Comparing Alternative Methods of Adjusting U.S. Federal Fiscal Deficits for Cyclical and Price Effects

In this working paper, Resident Scholar Neil H. Buchanan statistically tests six alternative definitions of the federal budget deficit to determine if these definitions improve the results of econometric studies that use the deficit as an exogenous variable. Buchanan wishes to 1) evaluate how well Robert Eisner's conclusion that a price- adjusted deficit definition improves econometric results, and 2) compare alternative measures of the deficit. Buchanan's analysis begins with two definitions of the structural deficit published by the government: the Bureau of Economic Analysis's high-employment deficit, which is based on a constant standard of unemployment, and the Congressional Budget Office's standardized employment deficit, which is based on a varying standard of unemployment, namely, the NAIRU. He then compares two sets of price-adjusted structural deficit measures to the set of original definitions. Each original definition is adjusted by using two different calculations of a price effect intended to gauge the change in the value of outstanding government debt. One set of price-adjusted measures is obtained by subtracting the product of the year-end par value of outstanding debt and the annual rate of inflation from each of the original measures of structural deficits. The second price-adjusted set third set is obtained by subtracting a more complex derivation of a price effect, namely one that accounts for the timing of inflation's effect on prior debt. The results of initial regression analysis indicate that the method of adjusting for price effects is more important than the method of adjusting for cyclical effects. Using a variety of specifications, time periods, and data definitions, Buchanan's findings did not support the case that deficit spending stimulates GDP growth. However, a relationship was found between unemployment and the deficit, even when the non-price-adjusted measures of the deficit were used. The results of regressions using shorter, 15-year periods show a decline in the importance of price adjusting and a further weakening of the growth-deficit relationship compared to what was found in the initial regressions. The unemployment- deficit relationship, however, was stronger than in the full- period regressions