Introduction

With a few notable exceptions, such as Fischer (1983), Peled (1985) and Bohn (1988a, b, c), the literature on public debt has concentrated on positive issues, such as the neutrality of the debt (Barro, 1974; Tobin, 1971). In this paper I want to concentrate instead on welfare issues, in particular, the impact of debt policy on the efficiency of risk sharing.

As a prelude to the central part of the paper, Section 2 reviews the familiar issue of the neutrality of the debt. The classical Ricardian equivalence theorem assumes that markets are complete. Nonetheless, even if markets are incomplete, there is an analogue of the classical neutrality theorem. A theorem of this sort is proved in Section 2 for a generic economy with incomplete markets. It shows that changes in the size and composition of the debt are neutral as long as the set of debt instruments issued by the government is unchanged. This result is similar to the Modigliani-Miller theorem of Wallace-Chamley-Polemarchakis (see Wallace, 1981, and Chamley and Polemarchakis, 1984). On the other hand, if markets are incomplete, it is clearly possible for the government to have an impact on the economy by introducing new securities that expand risk-sharing opportunities.

There is a tension between these two results. It seems that a tiny amount of a new security has a large impact while a large change in the amount of an existing security has no impact at all.