June 1, 2003
Robert W. Arnold
This paper reviews the recent empirical literature on long-run growth to determine what factors influence growth in total factor productivity (TFP) and whether there are any channels of influence that should be added to standard models of long-run growth. Factors affecting productivity fall into three general categories: physical capital, human capital, and innovation (including other factors that might influence TFP growth). Recent empirical evidence provides little support for the idea that there are extra-normal returns to physical capital accumulation, nor is there solid justification for adding a separate channel of influence from capital to TFP growth. The paper finds evidence that human capital—as distinct from labor hours worked—is an important factor for growth but also that there is not yet a consensus about exactly how it should enter the model. Some argue that human capital should enter as a factor of production, while others argue that it merely spurs innovation. The forces governing TFP growth are not well understood, but there is evidence that R&D spending is a significant contributor and that its benefit to society may exceed its benefit to the company doing the spending—that is, it is a source of spillovers. The paper concludes with some examples of how standard models of growth could be modified to reflect some of the channels of influence identified in the empirical review.