Financial Stress Can Squeeze the ‘R’ Out of R&D

Featured in print Digest

This is a line graph titled, Research and Development Spending by US Businesses. The y-axis is labeled, Billions of 2012 USD, ranging from 0 to 350, increasing by increments of 50.  The x-axis is year, ranging from 1993 to 2015, increasing by increments of 2. There are three lines on the graph, Total R&D, Development, and Research. The total R&D line increases from the middle of the graph towards the top right. It is not a consistent increase but overall, the line is generally increasing over the sample period except for a couple of small dips. It starts at around 175 in 1993 before increasing to 250 in 1999, 300 in 2007 and finishing close to 350 in 2015. The Development follows the same exact pathway as the R&D line except at a lower level. It starts at 125 in 1993, almost 200 in 1999, 225 in 2007, and finishing above 250 in 2015.  The Research line experiences minimum variation and is near horizontal for much of the study period. It hovers around 50 from 1993 to 2003 before increasing slighting to 60 by 2007 and then finishing at 75 in 2015.  The source line reads, Source: Researchers’ calculations using data from the National Science Foundation.

A financial crisis can affect the economy’s long-term trajectory if it alters investment behavior. In Research and/or Development? Financial Frictions and Innovation Investment (NBER Working Paper 31521), Filippo Mezzanotti and Timothy Simcoe investigate the impact of the 2008 financial crisis on domestic research and development (R&D) expenditures by US firms. They find that firms in greater need of refinancing responded to the crisis by cutting back investment in R&D. Scientific research felt the brunt of the cutbacks, while product development was less affected.

The study analyzes a sample of around 1,100 large US firms and draws on census surveys for 2002–07 and 2008–12 that make it possible to disaggregate spending on research from that on development. Private firms account for 75 percent of total US domestic investment in R&D. Firms entering 2008 with a higher debt-to-liquid-assets ratio saw less R&D growth than peers within the same North American Industry Classification System category with a lower ratio. A 0.5 unit increase in the ratio was associated with an 8 percent decline in R&D spending. The decline came almost entirely from reduced spending on basic and applied research, and mostly from cutting employment.

Cutbacks in corporate R&D spending during the Great Recession fell most heavily on budgets for basic and applied research.

The decline was not associated with preexisting trends in the companies’ R&D activities. Moreover, the drop in R&D investment persisted beyond the crisis, perhaps because of the high cost of replacing highly skilled workers.

Basic and applied research is more vulnerable to the budget axe than product development or other activities with specific commercial applications. Its goal is to generate new knowledge in the hope of a long-term payoff, in contrast to development research, which offers quicker, less speculative returns. Spending on development often seeks to improve existing products or create new ones. The researchers suggest that development investment may be driven more by competitive considerations than by the financial environment. They conjecture that firms will reduce development investment in a crisis only if their rivals face similar problems, such as accessing credit.

The long-term economic impact of the R&D cutbacks can be seen in trends for patents and patent citations among firms with greater exposure to the financial crisis. Patents fell by 6 to 7 percent over a three-to-five-year horizon, but these differences are not statistically different from zero. However, the quality of these patents decreased significantly: patent citations fell by 16 to 19 percent over the same period. To explain these findings, the researchers show that development spending is more strongly tied to patents than research spending is. This evidence may explain why a change in R&D mostly achieved through a reduction in research may have small effects on the raw count of patents, but still deteriorate the quality of this output.

—Steve Maas

This research was performed at a Federal Statistical Research Data Center under FSRDC Project Number 1874. (CBDRB-FY22-P1874-R9607, CBDRB-FY23-P1874-R10456).