کمیت و کیفیت آموزش و رشد بهره وری نیروی کار ایالات متحده (1870-2000)

This paper accounts for the contribution of the quantity and quality of schooling to worker productivity growth in the United States from 1870 to 2000. Schooling investments rose dramatically over the period before leveling off around 1970. Schooling likely caused 30 to 40 percent of the fivefold rise in worker productivity from 1870 to 1970 and produced a “wave pattern” in productivity growth (previously attributed solely to the timing and diffusion of important technological innovations). The results suggest that about 1 percent of the century-long 1.6-percent growth rate in worker productivity is sustainable. Journal of Economic Literature Classification Numbers: O47, J24.

The behavior of labor-productivity growth during the last quarter of the twentieth century has been concerning and puzzling. From 1972 to 1995, the U.S. economy experienced a “productivity slowdown”—with an annual productivity growth rate of about 1 percent. This was almost a percentage point lower than the 1.86-percent growth rate from 1913 to 1972 (Gordon,2000, Table 1, p. 53). Over the last five years of the century, productivity growth rebounded to levels above 2.5 percent. This raises the question of what long-run labor productivity growth will be in this century. Is labor productivity likely to grow faster or slower than it did in the twentieth century? The answer to this question must begin with an explanation of growth during the twentieth century. Identifying the underlying sources of historical growth in the twentieth century was “transitional,” and therefore not likely to be sustained indefinitely into the next century.There is reason to suspect that a portion of the growth in the last century was unsustainable. The share of resources devoted to education increased dramatically over the twentieth century and then leveled off toward the end of the century. Since investment shares cannot rise indefinitely, this source of growth must be viewed as transitional. Thus, if education investment was an important source of past growth, one would naturally expect slower growth in the future. This paper attempts to identify the underlying sources of historical growth by using a neoclassical growth framework to decompose growth into transitional growth, resulting from physical and human capital accumulation,and sustainable growth due to exogenous technological change.Our approach differs from previous theoretical growth accounting work by (i) accounting for changes in both the quality and quantity of education investments in U.S. history, (ii) identifying the causal role of schooling, (iii) exploring whether micro-econometric estimates of the effects of schooling are consistent with macroeconomic observations of productivity growth and interest rates, and (iv) attempting to explain the rise in the quantity of schooling within a microeconomic model of the family.Many recent theoretical and policy papers assume school spending has important effects on worker productivity and economic growth.2 The empirical justification for this common assumption is questionable. Macroeconometric evidence on the causal link between schooling and growth is mixed and subject to problems of measurement and interpretation.3 After a critical review of this literature, Krueger and Lindahl (2000) doubt that cross-country growth regressions will be “very informative insofar as the benefit of education is concerned” (p. 37). The micro econometric literature has also reached mixed conclusions about the causal effects of school spending (see the review in Section 2). The numerical simulation model presented here helps to resolve the debate over school spending. Micro-econometric estimates of schooling effects can be incorporated to see if they are consistent with macroeconomic observations such as the rise in school spending, the rates of return to physical and human capital, and the growth in labor productivity over the century. Consistency with aggregate observations can then be used as a criterion to evaluate the validity of different micro-econometric estimates. Section 1 provides a historical summary of labor-productivity growth and the trends in education investments for more than a century.Section 2 develops the neoclassical growth model used to distinguish transitional growth due to capital accumulation from growth due to technological change. Section 3 discusses the rates of return to physical and human capital over the century. Section 4 conducts the growth accounting that identifies the role of human capital in productivity growth. Section 5 attempts to explain the rise in schooling investments of older children, and Section 6 attempts to explain the behavior of interest rates. Section 7 examines how results are affected by alternative specifications of the model. Section 8 discusses several questions raised about the analysis. Section 9 draws conclusions about growth rates in the twenty first century.

Schooling investments of time and goods rose dramatically from 1870 to 1970. The rising investment shares helped to mediate the diminishing rates of return to these investments. As a consequence, the trend growth rate in worker productivity remained relatively constant over the century. The precise timing of the investment rise generated a “wave pattern” over the period 1870–2000 (with relatively high growth rates in the middle years). This pattern was previously explained by the timing and prolonged diffusion of unusually important technological innovations during the period 1860 to 1900 (Gordon, 2000). Our results suggest that the historical patterns of schooling investments, in combination with diminishing returns, also played a role in creating the wave.The model indicates that much of the growth in the twentieth century was transitional and not likely to be sustained into this century. It was shown that a large transitional component to our historical growth is consistent with trendless productivity growth rates and interest rates. The rise in schooling-investment shares accounts for 30 to 40 percent of the fivefold increase in worker productivity from 1870 to 1970. This implies an underlying rate of exogenous technological change of about 1 percent. Jones (2001) suggests that most of the technological change occurring between 1965 and 1990 was due to increasing investment shares in R&D, and thus is also not sustainable. The long-run productivity growth rate in this century may well be less than 1 percent, rather than the 1.6 percent observed over the twentieth century. Viewing human capital as an intergenerational transfer to dependent children predicts that time investments of older children at the beginning of the century were about half of current values. The private return to schooling (and perhaps learning in general) remained well above the return to physical capital throughout most of the century. This can be explained by the fact that schooling investments are largely decisions made by families who are unable to borrow across generations. This caused schooling investments to be below privately efficient levels until the end of the century.