change in its financial decisions would improve stockholder wealth and, if so,
how the changes should be made.^13 As shown in Table 7.9, these results sug-
gest that the firm, given its asset base, was spending less than optimal levels
on R&D, capital replacement/expansion, and dividends, and that it incurred
excessive new debt. Under the assumption that its asset base would remain
the same, these results imply that the firm could increase its stock price in
1983 by increasing its expenditures on worthy R&D and capital projects, in-
creasing dividends, and reducing issuance of new debt. As pointed out in
Guerard and Bean (1987), this approach to maximizing stockholder wealth
(that of formulating optimal financial policies) was shown to be more effec-
tive than approaches that attempt to maximize return on equity or growth in
earnings per share, as discussed by Rappaport (1983). In addition, the ap-
proach presented here offers econometric support for the financial modeling
approach and extends the theoretical models of Carleton (1970), Hamilton
and Moses (1973), and Burton, Damon, and Obel (1979, 1984).
Given these interesting results from the Compustat data, the analysis
was replicated using the 1982 NSF/Census data; the results are shown in
Table 7.10.
Clearly, the two data sets produce somewhat different results. The
NSF/Census data call for smaller adjustments than the Compustat data.
However, the adjustments are all in the same direction relative to the firm’s
actual allocations, and with respect to the industry averages. Thus, in both
196 COMPARING CENSUS/NSF R&D DATA WITH COMPUSTAT R&D DATA
TABLE 7.8 1982 Allocations and Commitments ($ millions)Sample Firm Industry AverageR&D $ 68 $148
Capital Expenditures 770 799
Dividends 42 117
New Debt 577 231TABLE 7.9 Sample Firm—Results of Optimization Procedure Using 1982
Compustat Data ($ millions)
Sample Firm Optimal Levels Under/OverR&D $ 68 $199 $–131
Capital Expenditures 770 924 –155
Dividends 42 278 –236
New Debt 577 475 102