular dataset, women’s average wage is closer to the average wage of men at higher levels of education. So the gap between the average wage of men and the average age of women decreases as education increases. But the average wages don’t converge because the return to education (that is, the partial effect of education on wages) for men is constant but there is a diminishing marginal return to education (that is, the partial effect of education on wages gets smaller as education increases) for women. How would you test this hypothesis using a single regression? (Hint: Start by drawing a graph to illustrate the hypothesized relationship between wages and education for men and women.)

Answer 1

The gender pay gap in the United States is the ratio of female-to-male median or average (depending on the source) yearly earnings among full-time, year-round workers.

The average woman's unadjusted annual salary has been cited as 78% to 82% of that of the average man's. However, after adjusting for choices made by male and female workers in college major, occupation, working hours, and parental leave, multiple studies find that pay rates between males and females varied by 5–6.6% or, females earning 94 cents to every dollar earned by their male counterparts. The remaining 6% of the gap has been speculated to originate from gender discrimination and a difference in ability and/or willingness to negotiate salaries.

The extent to which discrimination plays a role in explaining gender wage disparities is somewhat difficult to quantify, due to a number of potentially confounding variables. A 2010 research review by the majority staff of the United States Congress Joint Economic Committee reported that studies have consistently found unexplained pay differences even after controlling for measurable factors that are assumed to influence earnings – suggestive of unknown/unmeasurable contributing factors of which gender discrimination may be one. Other studies have found direct evidence of discrimination – for example, more jobs went to women when the applicant's sex was unknown during the hiring process.