Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)

First Advisor

Thomas R. Beard

Second Advisor

W. Douglas McMillin


This dissertation examines the impact of the anticipated and unanticipated components of government debt on real output when Barro's equations for debt growth and money growth are used as the starting points in specifying anticipated debt and money growth equations using an atheoretical statistical technique. The unanticipated debt and money variables are defined as the residual of the anticipated debt and money growth equations, respectively. Output equations are specified using the same criteria that are used to specify the anticipated debt and money growth equations, and Barro's equation for real GNP is used as the starting point in the specification. As a result of Dickey-Fuller tests, these equations are specified in first-difference form instead of log-functional form. It was found that: (1) both anticipated and unanticipated debt matter when debt is measured at market value; (2) anticipated and unanticipated money growth has short-term but not long-term effects; and (3) real federal purchases matter. Moreover, the sign of the coefficient of the anticipated and unanticipated debt variables is negative instead of positive as is expected from conventional macroeconomic theory. Two plausible explanations for the contractionary impact of the debt on real output are: (1) a negative wealth effect and (2) that the discounted value of government bonds held domestically has become less than the discounted value of the tax liability associated with the debt as a larger percentage of the public debt has become foreign held. The results of this study do not resolve the controversy over the validity of the structural neutrality assumption, but help to verify the existence of a negative wealth effect.