پویایی رژیم نرخ ارز: رفع عیب ، شناور، و تلنگر

The impermanence of fixed exchange rates has become a stylized fact in international finance. The combination of the “mirage” view that pegs do not really peg with the “fear of floating” view that floats do not really float generates the conclusion that exchange rate regimes are, in practice, unimportant for the behavior of the exchange rate. This is consistent with evidence on the irrelevance of exchange rate regimes for general macroeconomic performance. Recent studies, however, show that the exchange rate regime matters. This can be understood by considering the dynamics of exchange rate regimes. We demonstrate that the “mirage” view is somewhat misleading and incomplete. Pegs frequently break, but many do last. Also, there is a high degree of flipping, that is, the re-formation of pegs that have broken. Thus, a fixed exchange rate today is a good predictor that one will exist in the future. We also investigate the quantitative effect of fixed exchange rates. While the “fear of floating” view suggests little actual difference in fixed and floating rates with respect to exchange rate volatility, we show that fixed exchange rates exhibit considerably greater bilateral exchange rate stability than flexible rates, both today and in the future.

The choice of an exchange rate regime and the consequences of this choice, traditionally represent a central topic in international finance. But, recently, research has called into question the relevance of this line of inquiry. This is demonstrated by two of the more evocative titles in international finance articles published in the last decade, “The Mirage of Fixed Exchange Rates” (Obstfeld and Rogoff, 1995), and “The Fear of Floating” (Calvo and Reinhart, 2002). Obstfeld and Rogoff (1995) suggest that fixed rates are not all that fixed, writing “literally only a handful of countries in the world today have continuously maintained tightly fixed exchange rates against any currency for five years or more.” [p. 87] Calvo and Reinhart (2002) argue that floating rates do not really float, rather governments that claim to allow market forces to determine the value of their currencies actually act to minimize exchange rate fluctuations. Taken together, these influential articles suggest that the exchange rate regime is, in practice, unimportant and perhaps irrelevant for the actual behavior of the exchange rate. This conclusion is bolstered by empirical research that finds little role for exchange rate regimes in determining major macroeconomic outcomes beyond the real exchange rate (see Mussa, 1986, Baxter and Stockman, 1989 and Flood and Rose, 1995). Altogether, based on these works, there is a general impression that exchange rate regimes — in spite of all the attention they receive — are in some ways unimportant. In this paper we bolster the traditional view of the importance of the exchange rate regime. One of our central results shows that fixed exchange rates are more substantial than mere mirages; while almost half of the fixed exchange rate spells do not last more than two years, the expected duration of a peg increases dramatically if it survives past that age. Consequently, at any one-time, the set of countries that are pegged includes a large proportion of those with a peg lasting for a relatively long duration. We also show that the distribution of floating exchange rate spells is similar to the distribution of fixed exchange rate spells, with a large number of short-lived floating exchange rate spells and a smaller number of long-lived floating exchange rate spells. An important implication of this is the flipping of pegged rates, that is, the end of a peg is often followed by the reformation of a new peg. This dynamic behavior of exchange rate regimes, when combined with an analysis of the duration of fixed exchange rate spells and floating exchange rate spells, paints a different picture than one would expect from the well-known “mirage” of fixed exchange rates. The mere classification of annual observations into those categorized as “pegs” and those categorized as “floats” (or, more accurately, non-pegs) would not matter if the “fear of floating” truly limited exchange rate flexibility when countries do not peg. We demonstrate, however, that the magnitude of bilateral exchange rate volatility between a country that has a pegged rate and its base country is quite distinct from bilateral volatility when a country does not peg.1 Countries that peg have lower multilateral volatility as well, mainly because these countries tend to avoid extreme bilateral volatility outcomes.2 Taken together, these two results, on the dynamics of exchange rate regimes and the implications of the exchange rate regime for volatility, implies that a country with a fixed exchange rate today can be expected to exhibit greater exchange rate stability both today and over the course of time. This helps explain some current evidence on the effects of exchange rate regimes, including the significant and positive effect of fixed exchange rates on trade (Klein and Shambaugh, 2006a), the limits that fixed exchange rates place on monetary autonomy (Shambaugh, 2004 and Obstfeld et al., 2005), and the effects of the exchange rate regime on the transmission of terms of trade shocks (Broda, 2004 and Edwards and Levy-Yeyati, 2005), the national price level (Broda, 2006) and inflation (Ghosh et al., 2002). The results presented in this paper also help explain why research has found that the exchange rate regime plays a role in determining the rate of growth (Aghion et al., 2006, Levy-Yeyati and Sturzenegger, 2003, Husain et al., 2005, Di Giovanni and Shambaugh, 2006 and Ghosh et al., 2002) and growth volatility (Ghosh et al., 1997). Thus, this paper demonstrates the extent to which exchange rate regimes really do matter for exchange rate outcomes, and helps explain the source of effects of recent research demonstrating trade and macroeconomic consequences of exchange rate regimes. It examines the nature of de facto exchange rate regimes themselves, studying their duration, dynamics, and the extent they affect the exchange rate. In essence, we try to provide a new set of stylized facts on exchange rate regimes that helps explain why the new wave of empirical evidence on their effects can occur

There has been a flurry of new research showing substantial effects of the exchange rate regime on trade and a variety of macroeconomic outcomes. These results seem inconsistent with widespread perception that pegs do not really peg and floats do not really float, that is, that exchange rate regimes do not matter for the exchange rate, let alone other outcomes. This paper provides a new set of empirical regularities regarding the relevance of exchange rate regimes for exchange rate related outcomes which, in an important way, is a prerequisite for understanding how these regimes affect any other outcomes. We find that, despite the fact that many peg spells break soon after they begin, a fair number last beyond five years. These longer spells are overrepresented in an annual cross-section, as compared to short spells. In addition, once a peg lasts longer that a year or two, the probability that it will continue for one more year, conditional on lasting up until that year, begins to rise dramatically. The implication of this is that the length of a particular peg is more important than the less specific knowledge of average peg duration for agents considering actions that are affected by the maintenance of the peg. We also show that float spells have properties similar to those of peg spells, especially with respect to the prevalence of many short duration spells. An implication of this is that many countries re-form pegs quickly after experiencing the end of a peg spell. We also demonstrate important quantitative differences in exchange rate volatility across exchange rate regimes. These results are obtained even when controlling for country and year fixed effects, and for inflation behavior and capital controls. The difference also persists into the future, and we find that a peg today predicts lower volatility for a number of years out. We also find that bilateral pegging lowers multilateral exchange rate volatility, although this comes from the avoidance of high volatility outcomes rather than through an effect across the wide distribution of pegs and floats. Finally, we show that extremely high volatility outcomes are not more likely to follow a year with a peg than a year with a float, although there are a disproportionate number of high volatility outcomes in the immediate wake of long duration pegs as compared to shorter duration pegs. There are a number of policy implications one could draw from these results. Pegging does promote greater exchange rate stability, but a newly initiated peg may not last and the first year after a peg has a significantly higher rate of volatility than other years during a float spell. In addition, a peg may not stabilize other bilateral exchange rates unless it prevents high volatility; but other stable economic policies or strong institutional structures are, perhaps, better positioned to do this. But once a peg has lasted for a few years, its likelihood of enduring increases and, even if it breaks, it is likely to re-form quickly. The results presented in this paper help resolve the puzzle of why recent empirical work has found the exchange rate regime can matter in a variety of contexts when the perception is that fixed exchange rates are a mirage and governments fear floating currencies. These results can also provide some guide for theory that focuses on exchange rate stabilization by highlighting the distinction between the effects of fixed exchange rates for multilateral and bilateral exchange rates. For example, current debate in the New Open Economy Macroeconomics literature revolves around the appropriate exchange rate regime choice, often focusing on the ability of pegs to stabilize prices and consumption across countries. However, in these two-country models, multilateral and bilateral volatility are the same. Our results show that the multilateral rate will be stabilized, but cannot be stabilized perfectly. The peg has a weaker effect on the multilateral rate and many pegs still have fairly volatile multilateral rates. Thus, we need to take care when using tractable two-country models to recall when results involving a peg require the peg to stabilize the bilateral rate or the multilateral rate.