The purpose of this paper is to describe the impact of investment in education on U.S. economic growth. Beginning with the seminal contributions of Becker , Mincer , and Schultz , economists have found it useful to characterize the benefits of education by means of the notion of investment in human capital.(1) This idea captures the fact that investment in human beings, like investment in tangible forms of capital such as buildings and industrial equipment, generates a stream of future benefits. Education is regarded as an investment in human capital, since benefits accrue to an educated individual over a lifetime of activities.
One of the most important benefits of education is higher income from participation in the labor market. This increase in income is the key to understanding the link between investment in education and economic growth. People differ enormously in effectiveness on the job. Substituting more effective for less effective workers increases output per worker. More highly educated or better trained people are more productive than less educated or poorly trained people. However, education and training are costly, so that substitution of people with more education and training requires investment in human capital.
The most common approach to compiling data on educational investment is to measure the inputs, rather than the outputs, of the educational system.(2) Data on the expenditures of educational institutions for teachers and other personnel, buildings and equipment, and materials can be compiled from accounting records. This information can be supplemented by estimates of the value of time spent by students (and their parents) as part of the educational process. Costs of schooling and the value of the time spent by students can be used to measure the flow of resources into schools and universities.
While the costs of education are highly significant in economic terms, the cost-based approach to measurement of educational investment ignores a fundamental feature of the process of education. This is the lengthy gestation period between the application of educational inputs–mainly the services of teachers and the time of their students–and the emergence of human capital embodied in the graduates of educational institutions. Furthermore, some of the benefits of investment in education, such as greater earning power, are recorded in transactions in the labor market, while others–like better parenting or more rewarding enjoyment of leisure–remain unrecorded.(3)
A measure of output is needed to put the education industry on par with other industries producing goods and services. This measure must reflect the fact that education is a service industry, but its product is investment in human capital. Since the effects of formal schooling endure through the lifetime of an educated individual, the authors employ the impact of education on an individual’s lifetime income as a measure of educational output. A second important idea is that education enhances the value of activities outside the labor market, such as parenting or the value of leisure time. The estimates of the output of the education sector incorporate the value of time spent outside the labor market.
In measuring investment in education, the first step is to compile data on the economic value of labor market activities. In Section II of the paper, it is shown that the constant dollar value of time spent working has doubled in the postwar U.S. The growth of this value has been greater or the decline has been less for women than for men at all levels of educational attainment. This reflects the rapid increase in labor force participation by women relative to men. The proportional increase in the value of market labor time has been greatest for college educated men and women. This corresponds to the substantial growth in levels of educational attainment.
The second step in measuring investment in education is to estimate the value of nonmarket labor activities. These include time spent in investment in education as well as time spent in the consumption of leisure. The authors infer rates of compensation for nonmarket activities assigned to consumption from market wage rates. The value of these nonmarket activities, measured in this way, exceeds the value of market activities, primarily because nonmarket time exceeds time in the labor market. However, the value of nonmarket labor activities has grown more slowly. The expansion of the value of nonmarket time has been more rapid for men than for women. These findings are discussed at greater length in Section II.
In Section III of the paper, the authors estimate lifetime labor incomes for all individuals in the U.S. population. These incomes include the value of both market and nonmarket labor time. The impact of increases in educational attainment on the lifetime incomes of all individuals enrolled in school is estimated. The authors find that investment in education, measured in this way, is greater in magnitude than the value of working time for all individuals participating in the labor force. Furthermore, the growth of investment in education has exceeded the growth of market labor activities during the postwar period. Investment in education has increased much more rapidly for women than for men, especially at the college level.
In Section IV of the paper, the authors measure the effects of changes in the educational composition of the labor force on the growth of the noneducation sector of the economy. This sector includes business, government enterprises, and general government but excludes private and public educational institutions. The authors present a new measure of changes in the quality of hours worked. This measure is based on the impact of substitution among workers with different characteristics–including age, sex, and education–on labor input per hour worked. The growth of labor quality with an increase in the proportion of more highly trained and better educated workers is then identified.
Second, the authors measure the inputs of the education sector, beginning with the purchased inputs recorded in the outlays of educational institutions. They also measure the inputs of time for all students enrolled in formal education. A major part of the value of the output of educational institutions accrues to students as increases in their lifetime incomes. Treating these increases as compensation for student time, one can evaluate this time as an input into the educational process. Given the outlays of educational institutions and the value of student time, one can allocate the growth of investment in education to its sources.
Finally, the authors aggregate the growth of education and noneducation sectors of the U.S. economy to obtain a new measure of U.S. economic growth. This includes the growth of the business and government sectors, as measured in the U.S. national income and product accounts. It also includes the growth of the educational sector, as measured by the new data on investment in education. Combining these measures of output growth with the corresponding measures of input growth, the authors obtain a new set of accounts for the growth of the U.S. economy. In these accounts, the scope of output is increased by investment in education, and the scope of input is increased by the value of student time.
The authors present the conclusions of this study in Section V. The most important finding is that investment in human and nonhuman capital accounts for the largest part of U.S. economic growth during the postwar period. This finding characterizes the noneducation sector of the U.S. economy as well as the economy as a whole, including the education sector. Since 1973, the growth rate of the U.S. economy has slowed by almost four-fifths of a percentage point, relative to the postwar average. A revival of U.S. economic growth will require the mobilization of increased capital and labor resources. Educational investment will continue to predominate in the investment requirements for more rapid economic growth.
II. Market and Nonmarket Labor Incomes
In order to measure investment in human capital as an output of the educational system, the authors have constructed a new database for measuring lifetime labor incomes for all individuals in the U.S. population. This database includes demographic accounts for the population in each year, cross-classified by sex, individual year of age, and individual year of highest educational attainment. The demographic accounts include data on the number of individuals enrolled in formal schooling and the numbers employed. These demographic accounts are based on annual population data from the U.S. Bureau of the Census.(4)
Table 1 presents the estimates of numbers of students between 5 and 34 years of age enrolled in school, cross-classified by sex and level of education.(5) Enrollments in grades 1-8 and high school peaked during the late 1960s or the 1970s and have gradually drifted downward through 1986, the last year for which the data are available. Enrollments in college flattened in the 1980s for both men and women and have begun to decline. Enrollments in primary schools have increased over the period 1947-86 as a whole, while enrollments in secondary schools have nearly doubled. Enrollments in higher education have risen very dramatically, especially for women.
To measure lifetime labor incomes for all individuals in the U.S. population, the authors begin with a database on market activities constructed by Gollop and Jorgenson [1980, 1983]. They derive estimates of hours worked and labor compensation for each sex by 61 age groups and 18 education groups for a total of 2,196 groups for each year. Table 2 presents the estimates of the value of time spent working, cross-classified by sex and educational attainment, for all individuals in the U.S. economy from 1948-87. Estimates of the value of labor time are given in this table in constant prices.
Labor time in constant prices is a quantity index number, defined in terms of annual hours worked for individuals cross-classified by age, sex, and educational attainment. To construct a quantity index of labor time, the authors weight these hours worked by average compensation per hour. They assume that labor time can be expressed as a translog function of its 2,196 components. The growth rate of the corresponding quantity index is a weighted average of growth rates of these components. The weights are given by the shares of each component in the value of market labor time. A quantity index of labor input is unaffected by inflation in rates of labor compensation.
The constant dollar value of market labor activities has more than doubled over the postwar period. However, the quantity index for workers with eight or fewer years of educational attainment has declined substantially. For workers with high school education, the quantity peaks in 1979 for males and rises throughout the period for females. Finally, working time in constant prices increases by more than four and a half times for college educated males and almost seven times for college educated females. The constant dollar value of working time for males with a college education exceeds that for high school
educated males, beginning in 1980, while the value for college educated females exceeds that for females with a high school education at the end of the period in 1987.
The authors turn next to the task of evaluating labor time spent in nonmarket activities, considering activities such as formal schooling, that enter into investment in human capital and activities that result in consumption. The importance of evaluating time spent in nonmarket activities is widely recognized.(6) For example, Nordhaus and Tobin  have incorporated measures of the value of these activities into their measure of economic welfare. Kendrick  and Eisner  have also imputed values for time spent outside the labor market. Five types of nonmarket activities are commonly distinguished in studies of time allocation–household work, human capital investment, travel, leisure, and maintenance–the satisfaction of physical needs such as eating and sleeping.(7)
The authors allocate the total time available for all individuals in the population among maintenance, work, school, and household production and leisure. Studies of time allocation show that maintenance time per capita has changed very little during the postwar period. They estimate that time spent in maintenance is 10 hours per day per person and exclude this time from the measure of the value of nonmarket activities. The time spent in formal education for all individuals enrolled in school is estimated and allocated to investment. Finally, the authors allocate the time not spent on maintenance, work, or school to consumption.
The authors impute rates of labor compensation for nonmarket activities assigned to consumption from wage rates for employed individuals with the same age, sex, and educational attainment. Market wage rates are reduced by taxes on labor incomes estimated by Jorgenson and Yun . They impute rates of labor compensation for time assigned to schooling from the value of investment in education. Almost all of the value of investment in education, except for tuition charges and out-of-pocket expenses such as books and school supplies, accrues to individuals enrolled in school as lifetime incomes increase. These increases, net of tuition and school expenses, comprise the labor compensation for the nonmarket time of students.
Table 3 gives the value of nonmarket activities assigned to consumption in constant prices, cross-classified by sex and educational attainment, for all individuals in the U.S. population for the period 1948-87. As before, nonmarket time in constant prices is a quantity index number, defined in terms of hours of nonmarket time for all 2,196 categories of workers. While nonmarket time in current prices reflects inflation in imputed rates of compensation, the quantity index number is unaffected by inflation.
The value of nonmarket activities assigned to consumption in constant prices exceeds the value of market activities by a factor of two. This is due to the fact that nonmarket time, as it is measured, is greater than time spent at work. For the population as a whole, the growth of the value of nonmarket time is roughly comparable to the growth of the value of work time. However, the distribution of this growth is considerably different.