CHAPTER
7Comparing Census/National
Science Foundation R&D Data
with Compustat R&D Data
T
his chapter^1 reviews the results of a project in which research and de-
velopment (R&D) data drawn from National Science Foundation
(NSF) and U.S. Census Bureau studies were substituted in several finan-
cial models for R&D data drawn from the 1975–1982 Compustat tapes.
The result using the Compustat data did not differ significantly from that
based on the NSF/Census data in the aggregate, but significant differ-
ences were observed for certain industries and certain years. This chapter
discusses the financial models and determinants of corporate R&D ex-
penditures using the different databases, and suggests further questions
for research.
The general objective of this chapter is to review and summarize
findings as to whether R&D expenditures should be included in a set of
financial decisions that influence the market value of a firm, as reflected
in the price of its stock (Weston and Copeland 1986). A rigorous theoret-
ical position, known as the perfect markets hypothesis, has guided re-
search on this issue for many years, as discussed in the preceding chapter.
The perfect markets hypothesis asserts that the value of a firm’s stock is
determined by the firm’s ability to invest in opportunities that will pro-
duce enhanced earnings, dividends, or cash flow. It further asserts that
dividend policy is independent of these investment decisions; however,
the investment decision and the decision to issue new capital stock are
interdependent. Thus, the perfect markets hypothesis asserts that the
stock market value of a firm is not affected by dividend policy, only by
opportunities for further returns, as described in Chapter 6. This chapter
considers whether federal financing of R&D influences a firm’s financial
decision-making process.