Fiscal policy will matter

Challenge, Jan-Feb, 1998 by Wynne Godley, George McCarthy

Our simulations have one feature that invariably stands out. The real fiscal stance now projected by the CBO for the next five years, though slightly more expansionary than in the recent past, is insufficient to generate an acceptable rate of growth in the medium term. Given that no impetus is likely to come (looking to the medium term) from further credit expansion and that trade performance may do no better than tag along, the official projections of federal expenditure and revenue seem to be far too restrictive. Taking the next five years as a whole and assuming no change in policy; we would expect a cumulative shortfall of real GDP of at least 6 percent between now and 2002 compared with what will be warranted by the growth of productive potential. If there were another debt deflation, the shortfall could be much larger. Either of these outcomes would inevitably lead to an increase of several percentage points in the unemployment rate.

The main conclusion is obvious. If growth is to continue at an acceptable rate over the next five years, the CBO's existing plans for federal expenditure and taxation will have to be drastically overhauled; either expenditure must rise more, or taxes must be cut.

Notes

1. To express this mathematically; if G = T, where G is government expenditure and T is taxes, and if T = t(GDP) where t is the average tax rate, then GDP = G/t.

2. That is, if G X = T M, where X and M are exports and imports, respectively; and defining rn as the propensity to import (M/GDP), the AFS is given by (G x)/(t m). Equality between the AFS and GDP does not guarantee that stocks of government debt and net overseas assets are individually stationary. There can, for instance, be growing "twin" debts. This is what appears to be happening at present in the United States.

3. R.M. Solow, "Comment: Keynes and the Management of Real National Income and Expenditure," p. 165.

4. The fact that GDP follows the AFS means that, in a growing system, the latter series normally exceeds the former on a simultaneous basis. An excess implies that the budget is in deficit. Indeed, the need to create financial assets is one way of explaining why the budget must normally be in deficit.

5. The figures and all simulations reported in this article were prepared using the MODLER software produced by Alphametrics.

6. The root mean squared error is a measure of how closely predictions track actual values of the variable being predicted. It is the average size of the discrepancy between "predicted" GDP and actual GDP in this one equation model.

7. Solow, "Comment," p. 172.

8. In the short term, however, there have been violent fluctuations in private-loan expenditure, as the figures indicate.

9. We do not for a moment suggest that any level whatever of output can be achieved just by raising the fiscal stance. Any attempt to raise aggregate demand above capacity will fail to raise output for many well-known (and obvious) reasons.

10. In making this comparison we used annual data because the quarterly data are so erratic. The estimate of this change is sensitive to the precise way it is carried out, but no reasonable alternative would undermine the story we tell.