Unexplained Gaps and Oaxaca-Blinder Decompositions

By Todd E. Elder, J. Goddeeris, Steven J. Haider. Social Science Research Network, 2009

204 citations recorded by the source corpus.

We analyze four methods to measure unexplained gaps in mean outcomes: three decompositions based on the seminal work of Oaxaca (1973) and Blinder (1973) and an approach involving a seemingly naive regression that includes a group indicator variable. Our analysis yields two principal findings. We show that the coefficient on a group indicator variable from an OLS regression is an attractive approach for obtaining a single measure of the unexplained gap. We also show that a commonly-used pooling decomposition systematically overstates the contribution of observable characteristics to mean outcome differences when compared to OLS regression, therefore understating unexplained differences. We then provide three empirical examples that explore the practical importance of our analytic results.

Subfields: Experimental and Cognitive Psychology

View Paper: https://www.semanticscholar.org/paper/7ce533f322cd25e230fb2c673fa6a49fcdc2139a

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Experimental and Cognitive Psychology

Citations in the corpus

Interpreting OLS Estimands When Treatment Effects Are Heterogeneous: Smaller Groups Get Larger Weights (2018, methodology): “The main proof of theorem 1 uses a result on decomposition methods from Elder, Goddeeris, and Haider (2010).”
Earnings differentials between the public and private sectors in China: Exploring changes for urban local residents in the 2000s (2012, background): “However, this assumption leads to the well-known “index number problem,” resulting in various weighting schemes being put forward in the empirical literature (Neumark, 2008; Jann, 2008; Elder et al., 2010).”
Noncognitive Skills, Occupational Attainment, and Relative Wages (2009, background): “This decomposition is not unique and the choice of the counterfactual, nondiscriminatory wage structure used in the decomposition inherently depends on assumptions about the nature of discrimination present in the labour market (Elder et al., 2009; Neumark, 1988; Oaxaca and Ransom, 1994).”
When Do Covariates Matter? And Which Ones, and How Much? (2016, background): “, Neumark (1988), Fortin (2008), and Elder et al. (2010). Various alternative specifications can be obtained by appropriately constraining the population values corresponding to the α̂ coefficients.”