Debt bailouts and constitutions

Economic Inquiry, July, 2008 by Emanuel Kohlscheen

I. INTRODUCTION

A number of countries have undergone several rounds of bailouts of subnational government debts in the last decade. The Brazilian government, for instance, assumed the debts of the federated states in 1989, 1993, and 1997. Also in its southern neighbor, Argentina, the line that separates provincial and federal budgets has become blurred a number of times. Seven Argentine provinces were granted a debt bailout between 1992 and 1994, and the central government took over deficit-ridden public pension funds of 11 provinces between 1994 and 1996. Bailout operations also occurred in 1995 and 2001. Such operations are not always explicit, however, and in some instances, hyperinflation may have been the ultimate bailout that eroded debt stocks. (1,2) While bailouts of subnational entities could certainly be efficient ex post, the above-mentioned recurrence of episodes could eventually undermine efforts of the center to achieve fiscal discipline.

In an attempt to strengthen the credibility of budget separations between the different tiers of government, several countries have changed the institutional setting for subnational borrowing in the last years. After the financial meltdown at the end of 2001, Argentina's Congress approved a law containing a commitment to the creation of a federal fiscal body and coordination mechanisms for provincial indebtness. (3) Brazil and Mexico have enacted legislation containing explicit no-bailout provisions. For instance, Brazil's Lei de Responsabilidade Fiscal, enacted in 2000, precludes any further credit operation between units of the federation. Brazilian states are now required to submit new bond issuances to the sequential approval of the Ministry of Finance and the Senate. Golden rule limits of indebtness for states and municipalities were also defined. (4)

This paper explores the interrelation between bailouts of subnational governments and Constitutional tax revenue sharing arrangements. It argues that revenue sharing mechanisms, which are typically engraved into Constitutions, might change incentives for demand-driven bailouts and possibly widen political support for bailout to units that a priori may have little to gain from the direct transfer of debt. A subnational debt bailout implies that, unless the federal government has the flexibility and willingness to cut back on its expenditures to fully absorb the cost, taxation is shifted from the state to the national level. When the Constitution mandates that a fraction of federal revenues be automatically distributed to the states, federal revenues must be increased by more than the stock of debts shifted to the Union. These excess revenues accrue to member states according to the formula set in the Constitution acting as side payments conditioned on a bailout being approved. As transfers are a direct function of federal revenues, states with low debts--that would naturally oppose a shift of the repayment burden of subnational sovereign debt to the central government--might not do so, as this shift ultimately increases their source of income. Hence, in the presence of federal revenue sharing, a debt/gross domestic product (GDP) distribution that is skewed to the right is no longer a sine qua non condition for a bailout to be supported by a majority of states. The reason a bailout occurs then is not driven by an externality arising from financial market interdependencies, as in the model of Inman (2003), but from the fact that the failure to bail out indebted states generates a negative externality on states that are net recipients of the revenue sharing arrangement. Politicians of remote states that have constituencies which rely heavily on transfers of a predetermined share of federal revenues will probably not oppose measures that ultimately "increase the size of the pot." Therefore, policies and institutional arrangements aimed at reducing regional income disparities should be carefully designed so as to not soften perceived budget constraints.

The implications of the model go well beyond Latin American federations. Rodden (2003a) gives an account of the failed attempt of a group of U.S. states to shift its debts to the central government in the early 1840s. He concludes that "one of the best explanations for the defeat of the assumption movement may simply be in the numbers--the majority of states did not have large debts, and outside of Maryland and Pennsylvania, most of the debtor states had small populations." With no Constitutional revenue sharing mechanism in place, the interested parties may have found themselves unable to set up the sizable compensations to less indebted states that would have been needed to make the proposal politically feasible (for a detailed discussion of the episode, see Wibbels 2003). The issue should also be of interest to European policymakers seeking to set up an institutional arrangement that makes the no-bailout provision in the Constitution of the European Central Bank time consistent.