Level of Development and Rates of Return to Schooling: Some Estimates from Multicountry Data

Abstract

Estimation of rates of return to education has been an important part of the human capital and the development literature, and many scholars have provided rate of return estimates for a wide variety of countries, economic settings, labor market sectors, and levels and types of schooling. George Psacharopoulos, for exampale, has produced comprehensive rate of return estimates and discussed the relationships between the rate of return estimates and the level of economic development.' Although many different kinds of data sets and procedures have been used to generate rate of return estimates, no attempt seems to have been made to obtain a global view of returns to education by using multicountry data. While intracountry data have some obvious advantages over intercountry data, the latter are better in some other respects. For example, if one considers the entire labor force, crosscountry differences in age, experience, and ability may not be a matter of major concern, and relatively simple specifications can be used. Moreover, multicountry data permit a direct comparison of rates of return to education in groups of countries that are at different levels of economic development.2 The main purpose of this study is to use an earnings-function approach to estimate rates of return to schooling from multicountry data on labor force education and real income. The work has been made possible by Psacharopoulos and Ana Maria Arriagada's compilation of data on labor force schooling for a large number of countries and the recent update by Robert Summers and Alan Heston of data on real GDP that have good cross-country comparability.3 Besides obtaining an estimate of the overall (global) rate of return to schooling, this article (a) compares estimates from the sample of