ÁñÁ«ÊÓƵ

Ever since the start of the "human capital revolution" in economics in the early 1960s, economists have been concerned with measuring the returns on education. However, the cost-benefit (calculated through increased productivity and earnings) and Jacob Mincer's adaptation of the cost-benefit method used to estimate the economic return on education since the 1960s have two problems, which Walter McMahon seeks to redress.

First, they estimate only an "economic" return. For those who believe, as McMahon does, that "true economic development does not occur unless pure economic growth is accompanied by other dimensions important to quality of life" - such as health, prosperity, clean air and water, and basic human rights - it is important to evaluate education's contribution to wider aspects of human welfare. Second, the existing methods fail to take into account the likely external effects of education. Endogenous growth theory emphasises that education raises not only a person's own productivity but also the productivity of others around them. The volume thus contributes in two key ways. By using cross-country growth regressions of aggregate country-level data, McMahon holds out the promise of capturing the various economic externalities of education so heavily emphasised in new growth theory. He also has a broader view of development, focusing on the total (economic and social) returns and not just the economic returns of education.

This book is an intelligent exposition on the relationship between education and development. Given that no structural approach has yet been employed that measures systematically the indirect benefits of education on growth, the suggestion of such a quantitative approach is a valuable contribution to economics of education. Education and Development represents a huge amount of computation work involving data from 78 countries and many regressions and policy simulations over a 40-year period.

Initially, the book concentrates on the impact of education on the national economic growth rates for countries in East Asia, Latin America and Africa, holding constant the usual determinants of growth such as initial per capita national income and investment in physical capital as a proportion of national income. It then examines the effect of education on a large number of social outcomes: infant mortality, life expectancy, fertility and population growth rates; human rights and political stability; urban and rural poverty and inequality; the environment, and different aspects of crime.

The last few chapters trace the direct and indirect effects of education on economic growth in the complete model. Estimates of education's total economic and social benefits are made with policy simulations that generate time paths using data specific to each country up to the year 2035. This is calculated first using past policies and then under the scenario that the country increases its investment in education as a percentage of gross domestic product by two percentage points. The difference between these time paths for each variable is taken as the measure of the total social benefits of education.

Any reader familiar with cross-country growth regressions will not be surprised by the lack of a sizeable and robust direct effect of primary, secondary and higher education on economic growth: the effect is sometimes negative, sometimes positive, sometimes statistically significant and sometimes not, and it depends on the measure of education used, the enrolment rate or the investment in education as a percentage of GDP. However, McMahon demonstrates what is less commonly known, that education has an impact on growth via its strong effect on investment in physical capital, which is a key determinant of growth.

While the book presents careful empirical analysis grounded in theory, there are some areas of concern. McMahon makes little attempt to reconcile his macro evidence on returns to education with extant micro evidence (except tentatively for the United States and the United Kingdom). He estimates the total social rate of return to education in the US to be 14 per cent and in the UK 15 per cent. As this is quite close to estimates of 11-13 per cent from conventional Mincerian approaches that take no account of the economic or social externalities of education, it suggests, anticlimactically, that the indirect effects of education are very small. It is also disappointing that McMahon does not produce estimates of total benefits of education for each of his sample countries.

Second, McMahon claims that the social rates of return to basic education are very high (24 per cent for primary and 18 per cent for secondary education in sub-Saharan Africa), quoting estimates of returns to education by George Psacharopoulos. These African estimates have been damningly questioned by Paul Bennell in World Development (1994) and new figures published in recent papers by others. These facts are not recognised or reconciled with the estimates in the book. Third, it seems wrong to present separate growth regressions for each region because they are all part of the same underlying model of growth. The number of observations in the regional samples is small: only 35 observations (seven countries over five time periods) in East Asia.

One final reservation - McMahon fails to mention the quality of the data used. Knowledge about the reliability of data can help assess the findings' integrity. Official enrolment rates are known to be greatly exaggerated at the primary level in India, a country included in the Asian growth regressions.

These drawbacks apart, the book is an approachable and interesting analysis of the determinants of development, with a range of pertinent cross-country regressions. It will be of interest to education, labour and growth economists as well as to policy-makers and development organisations.

Geeta Gandhi Kingdon is research officer in economics, University of Oxford.