Budget Deficits and Real Interest Rates: Updated Empirical Evidence on Causality
Cebula, Richard, Atlantic Economic Journal
In the first part of this issue, the focus was on Simon's rules for setting public policy. One of these rules advocates a balanced federal budget. This rule arose from Simon's concern with the ever present budget deficits, in every year from 1961-74 (except for 1969) and projected deficits in the coming decades. Simon's assessment of the deficit outlook turned out to be correct. Except for a brief experience with federal government budget surpluses in 1998-2001, budget deficits loom once more on the horizon.
The sluggish U.S. economy following the 2001 recession, the 2001 multi-year federal income tax rate cut, and spending increases to meet the cost of the two wars (war on terrorism and war with Iraq) make the prospects for budget deficits well into the foreseeable future a veritable certainty. William Simon would in principle have been most concerned with this deficit outlook and its implications for productivity and economic growth over time. Simon's concern for the deficit reflects his belief that deficits raise the cost of capital in the private economy. Budget deficits increase the rate of interest, absorb private saving, thus adversely affecting capital accumulation and productivity [see Orcutt paper in this issue].
The link between the federal budget deficits and interest rates in the U.S. has been investigated extensively by a number of researchers, including Barth et al. [1984, 1985]; Barth et al. ; Carlson and Spencer ; Cebula [1988; 1991; 1997; 2000]; Cebula and Belton ; Cukierman and Meltzer ; Darrat ; Evans [1985, 1987]; Feldstein and Eckstein ; Findlay ; Hoelscher [1983, 1986]; Holloway ; Johnson ; Mascaro and Meltzer ; McMillin ; Ostrosky ; Saltz [1998, 1999]; Swamy et al. ; Tanzi ; and Zahid . Most of these empirical studies are couched within open or closed IS-LM or loanable funds models or variants of same. Many of these studies find that the budget deficit raises longer term rates of interest (1) while not significantly affecting short term rates of interest. Because capital formation is presumably much more affected by long term than by short term interest rates, the inference has occasionally been made that these deficits may lead to partial crowding out [Carlson and Spencer, 1975; Cebula, 1985].
This literature for the most part neglects net international capital flows, thereby ignoring the potential direct or indirect interest-rate effects of these flows and thus raising a question of a possible omitted-variable bias [Penner, 1987]. Moreover, this literature fails to incorporate the effect of personal income tax rates, thereby raising the question of a possible omitted-variable bias on yet another level, involving private sector spending and savings decisions (including the choice between taxable and tax free bonds), federal government tax collections, and federal spending [Cebula and Belton, 1993; Tanzi, 1985].
More significant, perhaps, is the fact that the deficit measure adopted most commonly in the empirical analysis is either the N.I.P.A. deficit, the structural deficit, and or the cyclical budget deficit, all of which include interest payments on the national debt. This raises the question of a possible model mis-specification. As the interest rate is the dependent variable, and interest payments on the national debt enter as the right-hand-side variable, the budget deficit itself, with causality assured to run uni-directionally from the budget deficit to the interest rate. The present study avoids this pitfall by adopting the primary budget deficit as the deficit measure [Cebula and Rhodd, 1993; Ott, 1993]. Since changes in ex ante real long term interest rates may either directly or indirectly alter the growth rate of GDP through changes in real private sector purchases, especially capital formation, tax revenues, government transfer payments, and the government budget deficit may also be affected. …