Returns to Schooling in Bangladesh Revisited: An Instrumental Variable Quantile Regression Approach

Abstract

The paper focuses on estimation of returns to schooling in the Bangladesh context. Earlier studies which tried to quantify the returns were constrained by a number of factors including the limitations of the measurement techniques that were deployed. This paper revisits the issue and makes an attempt to build on earlier scholarly works through application of quantile regression and instrumental variable quantile regression methods. The paper finds that endogeneity problem leads to underestimation of the returns to schooling, and that the returns tend to vary along the wage distribution, which mean regression models fail to capture. The analysis shows that average returns to schooling for female is higher than that of male. The analysis also shows that returns to schooling tends to be higher as one moves along higher percentiles of wage distribution. This is found to be true both for male and female, as also for rural and urban labour markets.