Do Business Cycles Alter the Composition of Research and Development Expenditures?
Rafferty, Matthew, Contemporary Economic Policy
This article examines one channel linking business cycles and growth: the composition of firm-financed research and development (R&D) expenditures. The data suggest that expenditures on basic scientific research rise during recessions but development R&D expenditures fall during recessions. The opposite pattern occurs during expansions. This means recessions increase the rate at which the scientific frontier expands but slow the rate at which firms transform basic scientific knowledge into new products. Likewise, expansions allow the economy to exploit the existing frontier but slow the frontier's growth. These results raise the possibility of growth-- business cycle interaction, which has important implications for stabilization policy.
Researchers often focus on either the effects of cash flow or the opportunity cost of inventive activity as a means of linking business cycles and long-run growth. (Aghion and Howitt  and Saint-Paul  provide useful surveys of the existing literature on the interaction between growth and business cycles.) Himmelberg and Petersen (1994) found a significant relationship between firm R&D expenditures and internal finance using a panel of small high-tech firms. Indeed, Hall (1992) found that firms finance almost all of R&D through cash flow and almost none through debt. This research suggests that cash flow influences R&D expenditures and that these expenditures are procyclical.
Examples of the opportunity cost literature include Bean (1990), who finds that negative demand shocks stimulate human capital growth; Hall (1991), who argues that firms reorganize during recessions; and Saint-Paul (1993), who finds that negative demand shocks stimulate total factor productivity growth. The explanation is that the opportunity cost of inventive activity is countercyclical. Recessions temporarily reduce the return to production and expansions temporarily raise the return to production. Because most of the return to inventive activity is in the distant future, the business cycle does not significantly influence its return. Therefore, the relative return to inventive activity should rise during recessions and fall during expansions, causing firms to substitute between production and inventive activity over the business cycle. This line of argument implies that R&D expenditures are countercyclical.
Countercyclical behavior of firm-financed R&D expenditures might help explain some interesting empirical results relating total factor productivity growth and the business cycle. Gali and Hammour (1992) used aggregate U.S. data and found that negative aggregate demand shocks stimulate productivity growth. Malley and Muscatelli (1999) used industry-level data from the U.S. manufacturing sector and found that temporary negative shocks to employment can increase total factor productivity in the long run. One explanation for the countercyclical behavior of total factor productivity is increased firm-financed R&D during recessions.
Saint-Paul (1993) finds no relationship between his measure of the business cycle and aggregate firm-financed R&D expenditures. However, he treated firm-financed R&D as a homogenous quantity, which is not necessarily the case. Furthermore, it is possible that firms alter the mix of R&D expenditures while leaving total R&D expenditures unchanged in response to the business cycle. The results in this article suggest that types of R&D expenditures exhibit different cyclical behavior; more basic R&D expenditures are countercyclical, whereas development R&D expenditures are procyclical. The changing mix of R&D expenditures over the business cycle might help explain the why Gali and Hammour (1991) and Malley and Muscatelli (1999) find that negative demand shocks stimulate total factor productivity.
II. R&D AND THE OPPORTUNITY COST EFFECT
The opportunity cost effect may exist, but the margin at which firms can substitute between inventive activity and other activities is small. …