Abstract
This paper uses a new data‐set for cumulative national investment in formal schooling and a new instrument for schooling to estimate the national return on investment in 61 countries. These estimates are combined with data on the private rate of return on investment in schooling to estimate the external rate of return. In 1990 the external rate of return ranged from 10% in high‐income countries to over 50% in the lowest‐income countries. The external benefits of schooling are about equal to the private benefits in high‐income countries and three times the private benefits in the lowest‐income countries.
Acknowledgements
Thanks to Robert Lucas, Asa Janney, George Psacharopoulos, Steve Heyneman, and numerous referees who provided comments and assistance that contributed to the development of this paper.
Notes
1. Hanushek and Kimko (Citation2000) suggest that test scores on standard tests may be a more accurate indicator of relative levels of education than expenditures, but test scores are not available for low‐income countries and may not be representative of all students in a country.
2. The data shown in Figure exaggerate the upward trend in investment per year of schooling at higher levels of schooling attainment because countries with higher levels of schooling typically have older populations that include retired members with low levels of schooling. These data are not used in the empirical analysis.
3. Although this paper describes the relationship between Protestant affiliation and schooling only to justify the use of the Protestant variable as an instrument for schooling, the strength of this relationship is an indicator of the importance of cultural factors in determining the degree to which nations choose to invest in schooling.
4. The basic Solow growth model includes only income (Y), labor (L), physical capital (K), and a residual trend to account for unspecified factors: (Y/L)it = (K/L)it α [A 0 e gt ].
5. One potentially confusing finding in the micro studies is that the return on incremental years of schooling (the Mincerian return) is not subject to decreasing returns. The data in Tables and show why this is the case. Since the annual investment per pupil increases with years of schooling, the rising investment per year at higher levels of schooling offsets the diminishing returns per unit of investment.
6. Estimates of the income models using 10 years instead of 15 years of expenditures for the physical capital variable provided similar empirical results.
7. Psacharpoulos and Patrinos denote the return to the worker from the worker’s own investment – the ‘private return’ – and the return to the worker from the nation’s total investment in their education – the ‘social return.’ The rates of return designated the private rate of return in this article are their social returns.