Budget Deficits, Current Account Deficits and Interest Rates: The Systematic Evidence on Ricardian Equivalence

The recent reemergence of large U.S. government budget deficits has rekindled the debate as to whether deficits adversely impact real interest rates. The conventional “crowding out” hypothesis predicts that there would be such an adverse effect; the Ricardian “deficit neutrality” hypotheses predict no adverse impact. However, in a fully open economy, budget deficits affect the economy through the current account rather than the interest rate channel. Under the conventional view, current account deficits react one-for-one to budget deficits; under the Ricardian view, neither real interest rates nor current account deficits would react. In this paper, we examine the systematic evidence on the impact of budget deficits on both real interest rates and current account deficits. We test the conventional “crowding out” hypotheses against the Ricardian “deficit neutrality” alternative. Because our tests are structured on both interest rates and current account deficits, they have validity regardless of the extent of the openness of the U.S. economy over the time period of the analysis. We structure a series of tests that explore the impacts of both current and expected future budget deficits on (1) both interest rates and current account deficits; (2) across a variety of specification issues including data non-stationarity, lag structure, auxiliary variables, term structure, and functional form; and (3) over the period 1976 through 2002 which contains spectacular “natural experiment” variation in the data that should allow any links between budget deficits and interest rates or current account deficits to be easily detected. Our results are systematic and dramatic. Over the full structure of the specification space, there is no evidence of any positive effects of either current or expected future budget deficits on either real interest rates or current account deficits. Because of the large variation in the data generated by the natural experiment, economically significant effects can be clearly rejected. Moreover, when viewed over the full structure of the specification space, there appears to be a relatively significant and systematic negative effect of budget deficits on current account deficits (and sometimes but not systematically on interest rates). This latter effect is consistent with augmentations to an underlying Ricardian framework based on either uncertainty effects or intertemporal deadweight loss effects.