During the last several years, the issue of the effect of budget deficits on interest rates in the United States has attracted considerable attention. However, not only has the issue remained unsettled, but more recently has become even more clouded by conflicting evidence. For example, Cebula (1987) reported a significant and positive effect on ex post short-term real interest rate, Kolluri and Giannaros (1987) a significant and negative effect on ex-ante short-term real rate, Makin (1983) a 'mixed to weak' positive effect using short-term nominal rate and Hoelscher (1983), Motley (1983) and Evans (1985Evans ( , 1987 an insignificant effect, again, using short-term nominal interest rate as the dependent variable.A strict comparison of these studies is not possible because of the differences in the definition of the dependent variable employed, in the selection of the determinants, the data and the estimation procedures used. The aim of this paper is to test whether budget deficits influence interest rates regardless of whether they are nominal or real and if real, whether ex-post or ex-ante and in all cases, whether short-term or long-term, using the same set of data. A by-product of the study is that it sheds light on the 'inverted' Fisher hypothesis recently proposed by Carmichael and Stebbing (1983).
The model and the dataIn order to test the robustness of our results, as far as the dependent variable was concerned, the following equations were estimated: i = a 1 + bli~ + clU + dlDEF + elM + v I (1) i = a 2 + bzi~e + c2U + d2DEF + ezM + v 2 (2) * Thanks are due to Zheng Wei for research assistance.