CRS FINDS UNCERTAIN ECONOMIC RESPONSE TO INVESTMENT CREDIT.

                          Brian W. Cashell

SUMMARY

Following the recent presidential election, the economic consulting firm of Laurence Meyer and Associates produced an analysis of President-elect Clinton's economic program using a large scale macroeconomic model, including a separate assessment of the effects of an incremental investment tax credit. A subsequent study by the WEFA Group examined the consequences of a regular (nonincremental) investment credit. The Congressional Research Service has also simulated the effects of both an incremental and a nonincremental credit using the Data Resources, Inc. (DRI) macroeconomic model of the economy.

The simulation results presented here show different responses to the enactment of an investment tax credit. None of the models suggests that there would be a large short-run effect on the economy. These simulations suggest that there would be some modest gains in employment but not until the second year after the credit was enacted. But, especially in the case of a nonincremental credit, there would likely be a substantial increase in the budget deficit. Furthermore, much of any increase in investment is likely to be financed by foreign lenders, and a good portion of the investment goods would be imported from abroad.

INVESTMENT TAX CREDIT: USING MACROECONOMIC MODELS TO ASSESS SHORT RUN EFFECTS

Interest has developed recently in restoring the investment credit to stimulate the economy. In particular, attention has been focused on the incremental investment credit, which would apply to investment in excess of a base. There are many issues in evaluating an investment credit and an incremental credit; this study focuses, however, on the short run effects of the credit and its role in stimulating aggregate demand. Further, it discusses the use of large scale macroeconomic models to assess these short run effects.

Following the recent presidential election, the economic consulting firm of Laurence Meyer and Associates produced an analysis of President-elect Clinton's economic program using a large scale macroeconomic model. ¹ This study included a separate analysis of an incremental investment tax credit. The credit would be allowed at a ten-percent rate for equipment, but would be restricted to investment in excess of 80 percent of 1992 investment. This study suggested that the credit would increase output enough to pay for itself in the first year.

A subsequent study by the WEFA Group examined the consequences of a regular (nonincremental) investment credit. ² They did not report results for an incremental credit, but did indicate that an incremental credit would result in smaller effects.

The Congressional Research Service has also simulated the effects of both an incremental and a nonincremental credit using the Data Resources, Inc. (DRI) macroeconomic model of the economy.

These simulations assume that the credit is permanent, that there is a fixed original base, and that the credit is allowed against the alternative minimum tax (AMT). Presumably, the response to a temporary credit would be larger than the response to a permanent one, while not allowing the credit against the AMT would reduce the overall size of the credit substantially.

WHAT ARE MACROECONOMIC MODELS?

Macroeconomic models are both very sophisticated and yet, in many cases, very simplified. These are models that attempt to forecast the short run path of the economy and to examine the consequences of various fiscal and monetary policies. They allow for unemployed resources and generally find that stimulative policies, such as an investment credit, can have significant effects on employment and output, even though it is generally recognized that fiscal and monetary policies are unlikely to produce additional jobs in the long run. ³

They do not assess many of the important aspects of tax policies such as their long run allocative and distributional effects. ⁴ Rather they focus on the short run effects of policies in altering aggregate demand and employment.

Although these macroeconomic models have many equations and considerable sectoral detail, the estimation of relationships in the model is often based on relatively simple econometric techniques. Moreover, there are cases where the statistical evidence is quite uncertain as to the behavioral response to a change; while the models must incorporate a specific response, the true response may, in fact, be much larger or smaller.

The response of investment to a change in "price" variables such as interest rates and investment credits is almost certainly a case where considerable uncertainty exists. Empirical evidence has not revealed a clear relationship; indeed, in writing about the DRI model in a book published in 1983, Otto Eckstein, founder of DRI, acknowledged that the investment equation was not very "robust" -- that is, that the behavioral parameters tended to change with changes in specification. ⁵

There are several important issues that link an investment credit to job creation in the short run. One important issue is how much a given credit stimulates increased investment spending by firms. If the response is very small, then the credit will have a small effect on employment. Moreover, there is a possibility of delay in the impact of the credit because of planning lags by firms. Some economists concluded that, due to the response lags, the investment credit was not very effective as a device to stimulate the economy. ⁶

An investment credit can increase investment directly through not only the "price" effect, or incentive for marginal investment, but also through cash flow effects. Thus, the responses of the models to a credit, and to different types of credits, depend not only on the price response but on the power of cash flow to alter investment. The relative importance and magnitude of these two effects vary from model to model.

If a credit is translated by businesses into equipment spending, there is then a multiplier effect, as the income received by the factors of production is spent. Because substantial portions of equipment and of other goods are imported, some of the fiscal stimulus is offset and the multiplier (increase in aggregate output divided by the initial increase in spending) tends to be relatively small. Furthermore, some of any increase in investment will, in the absence of higher U.S. saving, be financed by capital flows from abroad (as interest rates are driven up). Thus, some of the increase in investment spending will be financed with foreign capital, and foreigners will have a claim to some portion of the increase in income of the larger capital stock. This inflow of capital causes appreciation of the international value of the dollar, which ultimately depresses imports relative to exports and undercuts the stimulus from the credit.

RESULTS FROM THE MACROECONOMIC MODELS

Using the DRI model, the Congressional Research Service examined both an incremental and a nonincremental tax credit on purchases of business equipment.

In the first simulation (DRI-1), a 10-percent tax credit is assumed to take effect at the beginning of 1993 on all equipment purchases in excess of 80 percent of 1992 levels. This incremental credit is less likely to benefit investment spending that would have been made even in the absence of a tax credit.

In the first year of the credit, the most pronounced effects are on interest rates and the budget deficit. Not until the second year are there significant increases in economic growth and in employment.

Because of the rise in interest rates, U.S. investments are more attractive to foreigners, and there is an increase in inflows of foreign capital. In fact, it is likely that a significant share of the increase in domestic investment would be foreign owned. An increase in demand for dollars as a result of these capital inflows tends to drive up the exchange value of the dollar. This makes imports cheaper, exports more expensive, and widens the trade deficit. More than half of the increase in domestic investment is financed by increased foreign investment.

In the second simulation (DRI-2), the tax credit benefits all equipment purchases. The key difference between this simulation and the first one is that the budget deficit rises by more because the credit applies to a larger proportion of investment spending. As was the case in the first simulation, investment spending and economic growth pick up, but not until the second year. Again, more than half of the increase in domestic investment was financed by increased inflows of foreign capital.

The results of the simulations using the DRI model can be compared to the analysis by the WEFA group and the study by Lawrence Meyer and Associates. Recall that the Meyer study considered an incremental credit, as did the first DRI simulation (DRI-1). WEFA considered a nonincremental credit, as did the second DRI simulation (DRI-2). (In both cases, a substantial amount of the increased investment was financed by imported capital: about half in the Meyer model and more than half in the WEFA simulation.)

Two basic types of percentage changes from the simulations are reported: the change in investment in equipment and the change in employment. In all cases the reported changes are relative to a base case which differs only in that there is no investment tax credit.

Table 1 shows the percentage change in real investment in equipment. This table provides some notion of how the subsidy affects investment, although after a period of time investment also grows because of increases in aggregate demand.

     TABLE 1: ITC -- Percentage Increase in Equipment Spending /*/

The predicted response to the investments credits varies substantially across models and years. The Meyer model has a very large response initially relative to the other models. The DRI model appears to be more sensitive in the longer run than the other models; indeed, its response in the long run would be considered to be at the high end of the range of empirical estimates. WEFA occupies a more central position in the short run and has the smallest second-year effect. In the third year (reported only for WEFA and DRI), the DRI response is over twice that of WEFA.

The effects of the regular and incremental credit are very similar in the DRI model, which indicates that cash flow does not play a very important role. If cash flow were more important, the incremental credit would have a smaller effect than the regular credit. For example, although WEFA did not report the effects of an incremental credit, their discussion indicated that they would find a smaller effect from the incremental credit because of the cash flow effect.

Table 2 shows the change in employment due to the imposition of an investment credit. In general, the percentage changes in employment are larger with larger percentage changes in equipment spending. In 1993, the incremental investment credit in the Meyer model causes employment to increase by two-tenths of a percent. WEFA reports a similar first year result for a much more costly nonincremental credit. The increase is, however, negligible for the DRI model in each case. In the second and following years, the credit appears to increase employment by about one half of one percent. (A tenth of a percentage point is slightly over 100,000 jobs.)

Note that despite the much larger equipment spending response in DRI in year 3, the effect on employment is only slightly larger than in the WEFA study. It is also important to recognize that economic theory would not suggest that a permanent long run effect on employment would be obtained from the investment credit or, indeed, any fiscal policy. Thus, although the short run models may predict employment effects, these effects become more doubtful as time goes on.

           TABLE 2: Investment Credit -- Percentage Increase

    TABLE 3: Investment Credit -- Increase in Deficit ($ billions)

Effects on the deficit also vary with the model. The Meyer model shows a small effect in the first year, of about $1 billion for the incremental credit, compared to a cost of $9.5 billion in the DRI simulation. The WEFA model shows a much smaller cost for the nonincremental credit than DRI-2 in the first year.

These results reflect, in part, the differences in level of economic activity and their feedback effects. They also reflect different assumptions regarding the timing of tax receipts. In some models, there is a long lag on the collection of receipts. This long lag apparently accounts for [a] striking difference in revenue cost in the first year between the WEFA and the DRI-2 simulation. It is not clear that such a lag is appropriate; indeed, our conversations with WEFA indicate that they may modify their receipts equation to have a quicker loss of revenue, which would be consistent with the frequency of estimated tax payments by corporations.

There is apparently a delay in the effect on receipts in the Meyer model and that, along with the more rapid response to the credit, accounts in part for the differences between effects in the two models in the first year. It is the combination of the long lag on receipts and the stronger economic response in the Meyer model that lead to the finding that the incremental credit virtually pays for itself in the first year.

CONCLUSIONS

The model simulations show quite different economic consequences for the credits, whether judged by effects on investment, employment, or the deficit. As noted earlier, there is substantial uncertainty about the economy's response to a credit. None of the models, however, appears to find a large short-run effect on the economy. Moreover, the small effect on the deficit found in the Meyer model reflects not only an investment response that is larger than the other models, but also the lag in receipts which, even if correct, is largely a timing effect.

² Economic Policy Simulations for the Clinton-Gore Economics Transition Team, November 25, 1992, by Robert F. Wescott and Kurt E. Karl, The WEFA Group, U.S. Macro Special Study.

³ For a discussion of this point see "Is Job Creation a Meaningful Policy Justification?," by Jane G. Gravelle, Donald W. Kiefer, and Dennis Zimmerman, Congressional Research Service Report 92-697 E, September 8, 1992.

⁴ For a discussion of some of these issues see Jane G. Gravelle, "Tax Subsidies for Investment: Issues and Proposals," Congressional Research Service Report 92-205 S, February 21, 1992.

⁵ Otto Eckstein. The DRI Model of the U.S. Economy. New York: McGraw-Hill, 1983, p. 140.

⁶ See Jane G. Gravelle, "Tax Subsidies for Investment: Issues and Proposals," Congressional Research Service Report 92-205 S, February 21, 1992 for a survey of the literature.