بدهی عمومی بالا در بحران ارز: اصول در مقابل اثرات سیگنالینگ

This paper examines how public debt, government credibility and external circumstances affect the probability of currency devaluations in a three-period, open-economy version of the Barro and Gordon (J. Polit. Economy 91 (1983) 589) model. Nominal public debt links current to future policy actions: resisting a crisis may either enhance or undermine the sustainability of the exchange rate regime depending on whether what matters is the government’s reputation or fundamentals, i.e. the level of the debt. The focus is on the impact of public debt, debt maturity and the government’s credibility on the expected devaluation in the current and future periods. We can thus identify the factors that affect the short-term interest rate and the forward rate, hence derive predictions about the level and the slope of the yield curve.

The credibility of a fixed exchange rate regime depends on the government’s reputation (Backus and Driffil, 1985a and Backus and Driffil, 1985b) and on economic fundamentals (Dornbusch, 1986). A strenuous defense of the parity can signal the government’s commitment and thus enhance credibility, but it also worsens fundamentals, making the economy more vulnerable to adverse shocks. Drazen and Masson (1994) show that resisting a currency crisis can actually undermine rather than enhance the credibility of the exchange rate regime. Evidence from the EMS does indeed include episodes in which the interest rate rose (relative to the German rate) after a realignment and others in which it fell below the pre-crisis level. A rise in the interest rate is more commonly found in the early phase of the EMS and after unilateral realignments. Importantly, the interest rate fell in a number of realignments involving high-debt countries, such as Ireland and Italy, while in the same period it rose after French realignments. In this paper we inquire into the way public debt, debt maturity and the government’s credibility affect the probability of a devaluation in a model in which defense of the exchange rate worsens the fundamentals by increasing public debt. We take a three-period stochastic version of the Barro and Gordon (1983) model, where the decision to devalue or maintain the parity depends on the realization of an output shock. Because of uncertainty, a devaluation leads to unexpected inflation, which increases output, both through a standard price-output effect and through the reduction of the distortionary taxes associated with nominal debt service. We show that whether the exchange rate regime gains or loses credibility after a successful defense is uncertain. On the one hand, resisting a crisis enhances the credibility of the government and thus the expectation that the parity will be maintained. This effect is important when there is substantial uncertainty about the government’s cost of devaluation and when the level of debt is low. We term this the “signaling” effect. On the other hand, defending the parity and refraining from inflationary financing increases the debt burden, hence the likelihood of a forced future devaluation. This effect is important when the debt is large and there is little uncertainty about the government’s cost of devaluation. We call it the “debt-burden” effect. Which one prevails depends on the relative importance of the government’s reputation and the fundamentals. When the government’s preferences are publicly known, only the fundamentals matter; a devaluation always increases the probability that a future defense of the new parity will succeed. In this case the probability of a first-period devaluation increases with the size of the public debt and, more interestingly, with the share of it that is short-term. That short maturity impairs the government’s ability to defend the exchange rate is a familiar result in the literature on self-fulfilling crises (e.g. Giavazzi and Pagano, 1990, Obstfeld, 1994 and Obstfeld, 1996). We show that the argument for longer-term debt need not rely on intrinsic uncertainty: a large amount of maturing debt creates an incentive to devalue in order to roll over such debt at lower-than-expected interest rates. When the government’s preferences are not publicly known, the decision to devalue may reveal a weak government heading to further devaluation, which generates inflationary expectations and higher interest rates. This creates an incentive to defend the fixed parity both for a tough government, to signal its type, and for a weak one in order to appear tough. Thus, reputation motives increase the probability that the parity will be maintained. Our analysis offers a theory of the term structure of interest rates for high-debt countries vulnerable to currency crises, which is possible because, unlike in Drazen and Masson (1994), we derive the probability that the market assigns to a devaluation for both the current and the future period. The probability of a first-period devaluation decreases with the uncertainty about government preferences—i.e. with the difference between the devaluation costs of the two types of government—for any given expected cost. By contrast, the probability of a second-period devaluation does not depend on reputational considerations. Debt-burden and signaling effects are thus crucial in determining the slope of the yield curve. Reputational incentives lower the short-term interest rate but have no impact on the forward rate, which depends only on fundamentals. For any given level of debt, greater uncertainty about government preferences implies lower interest rates and a steeper yield curve. The effect of debt accumulation is not monotonic. When reputation is not important, the debt-burden effect prevails and larger debt is associated with higher interest rates and a flatter (or more downward-sloping) yield curve. In contrast, when reputational considerations outweigh fundamentals, a debt increase may reinforce the signaling effect of a defense of the parity and thus lead to a steeper yield curve. The paper is organized as follows. Section 2 presents evidence from EMS realignments. In Section 3, attention is restricted to the case in which the government’s preferences are publicly known. In Section 4 we introduce asymmetric information, and the credibility of the exchange rate regime is shown to depend on both fundamentals and signaling effects. Section 5 examines the effects of debt and debt maturity on the expected devaluation and derives implications for the term structure of interest rates. Section 6 concludes.

We have studied the role of fundamentals and signaling effects in currency crises in a model where a devaluation may increase rather than reduce the probability that the new parity will be maintained. Tight monetary policy to defend the currency worsens the fundamentals, thus impairing the ability to withstand future crises. In Drazen and Masson (1994) the persistence of unemployment provides a channel whereby current actions affect future policy. The present paper suggests public debt and debt maturity as additional factors. A devaluation may improve the prospects for the exchange rate regime by reducing the value of the public debt, but it may also lead to the expectations of further devaluations by signaling weak commitment to price stability. Examining the behavior of interest-rate differentials relative to Germany during EMS realignments in France, Ireland and Italy, we find instances in which the new parity was perceived as sustainable and others that led to a deterioration in market confidence in the EMS. This evidence suggests that both the debt-burden and the signaling effects play a role. We offer a first formalization of how these aspects interact in determining the expected devaluation in the current and future periods. The analysis offers a theory of the term structure of interest rates for high-debt countries vulnerable to currency crises. When improving reputation is important, the incentive to resist a crisis is strong and should decrease the probability of a current devaluation, thus leading to lower short-term interest rates. On the other hand, the prospects for the exchange regime, and hence, the forward rate depend only on fundamentals; i.e. the level of public debt. Debt-burden and signaling effects are thus important in determining the level and slope of the yield curve. For any given level of debt, greater uncertainty about the government’s cost of devaluation implies lower interest rates and a steeper yield curve. At low levels of debt the yield curve may also steepen because of debt accumulation, since a larger debt amplifies the signaling effect of a defense of the exchange rate. In general, however, debt accumulation produces higher interest rates and a flatter yield curve. These are interesting predictions to be tested empirically, a task which is left to future research.