داخل معمای موسی : رژیمهای سیاست پولی و نرخ واقعی ارز در یک اقتصاد کوچک باز

Industrial countries moving from fixed to floating exchange rate regimes experience dramatic rises in the variability of the real exchange rate. This evidence, forcefully documented by Mussa [Nominal exchange regimes and the behavior of real exchange rates: evidence and implications. Carnegie-Rochester Conference Series on Public Policy 25 (1986) 117], is a puzzle because it is hard to reconcile with the assumption of flexible prices. This paper lays out a dynamic general equilibrium model of a small open economy that combines nominal price rigidity with a systematic behavior of monetary policy able to approximate a continuum of exchange rate regimes. A version of the model with complete exchange rate pass-through is broadly consistent with Mussa’s findings. Most importantly, this holds independently of the underlying source of fluctuations in the economy, stressing the role of the nominal exchange rate regime per se in affecting the variability of the real exchange rate. However, only a model featuring incomplete exchange rate pass-through can account for a broader range of exchange rate statistics. Finally there exist ranges of values for either the degree of openness or the elasticity of substitution between domestic and foreign goods for which the baseline model is also consistent with the empirical insensitivity of output volatility to the type of exchange rate regime, as documented by Baxter and Stockman [Journal of Monetary Economics 23 (1989) 377].

For a long time economists have debated about whether fluctuations in the exchange rates reflect mere changes in relative money prices, as opposed to changes in the relative prices of goods or inputs. In a very influential paper, Mussa forcefully documents two facts:1 (i) Nominal and real exchange rates are strongly correlated; (ii) Industrial countries moving from fixed to floating exchange rate regimes experience dramatic rises in the variability of the real exchange rate. In Fig. 1, the German Mark–U.S. Dollar nominal and real exchange rates (top panel), the short-run variations of the real exchange rate (medium) and of its components (bottom) are plotted. The evidence is striking. Nominal and real exchange rate are almost perfectly correlated. A sharp increase in volatility stands out in the post-Bretton-Woods era for both the real and the nominal exchange rate, as opposed to a noticeably constant variability of the price level ratio. The decomposition of the real exchange rate into nominal depreciation rate and inflation differential shows a very weak correlation between these two components. Table 1 reports exchange rates statistics for several OECD countries. The volatility of the real depreciation rate is on average more than four times higher under floating than under fixed rates. The correlation between nominal and real exchange rate under floating is close to unity. Overall, the movements of the nominal exchange rate seem to play a dominant role in shaping the short-run variations of the real exchange rate. According to Mussa, these regularities systematically apply to every postwar exchange rate regime shift undertaken by an industrial country.At this point, it might seem that little room remains to argue that the exchange rates are a purely nominal phenomenon. Why, then, has this evidence often been treated as a puzzle? In principle the high correlation between nominal and real exchange rates may be rationalized in economies with perfectly flexible prices and a high incidence of real shocks, i.e., shocks originating in the goods market that require adjustments of the relative prices.2 Yet this would not explain why the volatility of the real exchange rate systematically starts to increase upon switching from a regime of fixed to one of floating rates. It could also be argued that the choice of the exchange rate regime is endogenous, and that it is indeed those countries experiencing large real shocks that choose to switch to floating exchange rates.3 The evidence on the change in volatility, however, is so overwhelming and extended over time that this does not seem a plausible explanation.4 Nor can the whole set of facts be rationalized in international real business cycle models, like the ones pioneered by the work of Backus et al. (1994), in which a switch in the exchange rate regime is simply not addressable. The last two columns of Table 1 report two additional exchange rates facts that are worth emphasizing. For one, the variability of the real always exceeds the one of the nominal depreciation rate. Furthermore, nominal depreciation rates and inflation differentials display a negligible correlation. There is in fact a large empirical evidence documenting that the pass-through of exchange rates to prices is low, a symptom of the failure of the law of one price (LOP) at the level of individual goods. Campa and Goldberg (2001) document that such deviations from LOP are larger for consumer than for import goods. Ghosh and Wolf (1994) provide evidence that incomplete pass-through may be the result of stickiness in the adjustment of import prices expressed in units of local currency. In Table 1 the sharp rise in volatility of the real exchange under floating stands in stark contrast with the apparent insensitivity of real output volatility to the change in the exchange rate regime. While on average output is under floating barely as volatile as under fixed rates, the real exchange rate is more than four times as variable under floating relative to fixed. This is consistent with the well-known evidence first reported in Baxter and Stockman (1989), who show that the business cycle properties of a broad range of real macroeconomic variables is independent of the underlying exchange rate regime.5 Along with the already described “Mussa facts”, this further evidence constitutes the empirical motivation of this work. I lay out a dynamic general equilibrium model of a small open economy characterized by two main features. The first one is a certain rigidity in the adjustment of prices, in accordance with Fig. 1. The second one is the commitment to (monetary) policies consistent with the maintenance of managed-fixed nominal exchange rates. This systematic component in policy stands in contrast with the supposed role of stochastic (real) shocks in explaining the change in variability of the real exchange rate after a switch in regime. The baseline version of the model assumes complete pass-through of exchange rate movements to prices. The key insight of the paper is twofold. First, the baseline sticky-price model with complete pass-through accords well with the Mussa facts reported in Table 1. More importantly, this holds independently of the underlying source of fluctuations, casting doubts on the theory that stresses the prominent role of real shocks. However, it is only a model featuring incomplete exchange rate pass-through that is able to fully account for all the exchange rate facts documented above. Namely, the “Mussa facts”, the observed ranking between real and nominal exchange rate variability, and the weak correlation between nominal depreciation and inflation differentials. Furthermore, the baseline model performs well in reproducing the Baxter and Stockman evidence on the insensitivity of output volatility to the shift in exchange rate regime. However this result is obtained only for specific parameterizations of the degree of openness and of the elasticity of substitution between domestic and foreign goods. A key goal of this work is to link the recent literature on sticky-price models of the open economy with the one on interest rate rules. The latter has seen contributions almost entirely confined to the closed economy.6 In an open economy context, however, where exchange rate regimes matter, it seems even more appropriate to think of monetary policy in terms of endogenous rules. A novelty of the present model lies, in fact, in the representation of a managed-fixed exchange rate regime by means of an interest rate rule assigning an increasing weight to the deviations of the nominal exchange rate from some theoretical parity. A central result is that a monotonic inverse relationship exists between the “degree of proximity” to a fixed exchange rate regime (measured by the weight assigned to the nominal exchange rate in the interest rate rule) and the volatility of the real exchange rate. Recently, several papers, rooted in an optimizing framework with nominal rigidities that has become the hallmark of the New Open Macroeconomics literature, have tried to address the related but distinct issue of the large and persistent deviations of the real exchange rate from the purchasing power parity (the “PPP puzzle”).7 Examples include Beaudry and Devereux (1995), Chari et al. (2000), Betts and Devereux (2000) and Kollmann (2001) among others. This literature, however, deals solely with the problem of explaining the absolute level of the real exchange rate volatility. Finn (1999) constructs a flexible price dynamic general equilibrium model in which the interaction of technology shocks with an accommodative role of money is able to generate the high empirical correlation between nominal and real exchange rate. My work focuses instead on the systematically different behavior of the real exchange rate across regimes. Furthermore all papers above do not analyze the role of interest rate rules. Hence monetary policy affects the business cycle only via its exogenous stochastic component. The remainder of the paper is as follows. Section 2 presents the basic model. Section 3 discusses the main findings on the relationship between real exchange rate and monetary policy regime, and the role of nominal rigidities. Section 4 presents an extension of the baseline model to include incomplete exchange rate pass-through. Section 5 concludes.

The formulation of monetary policy in terms of interest rate rules has attracted considerable attention in the recent macroeconomic literature. This tool has proved to be particularly useful in capturing the systematic role played by monetary policy over the business cycle. In this paper I have shown that a generalization of this setting to the open economy allows an evaluation of a central topic in international macroeconomics: the short-run dynamic effects of a change in the nominal exchange rate regime. Armed with this tool the paper rationalizes a series of facts characterizing the joint behavior of nominal exchange rates, real exchange rates and prices: the observed striking increase in volatility of the real exchange rate, the high correlation between nominal and real exchange rates under floating, and the very weak co-movement between the components of the real depreciation rate, namely the nominal depreciation rate and the inflation differential. The central result of the paper is that a model with sticky prices, endogenous monetary policy and incomplete exchange rate pass-through is best suited to match the exchange rates regularities described above. The key idea conveyed is that the choice of the monetary policy regime and therefore of the exchange rate arrangement is not neutral for the short-run adjustments of international relative prices. The present framework lends itself to several extensions. For example, a recent series of papers, in the light of work by Flood and Rose (1995), has reinterpreted the Baxter and Stockman results in a broader sense and argued about a more general exchange rate “disconnect puzzle”. According to this view it seems difficult, in optimizing rational expectations models, to relate the equilibrium movements of the exchange rate to the ones of the macroeconomic fundamentals. This holds in particular for the relationship between relative consumption and real exchange rate. I have abstracted from this issue here and concentrated on output only. Recent contributions by Devereux and Engel (2002), Corsetti et al. (2002) have made important steps in this direction.