The dividend (DIV) equation (Table 7.5) is also stable for the
1978–1982 period. However, it is noteworthy that the new debt variable
(EF), which was significantly and negatively related to dividends in
1975–1977, did not hold up in the 1978–1982 period. Thus, the data sup-
port the following dividend equation in all cases:
The new debt (EF) equation (Table 7.6) must also be modified because
of the time frame covered in this study. While both cash flow and dividends
were significantly and negatively related to new debt over the 1975–1982
time frame, cash flow does not hold up between 1978 and 1982, and divi-
dends waver somewhat. For the 1978–1982 period, the Compustat 303
data support the following model:
The longer Compustat period (1975–1982) model is:
Using the reduced Compustat sample, the equation for the 1978–1982
would become:
Using the NSF/Census data in the reduced sample, the 1978–1982
equation is:
These results provide a mixed picture of the new debt equation.
Clearly, new debt is issued to finance capital expenditure in all cases.
However, the direction of the dividend relationship seems sensitive to the
time frame issue. It is significantly and negatively related to new debt ex-
cept for the 1978–1982 time period using the Compustat 303-firm sam-
ple, when it is marginally positive. For the reduced Compustat sample,
CENS(78 82): EF CE, DE, DIV++
−=−
f()RCOMP(78 82): EF = (CE DIV+
−−
f ,)COMP303(78 82): EF = CE, DIV CF++
−−
f(,)COMP303(78 82): EF = CE, DIV++
− f()DIV = (DIV, NI+
f t−+
1 )