coefficients of the relative price term, while not statistically significant during the non-stabilization period, exhibits the correct sign in all four variants during the stabilization period and were significant at the 10 percent level in two of these cases. Coefficients of the relative price term in its noncompound form are generally not significant. Finally, the Koyck lag formulation again proves uniformly superior to that of the simple partial adjustment lag model. Constant Dollar Regressions-Using General Dummy Table 2 replicates the variants of the regression model using the change in constant dollar fixed investment as the dependent variable. The first set of eight regressions (equations 17-24) again uses a general dummy variable to capture the overall effect of the Stabilization Program on real fixed investment. As in the current dollar regressions, in each case the general dummy variable exhibits a positive sign and is highly significant. Similarly, as in the current dollar estimates, the accelerator and capacity utilization variables both have positive signs and are strongly significant in each of the eight variants. As in the current dollar regressions, the dividend term and the bond yield term remains generally insignificant with only one variant in which the bond yield registers significance at the 10 percent level. The relative price term in ratio form exhibits the correct sign in all four of the variants but is significant in only two cases. In testing the price and wage variables of this ratio separately, the price term uniformally exhibits the theoretically correct negative sign and is highly significant while the compensation per manhour term registers mixed signs and was uniformally not significant. In this formulation both the simple partial adjustment model and the Koyck version perform well for each of the variants. However, as in the current dollar estimates, the Koyck version appears to produce superior results. Constant Dollar Regressions-Using Slope Dummy Variables The final eight regressions (equations 19-24) test the model using slope dummies with the change in real fixed investment as the dependent variable. As in the current dollar estimates, the accelerator and the capacity utilization variables registers the theoretically correct sign and remain highly significant. Once again, the bond yield variable remains generally insignificant with only two of the eight possible variants significant at the 10 percent level. The slope dummy variable for the bond yield, however, while not significant, does exhibit the expected negative sign. As in earlier equations, the profit variable is uniformally insignificant in each of these eight variants for both the non-stabilization and the stabilization period. The dividend variable, again similar to the current dollar estimates exhibits-during the non-stabilization period- the theoretically incorrect positive sign and is insignificant; but during the stabilization period the variable exhibits the correct sign in all four cases. However, the dividend variable is significant in only two out of the four regressions. The relative price term in its ratio form, as in the models discussed above exhibits-during the non-stabilization period-the theoretically correct negative sign but is insignificant in each of the four variants tested. However, during the stabilization period the term exhibits both the correct sign and is significant in all four cases. When estimated separately, the results are similar to the current dollar estimates. The price term showed the correct sign and was significant during the non-stabilization period in each of the four cases; however, the results indicate this variable is not significant during the stabilization period. The compensation per manhour variable shows mixed results but has the correct sign and is significant in two of the four variants during the stabilization period. For the adjustment structure, the Koyck formulation again proves superior to the simple partial adjustment model. SOME IMPLICATIONS OF THE REGRESSION As noted in the earlier theoretical discussion, the measurement of the impact of so complex a program as the Economic Stabilization Program on so complex a variable as business fixed investment, is extremely difficult. Barring difficulties in estimation techniques, a more precise measuerment would undoubtedly require a "systems" approach wherein direct and indirect feedback effects of a simultaneous equation model would be possible. However, within the limitations of the single equation, partial equilibrium regression models tested above, some tentative conclusions might still be reached. As noted above, in all of the 16 variants using the general dummy formulation, the overall Stabilization Program appears to have had a positive and statistically significant influence on business investment during the stabilization period. However, when approaching this problem by examining the impact of the Program on individual determinants of investment behavior it is more difficult to reach such a clear cut interpretation. In terms of the general influence of the Program, the overall positive impact on investment spending might be attributed to what businessmen saw initially as a resolution of Federal policy towards dealing with inflation. Much of their concern prior to the controls, centered on restraining wages, as evidenced by the session of the Business Council in Hot Springs in October, 1970, when an overwhelming vote endorsed wage controls. Another factor might have been that businessmen saw an end to the sluggish 1970-71 recovery as the government announced its intention to strongly stimulate the economy. Hence, investment spending plans that may have been deferred earlier due to the slow pick-up in the economy were implemented during the stabilization period. Businessmen's expectations, in addition, may have been reinforced as signs of consumer confidence increased during late 1971 and 1972. By mid-1973, this confidence by both businessmen and consumers was no longer apparent. However, investment spending during the last 1973-early 1974 period probably reflected decisions made prior to this period. In any case, tests performed to split the stabilization period into Phases I-II, and III-IV to try and capture this "change in mood" did not produce meaningful statistical results. Examination of the slope dummy regression results suggests that the Program did have a marginally positive impact on investment spending through the increased capital available as a result of the voluntary dividend guidelines and by holding down prices of capital goods relative to wage rates. On the other hand, the Program did not appear to significantly influence investment by its controls on profit margins or by the CID measures calling for restraint on interest rates. With respect to the bond yield variable, an examination of interest rate movements during the Economic Stabilization Program period gives some explanation for this outcome. Short-term interest rates during this period swung widely-first declining from the high rates prevailing during the 1969-70 credit crunch and then sharply increasing throughout 1972 and 1973 in response to increased loan demand, and subsequently to the high rates of inflation and actions. by the Federal Reserve. Long-term rates, on the other hand, which more directly relate to long-term investment financing, remained relatively high and stable in the range of seven to eight percent throughout most of this period. In addition, it might be noted that the bond yield variable generally has not shown up in the empirical literature as a strong factor in investment spending decisions. Thus, it is not surprising that this variable did not register significantly in the equations tested either during the stabilization or non-stabilization periods. The failure of the regression results to show a more significant role for the profits variable is perhaps more surprising. Many of the strongest complaints against the Program were leveled at the profit margin regulations, charging in particular that this restraint inhibited investment spending. An examination of profits data during this period may give some indication for this outcome. During the early part of the Program, most companies had ample room with respect to their profit margin limitation. As noted eariler, the profit margin regulations were specified in terms of profits to sales rather than the return on investment or equity capital. During late 1971 and 1972, firms strongly expanded their volume of sales and hence were able to keep within their profit margin ceilings. It was only during the latter part of the Program, when these volume increases were not as strong, that this regulation might have had some impact on profits. In terms of the regressions tested in this paper, it should be noted that profits do not appear to influence investment spending as strongly as an increase in demand or a change in relative pricesa result generally borne out by the empirical literature. For this reason, even if the Program were to have had a stronger impact on profits, it would appear that such an impact would not have significantly influenced investment spending. Some positive impact on investment spending through the Program's influence on individual variables is suggested in the coefficients of the dividend and relative price terms. Some businessmen would argue that by holding down dividends, equity prices were adversely affected thereby limited firms' ability to raise equity capital and hence increase investment. However, most theoretical and empirical discussions would argue that, by and large, dividend behavior only marginally affects investment decisions-a position supported by the regression results of this paper. However, these results also suggest that during the stabilization period, dividend behavior took on a more significant role. Half the regressions that used the slope dummy variables showed that dividend behavior significantly influenced investment spending during the stabilization period. Compliance with the voluntary restraint on dividend payments during the Program was high, and, coming at a time when external financing was increasingly expensive, the added cash flow afforded by the dividend restraint apparently was welcomed by many companies. A similar interpretation is suggested by the regression results for the relative price term. During the stabilization period, the significant coefficients of the relative price dummy variables imply that the behavior of this variable took on greater importance during the controls period than during prior periods. The implication is that by restraining the price of capital investment goods relatively more than wages, the Program positively influenced investment decisions during this period. However, tests to determine whether the primary influence in this relative price term was the relatively greater restraint on capital goods prices or relatively less wage restraint proved indeterminant. Conclusion An examination of the regression results in this paper suggests that the Stabilization Program-by initially strengthening business and consumer confidence-appears to have stimulated investment spending during the controls period. The Program initially appeared to have provided a solution to the problem of inflation, and throughout Phase II, the economy experienced strong real growth along with low rates of inflation. During this early period, it is probable that investment commitments were made extending through 1973 and beyond, so that the subsequent disenchantment by businessmen with the Program was not reflected until the period beyond the Program's expiration in April, 1974. The results of the regressions that examined the individual variables that the Program sought to influence, on the other hand, suggest that the overall stimulative impact of the stabilization period on investment spending was not due primarily to measures directed at influencing these variables. While dividend restraints and controls on relative prices appear to have marginally contributed to increased investment spending, the mixed results suggest that the individual controls measures were not a primary influence on investment spending patterns during this period. Again, this result is not surprising. Economic theory clearly implies that business investment is a relatively long-term process and is not much influenced by short-term action, such as the controls program. This "short-run" aspect of the Program was frequently reiterated by the Administration and reinforced by the constantly changing phases and regulations. Perhaps the influence of such controls measures would have been greater had they been a single, constant set of regulations operating throughout the entire stabilization period and/or had they been perceived as permanent. While the evidence suggests that the Economic Stabilization Program, to some degree, positively influenced investment spending, it |