No open-access location is known for this record yet.
No open-access location is known for this record yet.
Census data from 1960 and 1970, augmented with schooling data (expenditures per pupil as well as other measures of school quality) from the various Biennial Surveys of Education, are utilized to analyze the impact of school quality on earnings and to investigate secular changes in rates of return to school quality. This research indicates that the effect of expenditures per pupil on earnings is smaller than that estimated by previous researchers. However, there has been a substantial improvement for blacks in the last decade, while the returns to the investment in the quality of education of white males have been unchanged. The major focus of research on the economic benefits of education has been on the returns to additional years, that is, the quantity, of schooling. Comparatively little emphasis on the returns to school quality can be found in the literature, primarily because of the lack of adequate data. However, in recent years a literature that uses data on the resource inputs into schooling as a proxy for quality has emerged. Most of this literature uses average per pupil school expenditures as a measure of school quality. The returns to school expenditures have been examined with small and/or specialized data sets by Morgan and Sirageldin [12], Johnson and Stafford [7], Morgenstern [13], Ribich and Murphy [14], Link and Ratledge [10, 11], Wachtel [17, 18], and most recently in this Journal by Akin and Garfinkel [1]. In addition, several authors have examined the returns to other measures of school quality such as teacher-pupil ratios and the length of the school year (see Wachtel [17] and Welch [19]). Rizzuto is a member of thefaculty of the University of Den ver. Wachtel is a member ofthe faculty of the Graduate School of Business Administration, New York University. [Manuscript received June 1978; accepted February 1979.] The Journal of Human Resources * XV * 2 0022-166x/80/0002-0240 $01.00/0 ? 1980 by the Regents of the University of Wisconsin System This content downloaded from 207.46.13.62 on Fri, 14 Jul 2017 17:38:56 UTC All use subject to http://about.jstor.org/terms Rizzuto and Wachtel | 241 In general, the conclusions to be drawn from this literature can be summarized as follows: First, expenditures per student exert a positive and significant influence on earnings. Second, the returns to school quality are substantially greater for blacks than for whites. Third, there are diminishing returns to per pupil expenditures on schooling. Fourth, there exists a tradeoff between the quantity and quality of education (i.e., years of schooling and expenditures per pupil can be viewed as substitutes for one another). Fifth, society's marginal rate of return to the investment in school quality is at least as large as its marginal return to investment in additional years of schooling. Finally, expenditures per student have a positive effect on the level of educational attainment.' In spite of these findings, there are several gaps in the research on school quality. First, the finding that returns to school quality for blacks are substantially greater than those for whites must be viewed as tentative because they are often based on small samples of blacks. Second, because of the lack of comparability among the samples utilized, there has been no discussion of the secular movements in the marginal social rates of return to school quality and, in particular, the change in rates of return of blacks relative to whites. Finally, most results are based on expenditures per pupil as a measure of school quality, which is likely to be an imperfect proxy.2 In this paper, these issues are investigated with earnings data from the U. S. Census of Population for 1960 and 1970 which is augmented, as discussed below, with schooling data from the various Biennial Surveys of Education. This data set is particularly useful in addressing these issues because it provides large and representative samples of the population for both blacks and whites. In addition, this unique data set provides comparable samples for 1960 and 1970 so that secular changes in the returns to schooling can be examined. 1 However, Morgenstern [13] and Ribich and Murphy [14] found little or no relationship between expenditures per student and earnings (i.e., average hourly wage for Morgenstern and lifetime earnings in the case of Ribich and Murphy). Aspects of the methodology utilized in each case may be suspect. Morgenstern's school-quality variable was expenditure per pupil in the state where the person spent his formative years divided by the national average of expenditures per pupil for the same time period. The division by the national average was supposed to remove the effect of inflation in school costs, but may also introduce errors in variables. Ribich and Murphy use data on individual earnings and occupation obtained five years after scheduled high school graduation to construct estimates of lifetime earnings. Not only are a single year's earnings and occupation a poor proxy for future earnings, earnings at such an early age also are likely to be lower since individuals with large investments in education tend to have jobs with more on-the-job training. 2 This is because it is difficult to deflate expenditures for both changes in costs over time and differences in costs among areas. Of the studies mentioned, only Akin and Garfinkel [1] do both. Furthermore, a simultaneity problem makes interpretation of the earnings/expenditure relationship difficult. Areas with high average incomes are likely to have high average schooling expenditures, so the direction of causality is uncertain. This content downloaded from 207.46.13.62 on Fri, 14 Jul 2017 17:38:56 UTC All use subject to http://about.jstor.org/terms 242 THE JOURNAL OF HUMAN RESOURCES MODEL SPECIFICATION AND DATA The Structure of the Model The basic model that was utilized to investigate the relationship between earnings and human capital investment, in the form of schooling, is the following: lnY = ao + a1EXP + a2EXP2 + a3lnW + a4S + a51nQ + a6URB + U where lnY = the natural logarithm of annual earnings, EXP = years of labor market experience, EXP2 = years of labor market experience squared, lnW = the natural logarithm of the number of weeks worked, S = years of schooling, lnQ = the natural logarithm of annual expenditures per elementary school pupil in average daily attendance, URB = one/zero dummy variable to represent urban residence at the time of the Census, and U = the residual. The conceptual framework underlying this earnings function as well as its functional form have been well documented in the human capital literature.3 Briefly, years of experience (EXP) and the experience squared term (EXP2) reflect the net returns to investment in on-the-job training. The weeks worked variable (lnW) is a way of standardizing for variation in work effort, chiefly unemployment and work-leisure preferences. The years of schooling (S) and the school quality (lnQ) variables represent measures of the investment in human capital. The expenditure variable is in logarithmic form so that its coefficient is an estimate of the elasticity of earnings with respect to school expenditures. Finally, urban residence (URB) has been added since earnings tend to be greater in urban areas.
Loading…
Stored only when both papers are in the corpus — 68 citations exist upstream.
Papers that cite the same references — the relational fallback while embeddings are still backfilling.
Imported from OpenAlex (CC0) — nothing here is generated.