The Determinants of Business Fixed Investment: a Comparison of Models and Policy Changes.

Abstract

This study investigates whether there are tax policy changes which could increase the proportion of business fixed investment (BFI) in real GNP to the 12 1/2 to 13 percent range. This is the proportion estimated by the CEA which is necessary in order to have substantive positive effects on the economy. Five models of BFI are compared according to historical fit of the equations, forecasting accuracy, and structural stability. These tests are needed because there is no generally accepted model of investment. The models studied are: (1) accelerator, (2) cash flow with accelerator term, (3) neoclassical, (4) securities value, and (5) Data Resources, Inc. Several of the models perform well under the first two criteria, especially the cash flow with accelerator term model, but none of them show structural stability for both equipment and structures investment. This creates serious doubts about the ability of any of these models to forecast policy changes accurately. The instability is probably due to misspecification brought about by omitted elevant explanatory variables. Three tax policy changes are simulated from 1981 to 1989: (1) a cut in the corporate income tax rate from 46 to 40 percent, (2) an increase in the investment tax credit (ITC) on equipment from 10 to 20 percent, and (3) a five year phasing-in of the 10-5-3 proposal for accelerated depreciation. On the basis of most dollars of investment per dollar of Treasury tax revenue loss, the investment tax credit appears to be the most efficient method of increasing investment, with 10-5-3 the least efficient because of its large revenue loss in future years. The ITC also results in the largest BFI-GNP ratio, although it appears that tax cuts somewhat larger than those simulated here, in addition to a conducive climate for investment, are necessary to raise the BFI-GNP ratio to the 12 1/2 to 13 percent range.