Abstract
Using Portuguese data from the 2001 wave of the European Community Household Panel, we analyse to what extent the endogeneity of schooling affects the estimation of the total impact of schooling on within-groups wage inequality by means of quantile-regression techniques. We conclude that the standard techniques assuming schooling-exogeneity may underestimate the total impact of schooling.
Acknowledgements
Financial support by the European Commission (EDWIN Project, HPSE-CT-2002-00108) is gratefully acknowledged. For their useful comments, I am indebted to Santiago Budría, Günther Lang, Joop Hartog, Pedro Telhado Pereira, Rudolf Winter-Ebmer and Robert Wright. The usual disclaimer applies.
Notes
1 Jobs in some industries may require more years of schooling than jobs in other industries.
2 Pereira and Martins (Citation2004) properly argue that in order ‘to obtain the full impact of education on wages, one should be careful not to include in the wage equation covariates whose value can depend on education. In the extreme case one should only regress the ln(wage) in education.’ (p. 526). See also Andini (Citation2007).
3 This explains why we consider declared schooling-years instead of education levels. Specifically, the use of education levels would have implied the estimation of a first-stage linear-probability model and therefore the implementation of maximum-likelihood techniques, which are not allowed by assumption (see model AHS). Alternatively, we might have transformed individual levels of education into equivalent (successfully completed) schooling years, but such procedure would have implied an arbitrary treatment of unobservable measurement errors.
4 These graphs are obtained using STATA and the GRQREG module due to Azevedo (Citation2004).