Gender Gap Decomposition
Also known as: Oaxaca-Blinder Decomposition, Blinder-Oaxaca Decomposition, Wage Gap Decomposition
Gender gap decomposition, most often implemented as the Oaxaca-Blinder decomposition, splits the mean difference in an outcome such as wages between men and women into a part explained by differences in measured characteristics (education, experience, occupation) and an unexplained residual part attributed to differences in how those characteristics are rewarded. Introduced independently by Ronald Oaxaca and Alan Blinder in 1973, it is the workhorse method for quantifying how much of the gender pay gap reflects composition versus differential treatment.
Key highlights
- Produces an intuitive, policy-relevant split of a raw gap into an explained characteristics component and an unexplained returns component.
- Builds on ordinary regression, so it is straightforward to estimate and widely supported in standard statistical software.
- Extends naturally to detailed decompositions that attribute portions of the gap to individual variables.
- Has decades of methodological development addressing reference choice, selection, and extension beyond the mean.
Intuition
This section is available to Pro members. Upgrade to Pro
How it works
This section is available to Pro members. Upgrade to Pro
When to use it
Use gender gap decomposition when you have individual-level data on an outcome (most often wages) plus characteristics for two groups and you want to quantify how much of the mean gap is due to differing characteristics versus differing returns. It is standard in labor economics for pay-equity analysis, but applies to any group comparison — by ethnicity, region, or sector. It is less suitable when the outcome is highly nonlinear, when selection into employment is severe and uncorrected, or when the interest is in gaps across the whole distribution rather than at the mean, where quantile-based extensions are preferred.
Strengths & limitations
- Produces an intuitive, policy-relevant split of a raw gap into an explained characteristics component and an unexplained returns component.
- Builds on ordinary regression, so it is straightforward to estimate and widely supported in standard statistical software.
- Extends naturally to detailed decompositions that attribute portions of the gap to individual variables.
- Has decades of methodological development addressing reference choice, selection, and extension beyond the mean.
- The unexplained component is not a clean measure of discrimination: it absorbs all omitted, unobserved, and mismeasured variables.
- Results depend on the choice of reference coefficients (the index-number problem), so the explained/unexplained split is not unique.
- The classic linear form decomposes only the mean and assumes a correctly specified linear wage equation.
- Detailed decompositions of categorical variables can be sensitive to the choice of omitted base category unless normalized.
Common pitfalls
This section is available to Pro members. Upgrade to Pro
Applications
This section is available to Pro members. Upgrade to Pro
Frequently asked
Does the unexplained gap measure discrimination?
Not directly. The unexplained component captures differences in the returns to characteristics, which can reflect labor-market discrimination but also unobserved or mismeasured productivity, omitted variables, and specification error. It is best described as an upper-bound proxy that warrants careful, caveated interpretation rather than a clean estimate of discrimination.
What is the index-number problem in this decomposition?
The decomposition requires a reference coefficient vector representing the 'nondiscriminatory' wage structure, but there is no unique correct choice — men's coefficients, women's coefficients, or a pooled combination all give different splits between explained and unexplained. This sensitivity is the index-number problem, and good practice is to report the reference used and ideally check robustness across alternatives.
How do I decompose the gap beyond the mean?
The classic Oaxaca-Blinder decomposes only mean wages. To examine the gap across the distribution — for instance, glass-ceiling or sticky-floor effects — researchers use the Firpo-Fortin-Lemieux recentered influence function (RIF) regression approach or the Juhn-Murphy-Pierce method, which extend the explained/unexplained logic to quantiles and other distributional statistics.
Sources
- 1.Oaxaca, R. (1973). Male-female wage differentials in urban labor markets. International Economic Review, 14(3), 693–709.
- 2.Blinder, A. S. (1973). Wage discrimination: Reduced form and structural estimates. Journal of Human Resources, 8(4), 436–455.
- 3.Jann, B. (2008). The Blinder-Oaxaca decomposition for linear regression models. The Stata Journal, 8(4), 453–479.
You have read it. What now?
Cite this page
ScholarGate. (2026, June 22). Gender Gap Decomposition. ScholarGate. https://scholargate.app/gender-studies/gender-gap-decomposition