عدم توازن مالی و پویایی بحران ارز

This paper analyzes links between the fiscal theory of the price level (FTPL) and the first generation models after Krugman (Journal of Money Credit and Banking 11 (1979), 311–325), exploring the idea that a synthesis between the two can become a new framework to analyze the fiscal dimension of currency crises. Working in a simple synthetic framework, we show how external nominal shocks can cause a fiscal imbalance and undermine currency stability, resolve two well-known paradoxes of the first generation model, discuss the role of seigniorage revenues, and illustrate how fiscal and interest rate policies interact to determine the magnitude and the timing of speculative attacks and devaluations.

In his macroeconomic lectures at Yale, Christopher Sims often remarked that the “first generation mode” of currency crises due to Krugman (1979) and the fiscal theory of the price level (FTPL)1 were “close cousins”. In this paper, we explore formally links between the FTPL and the Krugman model.2 We specify a simple, tractable model that we see as a step toward a more general, synthetic model of fiscal imbalances and currency crises. Our analysis yields a number of new results and clarifies some aspects of the literature. Although the Krugman model appeals informally to fiscal considerations, a currency crisis arises in that model exclusively as a consequence of domestic credit creation by the central bank. A strength of the Krugman model was its simplicity, but the model proved difficult to reconcile with evidence from recent currency crises. The observation that the currency crises in East Asia were not preceded by domestic credit expansion was a factor that motivated the first recent extension of the Krugman model. In Corsetti et al. (1999), a currency crisis is caused by anticipated future growth in domestic credit. Soon it became apparent that this extension per se would be insufficient to make the first generation model consistent with the recent crises. Burnside et al., 2001, Burnside et al., 2003a and Burnside et al., 2003b, henceforth BER, documented that while the devaluations in Latin America, East Asia and Turkey were associated with large fiscal imbalances, governments financed only moderate fractions of the imbalances with seigniorage revenues. A decrease in the real value of nominal government liabilities played a quantitatively more important role than seigniorage revenues. In an effort to match the evidence BER extended the first generation model, including in it nominal government liabilities (such as domestic-currency debt and spending commitments not indexed to the exchange rate). Daniel, 1998 and Daniel, 2001a applies more directly than BER two insights from the FTPL in a model of currency crises. Given a fiscal imbalance, the government budget constraint seen as an equilibrium condition implies that the price level increases and the exchange rate depreciates. The FTPL highlights the role of maturity of public debt in macroeconomic dynamics: Long-term nominal debt smooths out the effects of fiscal shocks on the equilibrium price level ( Cochrane, 2001; Woodford, 1998). In Daniel's work long-term nominal debt helps delay devaluation, consistent with the FTPL. The contributions of Corsetti, Pesenti and Roubini, BER and Daniel are sometimes seen as an alternative to the Krugman model. In light of Sims's remark, however, we see them as steps toward a new, synthetic model of fiscal imbalances and currency crises that will combine the insights of Krugman with those of the FTPL. In this paper, we take a further step towards building such a synthetic model. Our starting point is the intuitive experiment of Krugman. We assume a fixed exchange rate and postulate an exogenous shock that decreases the present value of government's real primary surpluses relative to its outstanding liabilities. We then analyze the dynamics of adjustment. Different from the first generation model, we conduct our analysis in an economy with public debt of different denomination and maturity (domestic- and foreign-currency and short- and long-term). The key to the adjustment is that devaluation causes a wealth transfer from holders of nominal public liabilities that finances the fiscal imbalance. Unanticipated devaluation reduces the real value of short-term nominal debt; anticipated future devaluation creates a wealth transfer from holders of long-term nominal debt. New results from our analysis can be summarized as follows: (1) Shocks that cause a collapse of a fixed exchange rate can be nominal, and need not imply a deterioration of government's real primary surpluses. This is because nominal shocks affect the real value of debt and are thus capable per se of causing fiscal strain. Thus a foreign deflationary shock (e.g., in the case of a country pegging to the euro, an appreciation of the euro relative to the dollar) can jeopardize the sustainability of a currency peg independently of changes in relative goods prices or “competitiveness” problems. Consider prospective entrants to the EMU, expected to keep their currencies stable vis-à-vis the euro prior to entry. Our model makes it clear that a stable path of real primary surpluses, and their sufficient reaction to external real shocks, will not be enough for the pegs to be viable. What is required in addition is the ability to adjust real primary surpluses in reaction to foreign nominal shocks. (2) The literature explains the timing of a currency crisis via a lower bound on exchange reserves (which corresponds to an upper bound on public borrowing). We show that the timing is also uniquely pinned down by a policy rule specifying the behavior of the short-term nominal interest rate. We see this shift of emphasis as realistic, since policymakers view themselves nowadays primarily as setting the short-term nominal interest rate. Moreover, the new focus helps avoiding two well-known paradoxes of the first generation model: The currency appreciates in that model if the exchange rate parity is abandoned “too early”, despite the fiscal crisis, and there can be trade at off-equilibrium prices. Our specification with an interest rate rule avoids any such odd implications. (3) The cause of a currency crisis is fiscal, with expansion of domestic credit and seigniorage revenues being a sideshow. To illustrate this most clearly, our model abstracts from base money. We thus show that a fiscal imbalance can cause an abandonment of a currency peg independently of any need for seigniorage revenues. If a currency crisis can be consistent with zero seigniorage revenues, it is a fortiori consistent with modest seigniorage gains seen in the recent episodes. Abstracting from money does not mean that we cannot use the model to analyze the effects of policy rules that change the interest rate on short-term nominal government bonds. Interest rate rules constitute monetary policy in our model. In Corsetti and Maćkowiak, 2000 and Corsetti and Maćkowiak, 2005 we show that the model with money, introduced in a standard way, behaves in the same way and all our results go through. Here we omit money to make the model simpler, in effect deleting an equation that would mechanically define the quantity of base money demanded for a given short-term nominal interest rate. The model clarifies what determines the size of a currency crisis. The magnitude of the crisis in the model is nonlinearly increasing in the fraction of public liabilities denominated in a foreign currency. The nonlinearity lets us understand why a government that borrows heavily in a foreign currency, like many emerging markets, is exposed to a devaluation of dramatic size. Our model clarifies characteristics of the post-crisis steady state and, in particular, the relation between the magnitude of devaluation and the rate of inflation in the post-crisis steady state. The post-devaluation regime in the Krugman model exhibits chronic depreciation (inflation). Furthermore, the first generation model implies a positive relation between the size of the currency crisis and the rate of chronic inflation. In contrast, in the recent episodes we witnessed one-time devaluations of large magnitude followed by little ongoing inflation. Our model predicts that a fiscal imbalance causes a one-time devaluation. Post-devaluation policy can set the nominal interest rate at any desired level, not necessarily switching to a regime of chronic depreciation. Furthermore, the relation between the magnitude of devaluation and the inflation rate in the post-devaluation steady state is negative. 3 The reason is that higher long-run inflation implies a larger capital loss to holders of long-term nominal public debt. Thus a currency crisis of striking proportions is consistent with zero inflation in the post-collapse steady state. 4 The rest of the paper is organized as follows. Section 2 presents the model, discussing shocks that can cause a fiscal imbalance as well as the fiscal rule and the interest rate rule. Section 3 uses the model to analyze the dynamic adjustment to a fiscal imbalance, discussing in turn determinants of the equilibrium exchange rate, the timing of speculative attacks, and the resolution of the paradoxes of the first generation model. Section 4 concludes, completing the discussion of the relationship between our framework and the first generation literature. In Appendix A, we describe the behavior of the model when post-devaluation interest rate policy is more general than a simple interest rate peg we consider in the main text.

This paper presents a simple framework to analyze the fiscal dimension of a currency crisis. An advantage of our framework is that it allows a discussion of the role of public debt and interest rate rules, two topics at the center of policy discussions following the recent crises. We have shown how external nominal shocks can produce a fiscal imbalance and undermine currency stability, resolved two well-known paradoxes of the first generation model, discussed the role of seigniorage revenues, and illustrated how fiscal and interest rate policies interact to determine the magnitude and the timing of speculative attacks and devaluations. Let us complete the discussion of the relationship between our framework and the first generation literature. The first generation authors emphasize the unique timing of devaluations caused by a fiscal imbalance – indeed, one often refers to uniqueness as one of the main insights from that literature. We note that according to the logic of the FTPL an imbalance per se implies only an upper bound for the timing. What explains this difference? The unique timing in the first generation models is due to an assumption of a minimum level of reserves – models with debt in effect assume that investors impose an exogenous borrowing constraint on the government. A different modeling strategy, an exogenous interest rate threshold, delivers the uniqueness result in our framework. It is apparent that the exogenous limit on reserves in the Krugman model is a special case of the constraint on interest rate policy we use. If one abstracts from the possibility of government borrowing – as Krugman does – his assumptions of an exogenous rate of domestic credit expansion and a lower threshold on reserves imply, taken together, a lower bound on the monetary base, or an upper limit on the interest rate. This paper points to a number of questions for further research. First, we have seen that small changes in the extent of “dollarization” can cause large differences in the magnitude of devaluation. This raises the question what determines the denomination of public debt. This issue is related to the recent literatures on “original sin” (e.g. Eichengreen and Hausmann, forthcoming), “sudden stops” (e.g. Calvo, 2003) and public bailout guarantees for private contracts in a foreign currency (e.g. Burnside et al., 2000). Second, what factors constrain fiscal reform and monetary policy after a fiscal shock? Political economy and concerns about financial stability are promising areas of analysis. Third, many devaluations coincide with both fiscal stress and sharp changes in output. The focus of this paper is on the former, while some recent models emphasize the latter.25 We regard our contribution as a building block toward future models that will simultaneously account for both aspects of crises.