Most empirical studies of economic growth attempt to determine the relative importance of increases in inputs and advances in technology in the achievement of growth in per-capita output. This approach is motivated by a desire to explain the sources of that output growth: How much less rapidly would the U.S. economy have expanded in the last fifty years if it had continued to operate with 1918 levels of technology or if technology had advanced but no net investment in tangible or human capital had occurred? Answers to these questions help us to maximize our future rate of growth by guiding policymakers to an optimal allocation of resources among investment in tangible capital, in education, and in technology-increasing activities, and they help in explaining the reasons for international differences in per-capita income.

Since the mid-1950s a common technique for the separation of the respective contributions of input growth and advances in technology has been the calculation of indexes of total factor productivity. Pioneering studies by Solow (1957, pp. 312–20) (Abramovitz, pp. 5–23; Kendrick, 1956; Schmookler, 1952, pp. 214–231) and others have suggested capital played only a minor role in per-capita growth, and that most of the long-term increase in U.S. output per capita was due to an increase in the output obtainable per unit of appropriately weighted input. While it was recognized that some of this increase in total factor productivity or “the residual” might have been due to the spread of education, most of it was assumed to have represented technical change.