Comments and Discussion

Nouriel Roubini:

This is an interesting, thought-provoking, and important paper on why emerging market economies may have reason to view floating exchange rates with concern—what the authors memorably term a "fear of floating."

In evaluating the authors' findings, it is noteworthy that the currency and financial crises of the 1990s have been crises of soft-peg (or intermediate exchange rate) regimes: the European Rate Mechanism in 1992–93, Mexico in 1994–95, East Asia in 1997–98, Russia in 1998, and Brazil in 1999.

The crises of the 1990s point to several conclusions about fixed rate regimes in emerging market economies:

• mdash Fixed rate regimes are fragile when such regimes are not consistent with economic fundamentals and with monetary and fiscal policies.

• mdash They do not necessarily provide monetary or fiscal discipline.

• mdash They may lead to currency overvaluation, which widens current account imbalances and eventually leads to currency collapses.

• mdash There are often real national or idiosyncratic shocks that require exchange rate flexibility.

• mdash In the absence of two-sided exchange rate risk hedging, they may lead to the buildup of short-term foreign currency borrowing (as in Asia). This makes countries vulnerable to liquidity shocks and balance sheet shocks. Two-sided exchange rate risk instead gives incentives to hedge.

• mdash Finally, fixed rates may lead to moral hazard, that is, excessive foreign currency borrowing with the expectation of a bailout in the event of substantial loss.

Citing the same evidence, however, many observers suggest that intermediate exchange rate regimes are neither feasible nor desirable and recommend that emerging economies adopt so-called corner solutions: independently floating exchange rates or fixed exchange rates (in the form of currency boards or currency pegs). Critics of the growing consensus for corner solutions cluster around two arguments: one school of thought holds that intermediate regimes, in the form of managed floats or crawling pegs or bands, are viable solutions: their position is that "no single exchange rate regime is right for all countries or at all times."¹ Others argue that the flexible exchange rate corner is not a realistic solution for emerging economies and suggest that such economies establish currency boards or even abandon their national currency and adopt the dollar or another strong currency ("dollarization").²

Being in favor of the flexible exchange rate corner, however, is not necessarily an argument for a pure float. The United States intervenes in foreign exchange markets from time to time, although much less frequently than Japan and the European Union. Indeed, there are good reasons why emerging economies may be concerned about excessive exchange rate volatility. Occasional intervention or policies that target domestic interest rates may be necessary in order to smooth excessive exchange rate movements. This view is fully consistent with supporting "dirty floats." Nonetheless, there is substantial difference between soft- or semi-pegs and a dirty float, which is less subject to abrupt changes in the value of the domestic currency since monetary authorities, while trying to smooth excessive movements in exchange rates, will not try to prevent changes in exchange rate trends.

During the last several years, many emerging economies have moved away from pegs, soft-pegs, and crawls to regimes of greater exchange rate flexibility, essentially regimes of "dirty floats," including Mexico (December 1994), South Africa (March 1995), Thailand (July 1997), Indonesia (August 1997), South Korea (December 1997), Brazil (January 1999), Israel (1999), Chile (September 1999), Colombia (September 1999), and Poland (April 2000). Although as yet there are no systematic, country-specific studies of dirty floats, a good number of national monetary authorities have found that such regimes have in fact been performing quite well. These regimes have provided flexibility to monetary authorities even if, given the international financial turmoil of the last few years, interest rate tightening was required at times to signal policy credibility and avoid excessive exchange rate movements.

The flexibility of exchange rates has helped in this regard, for the exchange rate becomes a partial absorber of external shocks. Given shocks to international interest rates, or access to international capital markets, or to terms of trade, the adjustment in these countries has occurred partly through higher interest rates and partly through a weaker currency. Thus exchange rate flexibility has helped to adjust to and absorb external shocks. Regardless of the exchange rate regime, however, emerging market countries need to maintain sound economic policies to protect themselves from financial crises.

Many of the results that Guillermo Calvo and Carmen Reinhart present in their paper represent an indictment of soft pegs that eventually collapse rather than a critique of exchange rate flexibility. Indeed, there is a vast difference between a fixed exchange rate regime that is not sustainable (that is, one that collapses through a sharp and discrete devaluation), leading to a currency or financial crisis, and a flexible exchange rate regime in which external shocks lead to changes in the equilibrium exchange rate in real time. Many of the effects found in this paper apply to fixed regimes that have collapsed, resulting in the severe consequences found by the authors.

Let us consider the authors' findings individually.

1. Devaluations are contractionary in emerging economies

The finding may well be true, but it implies that emerging economies may be better off with flexible exchange rate regimes than with unsustainable fixed rate regimes. A devaluation is not necessarily contractionary if it leads to an "expansionary" real depreciation. The main reason for a contraction is a balance-sheet channel: when a large stock of unhedged liabilities is denominated in foreign currency, a devaluation may be contractionary. Many observers have suggested that, historically, fixed rate regimes (where there was no two-sided exchange rate risk) have led to excessive unhedged borrowing in foreign currency by governments, financial firms, and corporations by distorting borrowing choices and leading to an excessive accumulation of foreign currency debt. Thus fixed rate regimes may pose moral hazard, because the guarantee of a peg distorts borrowing decisions by the private and public sector.

Flexible exchange rates, by contrast, provide two-way exchange rate risk and force borrowers to hedge more or to assume less foreign currency debt (or both). Under flexible exchange rate regimes, adjustments to internal or external shocks are smoother, more continuous, and less discrete than they are under pegs: sharp economic contractions, such as those associated with an abrupt fall in the currency when a peg collapses, are less likely to occur under a flexible regime.

Thus exchange rate movements under a dirty float should not have the same disruptive balance-sheet effects that they do under fixed rates: they should be less contractionary. Indeed, currency depreciation may enable economies to adjust to external disruptions (such as terms of trade shocks) and can have an expansionary effect.

2. The adjustments in the current account following devaluations are far more acute and abrupt in emerging market economies than they are in advanced economies.

This result may be true but is again partly due to the consequences of unsustainable soft pegs. Over time, such pegs may lead to real appreciation, a loss of competitiveness, and an increase in the current account deficit, leading, in turn, to excessive accumulation of short-term foreign currency debt. The ensuing collapse of the currency is often associated with financial distress, a sudden cutoff of access to international capital markets, and a painful need to restore the balance of trade by forcing a contraction in domestic demand, sharply reducing imports, and restoring competitiveness by means of large real depreciation.

Pure floats or dirty floats can help to attenuate these effects in various ways: they may prevent in the first place excessive real appreciation and unsustainable current account deficits; the smoother adjustment of nominal exchange rates, moreover, does not lead to the financial distress associated with abrupt changes in currency values and the ensuing cutoff of access to international capital markets.

3. Credibility and market access, as captured in the behavior of credit ratings following devaluations, are adversely affected by devaluations.

This result, again, suggests the dangers of soft pegs. Devaluations that follow unsustainable pegs, overvaluation, large external imbalances, and buildups of foreign currency debt lead to currency crises and financial crises. The ensuing loss of market access and sharp downgrade in credit rating is a direct result of having an unsustainable exchange rate regime. Conversely, sound economic policies and flexible exchange rates may reduce the risk of a sudden cutoff of access to international capital market that follows a sharp currency devaluation.

4. Lack of credibility gives rise to marked volatility in domestic interest rates; monetary and fiscal policies are procyclical.

External shocks (such as abrupt changes in global interest rates, sudden reductions of access to international capital markets, or terms-of-trade shocks) will generally be contractionary and lead to procyclical policy in emerging markets: they reduce economic activity and force national monetary authorities to raise interest rates in order to signal credibility and avoid excessive capital outflows and excessive currency depreciation. This does not imply, however, that a dirty float exchange rate regime will fare worse following such shocks than will a fixed exchange rate regime. To the contrary, although several countries with floating exchange rates were forced to raise interest rates during the 1997–99 crises, countries with fixed exchange rate regimes (even institutionally fixed regimes such as currency boards) were forced to do the same. Interest rate hikes and increases in country spreads were as high in Argentina and Hong Kong during the global crisis as they were for floaters.

The contractionary effects of interest rate tightening following external shocks, moreover, were of a higher magnitude in countries with fixed rate regimes than in countries with floating exchange rates. Compare what happened to output in Hong Kong with what happened in Taiwan and Singapore (which had similar fundamentals but allowed their currencies to depreciate), or Argentina's policy with that of Mexico, Brazil, Chile, and Peru. Fixed exchange rates do not shelter economies from external financial shocks or from contagion.

Finally, although monetary policy among floating exchange rate economies during the 1997–99 crises was constrained by the need to show credibility following external shocks, monetary authorities used exchange rate flexibility to allow their currencies to depreciate and thus reduced the negative effects of external shocks on aggregate demand. The fact that output fell less for the floaters in Asia and Latin America than for countries with rigid pegs, such as Hong Kong and Argentina, has partly to do with exchange rate flexibility. While monetary policy may be partly procyclical following negative supply external shocks (it is usually so even in industrial countries in the presence of stagflationary shocks such as the 1973 and 1979 oil crises), this does not rule out countercyclical monetary policy. Shocks to aggregate demand and even to terms of trade can be partly absorbed through a currency depreciation.

In this respect, the experience of small OECD open economies is telling. Floating exchange rate economies such as Australia, Canada, New Zealand, Sweden, and the United Kingdom have monetary policy autonomy and have adjusted to external terms-of-trade and demand shocks by allowing their currencies to depreciate. Australia avoided the Asian crisis and sustained 4 percent economic growth in 1998 by allowing its currency to fall in the wake of Asian export demand shocks and the shock to its terms of trade (primary commodities).

The counterargument that OECD countries are different from emerging economies because they do not suffer from a lack of policy credibility that rules out borrowing in domestic currency—the "original sin"—is only partly valid. Original sin does not condemn a country to hell for eternity. Sustained sound macroeconomic and structural policies may eventually lead emerging economies to OECD Eden. Countries such as Chile and South Africa can and have, over time, been able to borrow long term in their own currency and have greater monetary policy autonomy given a history of sustained policy credibility.

Even economies with partial policy credibility (such as Mexico, Peru, and Brazil) have been able to use some degree of exchange rate flexibility (and less monetary tightening than otherwise) to adjust to shocks. In Chile, which suffered so visibly from fear of floating in 1998–99, the monetary authorities seem to have realized that they should have allowed the currency to fall more in order to absorb the effects of the fall in copper prices.

5. Exchange rate volatility appears to be more damaging to trade (and the pass-through from exchange rate swings to inflation far higher) in emerging economies than in developed economies.

Emerging market economies may well feel the effects of exchange rate swings more acutely than advanced economies, but one may wonder to what extent these results depend on including within the sample fixed rate regimes that are unsustainable and eventually crash, leading to disruption of trade via credit crunches (as in South Korea, Thailand and Indonesia during the 1997–98 crisis). Currency crashes in economies with fixed rate regimes may in fact be more disruptive of trade than exchange rate volatility in floating rate regimes. Under floating rate regimes, economic agents have an incentive to deal with exchange rate volatility through hedging, adjusting their profit margins, and denominating their exports in particular currencies.

One would, of course, expect that the pass-through from exchange rates to inflation would be larger in small, open economies such as the emerging markets. However, the experience of the 1990s crises is again instructive and somewhat different from Calvo and Reinhart's findings. What is surprising about the currency crises of the late 1990s is how small the pass-through of currency depreciation to inflation ultimately was—between 10 and 15 percent. In Thailand, South Korea, Russia, and Brazil, the pass-through was very small. In Indonesia, inflation surged in 1998 (and fell sharply afterward), but given the size of the nominal depreciation, the pass-through was considerably smaller. Only in the case of Mexico in 1995 was the pass-through of the domestic currency devaluation to inflation relatively large.

More broadly, how widespread and how justified is emerging fear of floating among emerging economies? I would argue, on both counts: not much. Flexible exchange rates have provided emerging economies with a degree of monetary autonomy and an ability to respond to external shocks. Flexible exchange rate regimes have thus successfully minimized the real effects of economic disturbances. In fact, evidence and experience with flexible exchange rates in recent years, as well as some recent academic research, suggest that the arguments against flexible exchange rates are exaggerated:

1. 1. Policy credibility is gained with sound policies, not with the choice of the exchange rate regime. Fixed rates do not necessarily provide monetary or fiscal discipline, as the collapse of many pegged regimes proves.

2. 2. There is only partial liability dollarization in emerging markets (and little in Asia and South Africa), and sound policies may over time lead to a reduction in the degree of dollarization. Brazil, for example, has more financial indexation than liability dollarization.

3. 3. Countries preserve some degree of monetary autonomy under flexible exchange rate regimes. Eduardo Borensztein and Jeromin Zettelmeyer find that floaters are less sensitive to interest rate tightening than fixers.³ During the crisis of 1997–99, it was appropriate for countries with floating exchange rate regimes to raise interest rates in response to external shocks, but even countries with fixed exchange rate regimes were forced to tighten their interest rates considerably.

4. 4. Devaluations are contractionary under fixed rates because this regime leads to a buildup of foreign currency liabilities. Depreciations are less likely to be contractionary under flexible exchange rates. Moreover, negative balance-sheet effects occur also in fixed rate regimes when there are shocks that require a real depreciation.⁴

5. 5. Flexible exchange rates provide some shock-absorbing functions when there are terms-of-trade shocks. Christian Broda has demonstrated that the real exchange rate and output fall less sharply under flexible exchange rates than under fixed rates,⁵ a finding consistent with the experience of recent years. (Compare the extent of economic contraction in Taiwan and Singapore with that in Hong Kong; or the contractions in Chile, Brazil, Peru and Mexico with that in Argentina.)

6. 6. Inflation targeting and other monetary rules have provided credibility and allowed emerging economies with floating exchange rate (for example, Brazil, Indonesia, Israel,Mexico, Peru, South Korea, and Thailand) to maintain low inflation rates.

7. 7. Although there is not yet a systematic empirical study, the growth, inflation, export, balance of trade, and overall economic performance of countries on a float during the last few years have been satisfactory, as the experience of Brazil, Indonesia, Israel, Mexico, Peru, South Korea, Thailand, and other emerging economies suggests. These countries have regimes that are closer to dirty float rather than pure float, but there is no evidence that emerging markets cannot live with regimes that are closer to floating rates than fixed rates.

In summary, exchange rate flexibility among emerging economies has allowed some monetary autonomy and permitted the exchange rate to perform its shock-absorbing function in the presence of domestic and external shocks (such as terms-of trade shocks, shocks to world interest rates and sudden stops in the flow of capital to emerging markets).

Calvo and Reinhart raise several other points that merit elaboration:

• mdash Even the analytical model in the paper shows the risks of pegged regimes: a devaluation may trigger a loss of access to international capital markets.

• mdash The discussion of lender-of-last resort function should distinguish between systemic banking crises with large fiscal costs where monetization of such costs can lead to high inflation and a Diamond-Dybvig model of bank runs (that is, liquidity crises), where the authorities can provide domestic liquidity without inflation or depreciation. In the latter circumstances dollarized countries may be in trouble because their lender of last resort may have very limited resources.

• mdash Liability dollarization may be exacerbated by fixed rates. Many emerging economies moved to dirty floats in spite of partial liability dollarization and have performed well (Mexico, Brazil, and Chile).

Emerging economies may have a fear of floating, but their recent experience with dirty floats has been relatively successful. Given real and financial shocks of the last decade, one would expect foreign exchange reserves and interest rates to be volatile in emerging markets. The arguments and results in the paper do not necessarily present a strong argument for currency boards or dollarization. Dollarization may entail several costs, and the criteria for optimal dollarization are quite stringent. Finally, many of the results found in the paper may have more to do with the implications of unsustainable soft pegs that eventually collapse than with the effects of regimes of greater exchange rate flexibility.

Ricardo Hausmann:

Guillermo Calvo and Carmen Reinhart have written an excellent paper full of surprising stylized facts and new theoretical insights. It will serve as a source of inspiration for many new papers to come.

The paper attempts to establish a set of stylized facts about the behavior of emerging markets that are particularly interesting because they are not a priori obvious:

• mdash It is hard to distinguish alternative exchange rate systems by the ex post amount of flexibility they exhibit. Among countries that float their currency, emerging markets allow less exchange rate flexibility than industrial countries: this "fear of floating" is surprising because a standard floating rate model would predict that because emerging economies are subject to larger real shocks, they should use more, not less, exchange rate flexibility.

• mdash Exchange rate movements in emerging economies do not help stabilize the domestic price of commodities, hence they cannot be said to be used to adjust to this kind of real shock.

• mdash Emerging economies with floating exchange rates allow a higher degree of interest rate volatility.

• mdash There is a positive covariance between depreciations and increases in interest rates in emerging economies.

• mdash Emerging economies show a positive relationship between country risk and depreciations.

• mdash Emerging economies suffer from "sudden stops," that is, crises occur in the context of a much larger swing in the current account.

• mdash Emerging economies have a higher pass-through of exchange rate movements into prices.

These stylized facts are problematic for a Mundell-Fleming interpretation of the world. Under that framework, one would expect countries with larger real shocks to exhibit more exchange rate flexibility. One would expect floating exchange rate countries to exhibit more stable interest rates and less volatile reserves than fixed rate regimes and that depreciations would be expansionary, not contractionary, under such circumstances. And one would not expect currency movements to affect country risk.

So the question is, What do we need in a model to account for these stylized facts? The paper tries to attribute these features to credibility problems in emerging economies. In order to do this, the paper is sprinkled with neat theoretical results relating lack of credibility to some of the stylized facts. For example, the paper shows that lack of credibility may create the expectation of a future rise in the money supply, causing both depreciation and a rise in interest rates in the present period. The paper also shows that depreciations would be contractionary in the absence of capital mobility. Moreover, if a depreciation were followed by loss of access to foreign finance, then its impact would be even more contractionary.

The implicit message of the paper is that the stylized facts are caused by lack of credibility in emerging economies One possibility is that emerging economies with floating exchange rate regimes allow less exchange rate flexibility because they fear that depreciations may signal lack of commitment to low inflation. Hence monetary authorities in these countries react to pressure on the currency by raising interest rates, but given that they lack the exchange rate regime as a commitment device, they are forced to raise interest rates more than more credible floaters or economies with greater exchange rate fixity. As a result these economies end up with a procyclical rather than a stabilizing monetary policy.

A high pass-through may also help explain some of the results. If the pass-through were high, the central bank would try to prevent exchange rate fluctuations by intervening in foreign exchange markets or interest rates. Alternatively, liability dollarization may also account for this result, as currency depreciations would weaken corporate balance sheets, leading to a credit crunch and a contraction of economic activity. In order to avoid this outcome, central banks in emerging economies would also seek more exchange rate stability than the standard model would predict. It is less clear how these latter two explanations relate to lack of credibility.

My main criticism of the paper is that it states that lack of credibility is the principal cause behind the anomalous stylized facts but does not account for the origins of this lack of credibility. The logic is tautological. If country A exhibits these traits, then it must lack credibility. Such a theory is not falsifiable. When I encounter a country B that exhibits different patterns of behavior, am I supposed to attribute that behavior to the fact that it does have credibility? And if country A changes its behavior, does it do so because it gained credibility? All of these patterns are explained through unobservable factors and hence can never be directly tested.

Moreover, this lack of credibility does not seem to have a clear source. It is easy to argue that Latin America's history of inflation and fiscal problems has bequeathed it low credibility and hence "fear of floating." But why should the East Asian miracles, with their history of fiscal probity and low inflation, exhibit fear of floating? Why should they be less credible than Australia or South Africa? Without a theory of what causes credibility, one is left with a very ad hoc interpretation of the stylized facts and a shaky ground on which to base policy.

One alternative is to empirically ground the causes of fear of floating. In a paper coauthored with Ugo Panizza and Ernesto Stein, we studied the behavior exhibited by a sample of thirty industrial and developing countries classified as either floating or having very wide bands.¹ We found strong evidence of fear of floating in developing countries in the sense that these nations float with very large and fluctuating reserves, with very stable exchange rates and very unstable interest rates—in stark contrast with the G-3 countries and even with other industrial countries (see table 1). Interestingly, this sample of countries excludes soft pegs and hence shows that contrary to Nouriel Roubini's comment, fear of floating is not driven by soft pegs.

But what variables can account for this? We were able to show that income per capita and years of experience with floating regimes (a proxy for the accumulated reputation) are essentially orthogonal to the issue. We found surprisingly weak association between fear of floating and pass-through, suggesting that it is not the major actor in this play.

Instead,we found a very strong association between fear of floating and the presence of international debt denominated in a country's own currency. Countries with a significant amount of foreign debt in domestic currency exhibit much more flexibility than countries with foreign debt denominated in foreign currency. The data are shown in table 2. It is constructed by taking all international placements of bonds and money market instruments, calculating all the debt outstanding in a currency, and dividing it by the total debt outstanding by residents of a country. For example, the ratio for the United States is greater than 1 because there is more debt issued in U.S. dollars than there is debt issued by U.S. entities, public or private. What is remarkable is the extent to which this ratio is concentrated among a few countries. More than 90 percent of all

Table 1.

Evidence of Fear of Floating

Table 2.

Ability to Borrow in Domestic Currency

international issues of bonds and money market instruments takes place in six currencies. For sixteen out of the thirty countries in the sample, the number is 0. For the United States, Japan, and Switzerland it is much greater than 1. Interestingly, the number is large for Australia, New Zealand, and especially South Africa, three countries that show no fear of floating.

It has become commonplace to say that countries expose themselves to currency mismatches because of moral hazard. Nouriel Roubini in his comment argues that soft pegs encourage dollarization because they offer an implicit currency guarantee. This logic would explain why one would observe less borrowing in domestic currency among emerging markets countries than among industrialized countries. It cannot explain why we do not observe international transactions in these currencies by any private corporation, domestic or foreign, or any investment bank anywhere in the world. The evidence is more compatible with inability (rather than unwillingness) to borrow in domestic currency, otherwise referred to as "original sin."

If we take as our exogenous variable a country's ability to borrow in its own currency, then a net foreign debt will imply a net currency mismatch, which simply cannot be hedged, as a dollar debt plus a hedge is equivalent to borrowing in local currency. If it were possible, major banks would offer local currency loans and would hedge the exposure themselves.

The aggregate currency mismatch would explain the contractionary nature of depreciations, through the balance-sheet channel, and would account for the fact that country risk increases and market access declines after depreciations. It would explain why a rational central bank would try to protect the economy by holding very large stocks of reserves and by using intervention with reserves and with interest rates in order to stabilize the currency. The results of our paper show a very strong empirical association between what we take as "ability to borrow" and the symptoms of fear of floating.

But why are some countries able to borrow in their own currency and others not? Is this not equivalent to simply asking why some countries are credible and others not? I do not think so. First, we now have a measurable market phenomenon that we can try to account for and that is not tautologically linked to fear of floating. Second, it shows surprising variance across countries and over time. South Africa's experience with the euro-rand market dates from the mid-1990s. Mexico and Chile have actively attempted to develop an international euro-peso market, albeit with scant success. These experiences may shed some light into what determines a country's ability to borrow in its own currency and hence lead to a further set of questions.

Guillermo Calvo and Carmen Reinhart have produced a paper that is full of stylized facts that need explanation. They provide theoretical hints about potential causes but do not try to show empirically the relevance of the alternative interpretations. The challenge now is to find empirically verifiable explanations of the causes of fear of floating.

General Discussion:

Frederic Mishkin pointed out that standard textbook treatments of exchange rate policies, which usually focus on issues from the perspective of industrialized countries, may lead to the wrong policies for emerging market countries. One of the paper's strengths, he noted, is how it illustrates that emerging market countries cannot afford to treat their currency's exchange rate with the benign neglect that industrialized countries can. But the issue, Mishkin argued, is not fixed versus flexible exchange rates; rather, the debate comes down to which monetary and fiscal policy institutions get policymakers to the right place. A key issue is how to constrain discretion properly. In some cases (such as South Korea and Thailand), inflation targeting might be a useful regime to constrain discretion. In the absence of appropriate monetary policy institutions, however, national economies will lack credibility and will be unable to secure the benefits of monetary autonomy.

Ralph Bryant suggested that the likelihood of contractionary devaluations depends on what kind of shock moves the exchange rate. He also argued that a country's ability to borrow in its own currency is very much an endogenous variable. Paul Masson echoed Nouriel Roubini's distinction between floating that involves no long-run commitment and the type of floating that the authors seem to view with concern: one that attempts to smooth fluctuations and reduce volatility over the short term.

Dani Rodrik found that the paper succeeds in raising doubts about whether a pure float could actually work for emerging markets but noted that the question of float is distinct from the question of how useful changes in the nominal exchange rate can be. Rodrik took specific issue with the authors' statement that "devaluations in developing countries have a history of being associated with recessions, not export-led booms," arguing that nearly all the significant growth spurts of the last four decades have been preceded (and to a large extent assisted) by significant currency devaluations. As examples he cited South Korea and Taiwan in the early 1960s, Turkey in 1980, Chile and Mauritius in the mid-1980s, Poland in 1990, and India in the early 1990s.

Alan Blinder took issue with the purported exogeneity of borrowing in local currency, arguing that such borrowing is, in fact, a choice. The more accurate characterization, he argued, is that it is much more difficult for countries that do not have strong currencies to borrow in their own currencies than it is, say, for the United States or the United Kingdom. The choice of a peg greatly encourages borrowing in foreign currency: people start pretending that one baht is just another name for four cents. When South Africa borrows in rand rather than in dollars, it has to pay a premium of approximately three hundred basis points, but this puts the risk in the hands of the people most willing and able to bear it. Blinder argued that it is the combination of a soft peg and excessive borrowing in dollars that poses the greatest risk. The combination of a dirty float and domestic currency borrowing, even though it will not be as easy for the borrower as it would be for the United States or the United Kingdom, makes for a much more stable environment and a safer system.

Carmen Reinhart added that the inability to borrow in one's own currency is not entirely exogenous but emphasized that borrowing in dollars leads to a vicious circle. Once a government knows that the private sector's capacity to hedge is limited, it may be reluctant (for a variety of reasons) to float its currency. Dollar-denominated obligations, moreover, are not easily satisfied by a country with a weak currency. Ricardo Hausmann noted that Calvo and Reinhart do not make an argument for borrowing in domestic currency and said that it is difficult for an emerging economy to develop a market when it in effect tells the rest of the world, "lend me a lot of money in a unit that I can manipulate." Jeffrey Frankel cited Andrew Rose's finding that adopting a currency union has a large positive effect on trade,¹ and he suggested that the finding implies that there may be some discontinuity at the corners of exchange rate regimes.

Reinhart concluded the discussion by drawing attention to the issue of interest rate volatility. In the United States, the probability that interest rates will rise by five hundred basis points is zero; in Mexico, the probability is 29 percent. She argued that interest rate volatility has deleterious consequences that should not be underestimated.