Decomposing the language pay gap among the indigenous ethnic minorities of Mexico: is it all down to observables?

Using the decomposition methods of Oaxaca and Choe (2016), we investigate the pay gap between indigenous language monolinguals (INL) and Spanish-indigenous-language bilinguals (BIL) among indigenous ethnic minorities in Mexico using the 10% sample of the Mexican Census 2000 and 2010. The decomposition fits linear models with municipal fixed effects for the case of males and correlated random effects Heckman sample selection models for the case of females (to account for potential sample selection bias). We find evidence of a positive return to bilingualism for males of 17% and of 42% for females. Over 60% of the pay gap is explained by differences in observable characteristics.