The gender pay gap is a myth

The claim

The 82-cents figure compares all full-time male and female workers without controlling for occupation, experience, or hours worked. When you control for these factors, the gap shrinks to near-zero. The remaining gap reflects choices about risk, work intensity, and career interruption — not discrimination. Women who make the same choices as men earn the same wages.

The mechanism

The “controlled gap” argument contains real methodology. When you hold occupation, industry, experience, and hours constant, the raw gap does shrink substantially — from 16 cents to approximately 2–5 cents in most analyses. But this narrowing creates a false dichotomy by treating the control variables as exogenous personal choices, rather than themselves being shaped by structural forces.

Occupation devaluation — the Paula England mechanism: Paula England and colleagues have documented a persistent pattern: when women enter an occupation in large numbers, wages in that occupation decline — even controlling for workers’ characteristics. Levanon, England, and Allison (2009, Social Forces, 88(2), 865–891) analyzed occupational wage changes using 1950–2000 US Census data and found substantial evidence for the devaluation thesis: occupations with a growing female share paid less, holding skill and education constant. The reverse also holds: when men enter previously female occupinations, wages rise. The classic example: computer programming was a female-dominated occupation in the 1950s and 1960s (Margaret Hamilton, ENIAC programmers); as it became male-dominated in the 1980s and 1990s, wages rose dramatically. The work did not change; the gender composition did. This is not personal choice — it is the market’s valuation of work based on who performs it.

The motherhood penalty and fatherhood premium: Shelley Correll, Stephen Benard, and In Paik’s 2007 American Journal of Sociology paper used an audit study with identical resumes: mothers were rated as less competent and committed than non-mothers (and than fathers), offered lower starting salaries, and less likely to be promoted. These differences appeared when hiring managers evaluated identical credentials. Michelle Budig and Paula England’s 2001 American Sociological Review paper analyzed NLSY data and found that mothers earn approximately 4–7% less per child than comparable childless women, while fathers earn a wage premium over childless men. The mechanisms include both discrimination (employers anticipating availability reductions) and actual career interruptions — but Correll et al.’s audit study captures discrimination directly, since the resumes are identical.

The “greedy jobs” mechanism (Claudia Goldin, Nobel 2023): Goldin’s career work on gender earnings convergence and remaining divergence culminated in her 2014 AER paper and her 2021 book Career and Family. Her central finding: the gender pay gap has narrowed dramatically as women’s education and labor force participation converged with men’s, but the remaining gap is concentrated in specific sectors and moments. The primary driver of the remaining gap is the “greedy job” premium — employers pay disproportionately high wages not just for more hours worked but for being available at unpredictable times (weekend calls, emergency travel, client coverage at 10pm). These premiums are largest in corporate law, finance, and consulting. The penalty for reduced availability falls more heavily on primary caregivers — who are disproportionately women, not because women are inherently less ambitious, but because the division of domestic labor (itself structured by wage gaps, cultural expectations, and the absence of paid leave and subsidized childcare) assigns primary caregiving to women. The policy implication: the greedy job premium would be reduced by workplace flexibility, paid leave for both parents, and subsidized childcare — structural interventions that change the underlying division of labor.

The residual gap after controls: Even after the most exhaustive controls, a 2–5 cent per dollar gap remains. A 2023 Glassdoor Economic Research analysis of 500,000 salary observations found a 5.4% controlled gap — “women earn 94.6 cents for every dollar men earn doing the same job at the same company.” This residual is statistically significant and consistent across multiple methodologies. Some portion likely reflects discrimination in the traditional sense — paying women less for the same work. Some may reflect measurement gaps (hours not fully captured; the exact same “job” having different duties for men and women at the same title). The important point is that the controlled gap is not zero.

Who benefits

Employers in sectors with large female workforces — retail, food service, healthcare support, education — have financial interests in maintaining wage norms that discount female labor. These interests appear in lobbying against the Paycheck Fairness Act (which would strengthen EPA enforcement and require salary transparency), against mandatory pay audits, and against policies that would equalize the caregiving burden (universal paid leave, subsidized childcare) — because those policies would reduce the greedy-job premium and narrow the gap.

The data

BLS Highlights of Women’s Earnings is published annually. The 2022 report (based on Current Population Survey data, published 2023) found: women working full-time, year-round had median weekly earnings of $978 vs. $1,143 for men — an 84-cent ratio. This figure has improved from approximately 57 cents in 1970 and 77 cents in 1990, reflecting real convergence in education, experience, and occupational distribution. The remaining gap is the contested terrain.

OECD.Stat provides comparable gender wage gap data across member countries. The 2023 figures (unadjusted, median earnings):

Sources: OECD Gender Data Portal 2023; OECD Family Database.

The correlation between national gender pay gaps and the availability of paid parental leave and subsidized childcare is strong and consistent across OECD countries — providing structural evidence that these policies reduce the gap by changing the underlying distribution of caregiving responsibilities.

Iceland’s mandatory pay audit law (2018): Iceland implemented Jafnlaunalög (Equal Pay Standard) requiring companies with 25+ employees to obtain certification that their pay practices are gender-equitable, or face fines. Since 2018, Iceland’s gender pay gap has narrowed from approximately 14% to 7.7% — the smallest in the OECD. The certification requirement created accountability for the residual gap that voluntary measures had not.

Comparators

Iceland’s pay audit law provides a natural experiment. Relative to other Nordic countries without mandatory pay audits, Iceland’s gap narrowed faster after 2018. This suggests that some portion of the “controlled gap” reflects discrimination that is identifiable and correctable when employers face regulatory scrutiny — not immutable personal choices.

Sweden’s generous dual-earner/dual-carer model (both parents encouraged to take paid parental leave through “use it or lose it” paternity months) produces a smaller gender pay gap than countries with primarily maternal leave, by equalizing the caregiving signal employers receive when hiring young women vs. young men.

The counter

The libertarian response: even if the motherhood penalty is real, it reflects employer rational statistical discrimination — women of childbearing age are, on average, more likely than men to take career interruptions, and employers price this risk. Requiring equal pay regardless of statistical availability differences would reduce employers’ willingness to hire women of childbearing age (a perverse outcome). This is a real theoretical argument. The empirical evidence weighs against it: countries with mandatory paid leave (including paternity leave) and subsidized childcare show smaller pay gaps AND higher female labor force participation — not less hiring of women. Employers who face a level playing field (where the caregiving interruption applies equally to men and women) cannot statistically discriminate on gender alone. The structural policy changes the underlying signal, rather than suppressing employers’ response to it.

References

Budig, M. J., & England, P. (2001). The wage penalty for motherhood. American Sociological Review, 66(2), 204–225. https://doi.org/10.2307/2657415

Bureau of Labor Statistics. (2023). Women’s earnings and employment by industry and occupation, fourth quarter 2022. U.S. Department of Labor. https://www.bls.gov/opub/ted/2023/womens-earnings-and-employment-by-industry-and-occupation.htm

Correll, S. J., Benard, S., & Paik, I. (2007). Getting a job: Is there a motherhood penalty? American Journal of Sociology, 112(5), 1297–1338. https://doi.org/10.1086/511799

Goldin, C. (2014). A grand gender convergence: Its last chapter. American Economic Review, 104(4), 1091–1119. https://doi.org/10.1257/aer.104.4.1091

Goldin, C. (2021). Career and family: Women’s century-long journey toward equity. Princeton University Press.

Levanon, A., England, P., & Allison, P. (2009). Occupational feminization and pay: Assessing causal dynamics using 1950–2000 U.S. Census data. Social Forces, 88(2), 865–891. https://doi.org/10.1353/sof.0.0264