Public Debt
Default risk

Many governments that are now seeking to reduce their debt/GDP ratios will require enormous budget surpluses to achieve this stabilization. In Discussion Paper No. 787, Francesco Drudi and Research Affiliate Alessandro Prati develop a model of costs of default on government debt. The public is uncertain about whether the government will favour the `poor' by levying a `surprise' tax on the debt or the `rich' by following the `sustainable' policy of repaying the debt in full.

Drudi and Prati show that the public's uncertainty about the government's type may or may not be resolved depending on the two types' relative preferences and initial reputation. Once a lasting surplus is achieved, however, a `defaulting' government has no incentive to mimic the `sustainable' policy, so a risk premium arises when it is on the verge of achieving such a surplus. If its initial reputation is `bad' or its total expenditure large, however, this premium may become `too high' for any debt to be sold; the public anticipates that either type of government will levy a surprise tax.

Drudi and Prati then relate their results to the Maastricht Treaty conditions to show that the deficit condition will distinguish `sustainable' from `defaulting' governments, but the debt threshold may be `too tight'. Under certain conditions it will exclude some governments that have `bad' initial reputations or face high interest rates, while ignoring their commitment to `sustainable' policies; it may also admit some with `good' reputations whose policies will in fact lead to default. Drawing policy implications requires caution, however: excluding countries that may default is not the only rationale for the Maastricht rules; the assumption that no risk premium is needed once the `sustainable' government signals its type may not apply if governments change frequently; and they can also meet the Maastricht rules by cutting expenditure instead of increasing taxes.

Signalling Debt Sustainability
Francesco Drudi and Alessandro Prati

Discussion Paper No. 787, May 1993 (IM)