Journal Issue

Education as an investment: World Bank education projects yield high economic returns

International Monetary Fund. External Relations Dept.
Published Date:
September 1982
  • ShareShare
Show Summary Details

George Psacharopoulos

Human capital is created and the quality of human input in production is significantly improved by spending on education. This is why countries, particularly those with low per capita incomes, invest such a large proportion of their budgets on education—and why, when the state does not, individuals do. It is highly rewarding, both to society and to the individual. It is difficult to measure all these rewards; many, such as the benefits to society of a better electorate or a more informed consumer body, though tangible, are quantitatively elusive. But the measurable evidence alone is sufficient to demonstrate sizable financial rates of return to investing in education.

The World Bank made its first educational loan in 1963, after it had been in operation for over 15 years. Since then, 224 such projects of increasing diversity have been financed throughout the world, mostly in poor countries. The Bank’s late entry into educational lending reflects the philosophy that prevailed after World War II until the late 1950s that massive financial investment in such physical assets as ports, roads, irrigation, dams, factories, and machines would generate higher output. But experience showed that some countries (like India) responded sluggishly to the aid inflow, whereas some others, particularly the war-devastated countries (like the Federal Republic of Germany and Japan) did very well without as much financial assistance. This apparent paradox pointed to the importance of human capital in economic development. To put it bluntly, no machine can be operated effectively or agricultural project executed properly unless it is appropriately staffed.

The methodology used for evaluating education projects is a much debated issue. For many years, the Bank has been using the “manpower requirements” approach, which identifies the number of skilled people a country needs to reach sectoral growth targets. Because manpower requirements are directly based on sectoral needs for qualified people, often in the modern sector of the economy, the approach tends to lead to investments in higher education and technical curricula. However, the Bank is also increasingly using information on the costs and benefits of different levels of education as a basis for decisions to invest. Typically this approach favors the lower levels of education and general curricula because of their low costs. Deriving costs is a relatively straightforward matter; arriving at private and social rates of return, particularly the latter, seems to pose more problems, both empirically and theoretically. This article will discuss the calculation of rates of return and will examine the results of using them to quantify the profitability of different levels of education in a broad sample of developing countries.

The findings have important policy implications. They show, first, that using this methodology, returns on investments in all levels of education in the countries reviewed are well above the 10 per cent normally used as a criterion for accepting an investment project. Second, returns are highest for investments in primary schools. Third, surprisingly, returns on education in general arts subjects are often higher than those for technical training, mainly because of the high cost of the latter. This evidence on quantifiable returns also captures some of the more elusive, un-measurable social benefits of education. Since most women are trained in general subjects, investing in general curricula also brings the wider social benefits of better nutrition, hygiene, and so on that are derived from women’s education, regardless of whether these women form part of the labor force or not. Recent studies indicate, too, that primary education has a significant impact on poverty and the distribution of incomes, in addition to being highly profitable.

Estimating rates of return

Estimates of the rate of return to a given level of education are calculated by comparing the discounted benefits over the lifetime of an educational investment “project” to the costs of such project. Thus, for the calculation of the private rate of return to four years of university education, benefits are estimated by taking the difference between existing statistics on the mean post-tax earnings of university graduates by age and those of a sample group of secondary school graduates. The earnings of the latter also represents the opportunity cost of staying in school. Direct costs are obtained from statistics on a student’s out-of-pocket expenditures that are strictly due to the costs of college attendance. Given these data, the rate of return to investment in a college degree compared with a secondary school qualification is the rate of interest that reduces to zero the net present value of the discounted difference between the costs and benefits. A simple equation for the private rate of return is

Note that this formula can be interpreted as the yield of a permanent constant stream of benefits (the difference in earnings appearing in the numerator) over a lump sum cost of projected earnings plus direct outlays (appearing in the denominator). Neither the permanent benefits assumption nor the lumping together of costs are critical in the calculation, since the latter occur within four years and the former extend over several decades.

A social rate of return to college education could be calculated in the same way, although earnings should be pretax (as taxes are a transfer from the point of view of society at large) and the direct cost should include the full amount of resources committed per student of higher education, rather than the usually smaller part of expenditure borne by the student.

Profitability of investment

The rate of return to investment in education can be estimated in the same way as returns to any other economic activity. A given cost is incurred for a number of years in the form of teachers’ salaries; the cost of using buildings and facilities, equipment, and materials; and forgoing production while the student is in school. Benefits come in the form of increased production over the lifetime of a person with more education and broader skills.

Because of the difficulties involved in directly measuring the marginal product of labor with different levels of education or the opportunity cost of someone who is staying in school rather than working, earnings are used as a proxy for productivity. Two types of profitability measures are usually estimated, private and social.

The private rate serves to explain and predict the private demand for places in certain types of schools. Social rates of return are used as guides for allocating public investments in education. (See the box for the quantitative expression of these costs and benefits.)

The private rate of return estimates the gains made by the educated individual; benefits are represented by the extra value education gives in the form of his or her lifetime earnings, after tax, and costs by the out-of-pocket expenses—tuition, fees, and other incidental expenses strictly related to the extra education received. For example, a fellowship would enter as a negative cost item in calculating the private rate of return to a college degree, and the direct private cost of primary education in a fully subsidized state school would be virtually nil in most countries. By contrast, a calculation of the social rate of return includes the full social cost of education, regardless of whether schools charge fees or not. The benefits (or earnings) are calculated before taxes, since taxes represent a transfer from the point of view of society.

These rates of return are the subject of much debate. Two of the main issues will be mentioned here; in fact, none of the caveats are sufficiently serious to invalidate the method of calculation.

The main criticism is that observed market wages frequently do not fully reflect the social benefits of education because in many countries wages are set by non-market forces such as social custom, unions, or employers (such as the government) who do not pay market rates. However, in most situations it is possible to correct for such distortions by basing calculations on shadow wages or earnings paid to people with the same education levels in the competitive sector of the economy. The social benefit of providing extra education to farmers is often not captured in their earnings, so the return is measured in terms of the extra rice or wheat grown rather than in dollars or pesos. The result of a recent Bank survey of farmers’ education and efficiency has shown that, on average and other things being equal, farm productivity increases by 7.4 per cent when a farmer has completed four years of elementary education. When this increased productivity is related to the cost of providing primary education, the resulting social rate of return to investment in primary schools is of the order of 30 per cent.

Another criticism of using wages to show the profitability of education is that the difference between the earnings of the educated and the uneducated may well be influenced by factors other than education, such as native intelligence or socioeconomic background. But when these factors are taken into account the net effect of education on earnings and productivity is only slightly reduced.

Some empirical evidence

Several estimates have been made over the last 20 years of educational costs and benefits, and, of course, the resulting rate of return. A recent Bank survey of 44 countries at different stages of economic development gave the social rates of return appearing in Table 1. (The private rates of return are several percentage points higher than the corresponding social rates, as will be discussed later.)

Table 1Social returns to education by level and country type(In per cent)
Country typePrimaryEducational level

Source: Based on George Psacharopoulos, “Returns to Education: An Updated International Comparison,” Comparative Education (October 1981).

Developing countries include Brazil, Chile, Colombia, Ethiopia, Ghana, India, Indonesia, Kenya, Malawi, Malaysia, Mexico, Morocco, Nigeria, the Philippines, Rhodesia, Sierra Leone, Singapore, South Korea, Taiwan, Thailand, Uganda, and Venezuela.

Intermediate countries are Cyprus, Greece, Israel, Iran, Puerto Rico, Spain, Turkey, and Yugoslavia.

Advanced countries are Australia, Belgium, Canada, Denmark, France, Germany, Italy, Japan, Netherlands, New Zealand, Norway, Sweden, the United Kingdom, and the United States.

Not computable due to lack of “control” group of unschooled.

Source: Based on George Psacharopoulos, “Returns to Education: An Updated International Comparison,” Comparative Education (October 1981).

Developing countries include Brazil, Chile, Colombia, Ethiopia, Ghana, India, Indonesia, Kenya, Malawi, Malaysia, Mexico, Morocco, Nigeria, the Philippines, Rhodesia, Sierra Leone, Singapore, South Korea, Taiwan, Thailand, Uganda, and Venezuela.

Intermediate countries are Cyprus, Greece, Israel, Iran, Puerto Rico, Spain, Turkey, and Yugoslavia.

Advanced countries are Australia, Belgium, Canada, Denmark, France, Germany, Italy, Japan, Netherlands, New Zealand, Norway, Sweden, the United Kingdom, and the United States.

Not computable due to lack of “control” group of unschooled.

Some salient features emerge from the table. First, the rate of return to investment at any level of education in developing countries is generally above the common 10 per cent normally required by the Bank to justify investment in other economic enterprises. The benefits accrue both to the individual, who is being educated and earns wages in accordance with his or her level of education, and to society, to the extent that the individual’s contribution to national income is more than the public resources committed to his or her schooling.

Second, the returns to investment in schooling decline as the educational level increases. This pattern suggests that top priority should be given to investing in primary schools. (It is also in accordance with the fundamental economic proposition of diminishing returns to investments at the margin, and some regard this as an ex post confirmation of the theoretical validity of the underlying estimation procedure.)

Third, the same declining pattern is observed across countries at different stages of development: the absolute returns to investment in schools being highest at all levels in poorer countries. This empirical finding has important implications for lending for education projects. It is the poorest countries that need most help in financing their educational systems, because money spent in this way exhibits the highest economic payoff.

The averages in the table hide considerable variations between and even within individual countries. Thus an early World Bank study has shown that the social returns to education in Kenya were 22 per cent for primary and 19 per cent for secondary. A later study in the same country showed that private returns for the same levels were of the order of 28 and 33 per cent. A study on Malaysia showed private returns to higher education (35 per cent) that were even higher than the returns to secondary education (33 per cent)—a deviation from the observed world pattern of declining profitability by level of education. Although social returns to primary education of the order of 82 per cent (Venezuela), 66 per cent (Uganda), and 51 per cent (Morocco) are not uncommon, the social returns to higher education in some countries are rather modest—thus, the Philippines has returns of only 9 per cent, Israel 7 per cent, Japan 6 per cent, Greece 5 per cent, and South Korea 5 per cent. It should be noted, however, that these relatively low returns are observed in countries with more or less highly developed systems of university education. Although one should not jump too easily from association to causation, it is tempting to attribute the high rates of economic growth observed in some of the countries listed above to previous investment in higher education.

The high social rates of return on education are echoed by the private rates. In fact these are higher, because the private rate balances the benefits of after tax earnings against the cost of schooling to the individual, which tends to be below the full cost to the public sector in most countries because of public subsidization of education.

High private returns exist despite the often substantial costs in terms of productive opportunities forgone while the individual is at school. To take the most extreme case, if the state covers 100 per cent of schooling costs, the additional earnings accruing to the educated individual would be obtained at zero personal cost, hence, the educational investment would exhibit an infinite private rate of return. However, the individual typically pays an opportunity cost, although it varies with the level of education invested in, as well as the country’s level of economic development. For example, the opportunity cost of someone studying for a doctorate in an advanced country is enormous because the individual forgoes what he or she could earn with a master’s or bachelor’s degree. On the other hand, there is virtually no opportunity cost associated with primary schools in advanced countries, because there are no employment opportunities for children and because the legal minimum working age is 14. However, the opportunity cost of primary education in poor countries is important because the productive contribution of children to the economy is both legal and substantial, particularly in activities such as agricultural production.

Nevertheless, students and their families continue to show a strong demand for places at all levels of schooling. Moreover, in many countries this demand is unsatisfied. The percentage of applicants obtaining university places is 34 per cent in India, for example; 33 per cent in Brazil; 33 per cent in Viet Nam; 25 per cent in Greece; and 15 per cent in Iran. Such low entry rates are not encountered in countries like the Philippines where most education is privately provided and private demand is almost completely satisfied. The unsatisfied private demand for education and the willingness of families to pay for the education of their children is prima facie evidence that there exists room for investment in education in developing countries.

Tapping private resources has been the typical solution in countries where the state budget is unable to provide as many school places or the kind of curricula the families want. Private schools, from the nursery to the doctorate level, have flourished in all kinds of countries to meet demand. Even where private universities are constitutionally prohibited, as in Greece, the safety valve has been to study abroad—it is estimated, for instance, that one out of every four Greek university students is enrolled at a foreign university at a substantial cost to the country’s balance of payments.

Benefits to society

At one time it was thought that the right kind of education to promote economic development should be of the technical-vocational type. However, recent evidence on the returns to this type of education has cast serious doubts on its value. More often than not, the rate of return to a general curriculum is higher than the return to a technical-vocational curriculum at the same level of education (see Table 2). The reason is that although engineers and technicians earn more, on average, than economics or humanities graduates, the cost of technical education is very much higher than the cost of general education—a fact that depresses the social returns to investment in technical education. Of course, this finding does not imply that engineering education should be curtailed. Simply put, it gives a signal that additional investment should go to arts subjects.

Table 2Social returns to university education by discipline in selected countries
CountryDisciplineRate of return

(In per cent)
Source: Based on George Psacharopoulos, “The Economics of Higher Education in Developing Countries,” Comparative Education Review (June 1982).

This finding also has other implications. The humanities and liberal arts have high female enrollments and are a socially profitable investment for two reasons: (1) the contribution these enrollments make to increasing female participation in the labor force and (2) the well-documented link between the education of women and social welfare factors such as better nutrition and lower fertility. Actually, the higher the level of educational attainment, the greater the participation of women in the labor force. Hence, more education, of any type, enhances the chance that it will be productively used in the market place. For example, in the Sudan the probability that a female will participate in the labor market increases from 8 to 38 per cent if she has completed secondary education. Similarly, in Singapore the chance of a woman participating in the labor force is 21 per cent if she has no schooling, 29 per cent if she has primary schooling, 68 per cent if she has secondary education, and 76 per cent if she has attended university.

But even if educated women never participate in the organized sector of the market (which is the most common case in developing countries), they certainly are economically engaged in agriculture and so-called “household production.” What this means is that more educated women produce more and better quality goods and services for the consumption of the members of their household. Such “goods” might take the form of more nutritious meals and a balanced diet, better sanitation conditions, and use of contraceptives. The link between education and fertility has been well documented in the literature of such issues. Other things being equal, more educated women desire and eventually have fewer children.

Investment in education, especially at the lower level, also has tangible effects on equity and the alleviation of poverty. In practically every country, labor earnings increase with educational attainment (see Table 3). The provision of universal primary education, apart from being highly profitable, has important egalitarian effects; it pushes people from the illiterate, low-income class into a higher income class. In the same way, higher education increases the supply of graduates, which tends to reduce the existing large income difference between university graduates and other types of labor by increasing the earnings of the latter. This proposition is based on the evidence of what has actually happened in advanced countries following a recent immense expansion in education (compare the relative dispersion of the index in the two columns of Table 3).

Table 3Earnings differences by educational level and country type
Country type
Educational levelAdvanced1Developing2
Primary (index base)100100
Source: Based on George Psacharopoulos, Earnings and Education in OECD Countries (Paris, Organization of Economic Corporation and Development, 1975).

Advanced countries include Canada, France, Greece, Italy, Japan, Netherlands, Norway, the United Kingdom, and the United States.

Developing countries include Ghana, India, Kenya, Malaysia, Nigeria, the Philippines, South Korea, and Uganda.

Source: Based on George Psacharopoulos, Earnings and Education in OECD Countries (Paris, Organization of Economic Corporation and Development, 1975).

Advanced countries include Canada, France, Greece, Italy, Japan, Netherlands, Norway, the United Kingdom, and the United States.

Developing countries include Ghana, India, Kenya, Malaysia, Nigeria, the Philippines, South Korea, and Uganda.

Again, the summary picture given in this table masks considerable variation in the structure of earnings between individual countries. Thus, in Ghana, the ratio of university to primary school graduate earnings is 9.1 to 1 and in Malaysia 7.7 to 1. However, in a country like the Philippines with a fairly well developed system of higher education, the corresponding ratio is 2.2 to 1.


The global proposition that spending on education is an investment with a high social rate of return is well supported by the accumulated evidence over the last decade. It can also be confidently concluded that education, especially at the lower level, contributes to the alleviation of poverty, a more equal distribution of income, and an improved social environment.

Although there appears to be considerable room for further investment in educational projects, often little is known about the factors determining the type of investment needed in individual circumstances. One reason for the delay in applying cost-benefit analysis, for instance, is that it requires time-consuming compilation of data. Age-earnings profiles, such as those discussed in the box, have only recently become available for a large number of countries. Several Bank projects now contain provision for the generation of such data in countries where they do not exist or where the Bank needs updated information for its operation. More research needs to be done on returns to educational lending in countries where studies have not yet been made, existing studies need to be updated, and more work needs doing on the relative profitability of spending on different curricula. The economic performance of those engaged in the nonformal or nonwage sectors in developing countries, especially in agriculture and self-employment, has only very recently begun to be seriously analyzed. A Bank-sponsored research program is currently underway to provide information on these and other issues in order to make the lending program more effective.

Related reading

Other Resources Citing This Publication