The Differential Effects of the Determinants of Household Education Expenditure in India : Quantile Regression Estimation

Abstract

Despite substantial government expenditure on education, scholarships, and financial aid to students to provide affordable education, the household education expenditure on children is sizable and varies widely on account of gross differences in the socioeconomic, demographic, religious, and cultural factors. This paper attempted to identify such determinants and analyze the differential effects of the determinants of household education expenditure on children in India using the 2014 NSSO 71st Round survey data by applying the quantile regression method. Unlike the standard regression method, the quantile regression method allows estimation beyond the average effects at different points of the distribution of household expenditure on education. The quantile regression estimates revealed that low-income households were more sensitive to changes in household income and government programmes than upper-income households. The proportion of household income spent on the education of children increased more in the lower quantiles than in the higher quantiles. Gender bias existed at the lower quantiles and was considerably less at the higher quantiles. The SC/ST households spent less than the non-SC/ST communities at the lower quantiles and the difference got reduced at higher quantiles. Compared to scholarships, the provision of educational materials had a higher impact on household education expenditure. More children from lower quantiles attended government institutions, and a substantial difference existed in household education expenditure between the students attending government and private educational institutions. Despite government policies and programmes for affordable education, the study observed that the lower-income households still incurred a considerable proportion of their income on the education of their children.