26 July

We hold against that mainstream view. We show that corner solutions are not as crisis-free as is often maintained. Pure floating has at times led to costly misalignment of exchange rates, devastation of unhedged balance sheets, and inflation import – reasons why it is rarely practised in developing countries. Hard pegs have become more popular, but we visit two important cases – Africa’s CFA experience and Argentina’s currency board – which warn against underestimating the risks involved. In maritime terms, no sensible sailor drops the anchor before the boat stops moving.

While hard pegs often confer initial gains in credibility and hence lower capital cost, these gains can be ephemeral when they are not supported by a sufficient degree of institutional development and economic flexibility. Both Africa and Argentina became trapped by an inappropiate anchor currency, inappropriate as the anchor did neither reflect their trade directions nor their cyclical needs. As there are few currencies available to borrow credibility from, this lesson will not be unique: it suggests either a basket peg or, if a realistic option, to build rather than borrow credibility.

The prospect of regional integration invalidates corner solutions as non-cooperative (float) and costly to exit (hard pegs), but it revives the intermediate exchange-rate regime. The EMS experience shows that target zones plus effective codes of conduct, wide enough to allow for sufficient flexibility, can indeed confer sustained credibility as to avoid large misalignments and to reduce crisis vulnerability. What they need to achieve these objectives goes beyond the public perception that the central parity is consistent with long-term fundamentals. Expectations need to be guided by mutually agreed and surveilled governance codes towards intensifying integration, based on visible progress in macroeconomic stability and regulatory reform.

The MSF (multilateral surveillance framework) has to be "owned" by, rather than imposed on, the countries concerned. It must therefore be supported by peer pressure and yardstick competition, both of which are built gradually. The "Eurocentric" approach to earning credibility on the way to monetary integration holds impressive successes in the former periphery. This is why the practical operation of the EMS provides important lessons for authorities struggling to implement sustainable exchange-rate regimes to support economic convergence. These lessons are beginning to spread beyond the European continent, reaching Asia and Latin America. While a EU style MSF is no panacea either, the fact that it does not seem to avoid difficult choices will certainly be welcome by reformist governments worldwide.

Identifying and implementing an appropriate exchange-rate regime is giving rise to heated debate and urgent search. Virulent and contagious financial crises have hit diverse developing countries repeatedly over the past decade, the last instance being Turkey, a founding member of the OECD. The recurrence of crises has redefined policy choices and trade-offs in a world of intense capital mobility. ‘Corner Solutions’, either moves towards purely floating regimes or hard pegs such as dollarisation or currency boards, became the flavour of the day in prescriptions if not in practice. Moreover, the introduction of the euro may have led to a period of higher fluctuations between the key currencies, dollar, euro and yen, ensuing risks of destabilising trade-weighted exchange rates in many countries facing several different export markets. Hence the quest for regional monetary integration, as a means to earn credibility in world financial markets and therefore to promote a sustained convergence of living standards.

We concede that pure floating is rarely practised in emerging markets, and for good reasons. Old and recent historical experience warns against adopting hard pegs or establishing monetary integration schemes without a sufficient set of institutional prerequisites built up beforehand. That experience will be highlighted here.We shall build upon the CFA experience and upon the Argentine currency board to argue that corner solutions are not a panacea.

Is there room for clean "dirty floating"? Much can be learned from the European Monetary System (EMS), especially from the latter part of its existence, after the 1992 crisis. Central parity with large bands seems to have worked reasonably well, in the midst of large uncertainties regarding the path to Economic and Monetary Union (EMU). The way out of the "don't fix, don't float" conundrum is the idea that you may have to float in order to fix. This is like a financial "cruel to be kind" (Macedo 2001a).

In practice, the operation of the EMS brought out a largely unwritten code of conduct through the Exchange Rate Mechanism (ERM). This turned what was essentially an agreement between central banks into an instrument of budgetary discipline and then into a vehicle for the convergence of Ireland, Spain and Portugal towards the criteria set out by the 1992 Union treaty. Acquiring financial reputation was an essential prerequisite for the socalled cohesion countries to catch up with the EU average (see, e.g. Reisen 1993a, Macedo 1996).

Co-ordination mechanisms among monetary and fiscal authorities like the ones found in the European Union (EU) have been the best response to threats of contagion of national crises. In effect, the current international system calls for more effective regional and global responses adapting EU methods in cases like the Central European Free Trade Association (CEFTA), the Association of South East Asian Nations (ASEAN) and Mercosul. The evolution of the swap arrangements between ASEAN, China, Japan and Korea towards more systematic macroeconomic consultations and the creation of a "macroeconomic monitoring group" in Mercosul are specific examples presented in Macedo (2000b). The Financial Times of 7 March 2001 reviews these arrangements under the title of "yearning for stability spurs surge of regional harmony"…. The ERM code of conduct reflects a particular governance response to the pressure of globalisation which impinged on the European economy. With respect to African economies, which stand to the very early stage of their institutional endeavor, this may be a safer way to monetary stability than the one-shot jump to full monetary integration.

The adoption of a regional multilateral surveillance framework (MSF) complementary to IMF global surveillance may promote the kind of "peer pressure" that has become associated with the ERM code of conduct. Just like the emerging markets which are moving towards more integration, African countries will find useful lessons in the practical operation of the EMS. Another painful lesson in CFA countries highlights the importance of an appropriate anchor currency.

The paper is organised into two parts and five sections. Part I deals with so-called "corner solutions" such as currency boards or dollarization. Section I.1 provides some theoretical background and a brief survey of recent contributions to the policy literature. Section I.2 deals with the CFA experience, explaining why the 1994 devaluation was inevitable in view of the appreciation of the real effective exchange rate. Similarly, section I.3 explains how Argentina could not manage a real exchange rate misalignment.

Part II aims at showing why a transition where a country "earns" its credibility is a crucial prerequisite to any fixed exchange rate arrangement. Section II.1 looks at the EMS experience, stressing the portability of the MSF that emerged from the ERM code of conduct. Section II.2 draws lessons for the choice of exchange rate regime in emerging markets and for international financial architecture. A general conclusion then interprets the "don't fix and don't float" conundrum chosen for the title of this volume as a rationale for intermediate regimes capable of earning credibility abroad through peer pressure.

The adoption of appropriate exchange rate regimes has become a live theme. Consider the following three important statements:

"the choice of an appropriate exchange rate regime, which, for economies with access to international capital markets, increasingly means a move away from the middle ground of pegged but adjustable fixed exchange rates towards the corner regimes of either flexible exchange rates or a fixed exchange rate supported, if necessary, by a commitment to give up altogether an independent monetary policy" Lawrence H. Summers (2000)

"I will argue that proponents of what is now known as the bipolar view – myself included – probably have exaggerated their point for dramatic effect. The right statement is that for countries open to international capital flows: (i) pegs are not sustainable unless they are very hard indeed; but (ii) that a wide variety of flexible arrangements are possible" Stanley Fischer (2001)

"Countries seeking both to maintain a flexibility and to avoid excessive volatility in exchange rates might consider intermediate regimes such as band arrangements" Third Asia-Europe Finance Ministers’ Meeting, Chairman’s Statement (2001)

Immediately after the Asian crisis, it had become an almost common view that intermediate or BBC regimes (bands, basket or crawling pegs) are not sustainable in a world of intense capital mobility. There were few efforts to revive the intermediate option (but see Williamson, 2000). Countries were advised, e.g. by the US treasury, to the corners of either firm-fixing or free-floating. As stressed by Fischer (2001), each of the major international capital market-related crises since 1994 had in some way involved a pegged exchange rate regime. Turkey in early 2001 added another victim to the list of pegs gone bust. But the Treasury advice also reflected the desire to keep capital markets open, as can be inferred from Figure 1.

[Insert Figure 1 here]

The logic behind the proposition in favour or corner solutions is the impossible trinity (see Frankel, 1999). Impossible trinity, because a country must give up one of three goals: exchange rate stability, monetary independence (useful to cope with slumps), or financial-market integration. It cannot have all at the same time. Countries can attain only two of the three goals simultaneously:

— a full float allows integration and monetary independence.

These trade-offs may appear to be an embarras de richesse. The many poor developing countries which are not yet on the radar screen of foreign investors and where domestic residents do not place their money at home anyway, are not concerned. But this comes at a large cost of forgone benefits of foreign flows which a recent study at the Centre has shown to be particularly significant for direct and portfolio equity investment (Reisen and Soto, 2001). Hence the focus of this paper is not only on emerging markets, but on those among the least developed countries which want to escape that condition by acquiring a financial reputation and therefore entering the radar screen of international investors.

As Guillermo Calvo rightly states in chapter 3 of this volume, there is widespread "Fear of Floating" in emerging markets. Governments in a typical emerging country can hardly tolerate massive overvaluation or undervaluation of their currencies. Calvo provides as reasons for fear of floating the transmission of inflation and devastating balance-sheet effects. We add to this list the growth disruptive effects of prolonged periods of misalignment (such as in New Zealand in the 1980s) which threaten growth strategies based on diversifying exports away from traditional crops towards non-traditional industries (see, e.g. Joumard and Reisen, 1992).

Whether hard pegs are a better alternative for open economies depends very much on institutional and regulatory prerequisites and on their degree of endogeneity with respect to the exchange rate regime (Eichengreen, 2000). These can be summarized as follows:

This is a long list which is not easily met. In order to see why one should not take for granted that these conditions will be met (somehow endogeneously), we will review in the next two sections what we take as two cases that bear their note of caution: the CFA experience and the Argentine currency board. In addition to Section II.1, where the EMS experience is brought to bear on this issue, chapter 2 by William Branson deals with the Pound sterling entry into the EMS and with the monetary preconditions for EU accession by EU applicant countries. Chapter 4 by Brigitte Granville deals with broader issues of exchange rate regimes in transition countries.

Agnes Benassy-Quere and Benoit Coeure (2000) have recently stressed the regional dimension of the debate on corner solutions. They argue that both pure floating and hard pegs make future regional cooperation more difficult. This is important in a world of regional trade blocs which look for ways to intensify cooperation. A float is an inherently unstable regime for countries competing on world markets for a similar range of products and hence sets incentives for beggar-thy-neighbor competitive devaluation. Floating induces non-cooperative strategies, especially when the competing neighbors face a common shock. Hard pegs are hard because it is so difficult to reverse them and because they lack an exit strategy. They are thus only suited for countries which aim at joining a monetary union with the anchor currency in not too distant a future (such as some countries in Central Europe). On the other hand, the perspective of joining or creating a monetary union can make intermediate regimes more robust in the mean time, as Section II.2 demonstrates.

With increasing financial integration, emerging markets may thus opt to give up on some exchange rate stability and on some monetary autonomy. There is ample choice on currency regimes inside the corners, although most of the regimes will amount to some sort of inflation targeting. Sebastian Edwards (2000) lists the range of regimes from 1) free float (with the exchange rate determined in the market alone) to 2) floating with a ‘feedback rule’ (indirect intervention which does not result in changes in reserves), 3) managed float (with intervention resulting in changes in international reserves), 4) target zone (floating within a band, with the central parity fixed), 5) sliding band (with adjustable central parity), 6) crawling band (backward- or forward-looking change of central parity), 7) crawling peg (passive or active adjustable peg), 8) fixed, but adjustable peg, 9) currency board, and 10) full adoption of another country’s currency. Fischer (2001) excludes fixed, but adjustable pegs and narrow band exchange rate systems as not viable in countries open to international capital flows.

Fiscally disciplined Southeast Asia succeeded for quite a while (1978-96) to reconcile stable exchange rates, low inflation and massive capital inflows without resort to capital controls. In the absence of developed money markets, the Southeast Asian central banks extended the open-market sterilisation instruments common in industrial countries through the use of public institutions such as social security funds, state banks and public enterprises as monetary instruments (Reisen, 1993b). But, from the mid-1990s, the regime ceased to keep the real effective exchange rate stable, and gradually turned into a dollar peg, as demonstrated by Benassy-Quere and Coeure (2000).

Further, exchange rate target zones with little intra-marginal intervention and moderate width have been persued quite successfully in Chile, Colombia and Israel in the early 1990s (Williamson, 1996), despite a large degree of financial openess in these countries: their crawling band help to achieve the trade-off between the conflicting objectives of reducing inflation and maintaining export growth. That they were given up in all cases, needs some explanation. A possible theoretical rationale is the complexity of basket pegs with bands, which hampers their verifiability, but is nevertheless needed for credibility (Frankel, Schmukler and Serven, 2000). In Section II.2, we argue that, once the effectiveness of the MSF is verifiable, there will be greater tolerance for intermediate regimes, so that the argument that they are "too complicated for locals and for Wall Street" need not apply.

Even when monetary unions survive, they can be harmful to growth when the effective (trade-weighted) exchange rate gets out of line, which can happen to a dollarised Latin America as it has happened to the CFA franc zone in Africa, which was pegged to the French franc since 1949. When the French franc started to appreciate in the mid 1980s against the US dollar, but at the same time a prolonged fall in commodity prices reduced the CFA zone terms of trade, the CFA zone countries recorded low growth relative to the rest of Africa, increased current-account deficits and a build-up in external debt. Alternative adjustment measures were unable to rectify the imbalances, and 1994 saw the end of the longest peg in memory: a 50 per cent devaluation of the CFA franc against the French franc. Table 1 presents a very clear-cut summary of what happened up to the massive devaluation in January 1994 and thereafter.

The CFA countries had grown by 3% more rapidly than their non-CFA counterparts in the decade 1975-85. They achieved this with (mildly) lower inflation and a larger current account deficit. This came at a time when the terms of trade were extremely favourable. Then came the bad times: in 1986-93 growth was 3% lower in CFA countries, in effect canceling all the benefits of the previous years. The fact that inflation was absent in CFA countries can be taken as evidence of a deflationary pressures of the price level, which begged to be solved by devaluation. The good years had created an overvaluation of the currency, which became unbearable when the dollar fell with respect to the franc. Although fiscal imbalances were indeed larger compared to the non-CFA performance, it is hard to see that such might have been the critical factor.

From this very brief overview of the CFA record, it is fair to say that an irrevocably fixed exchange rate system is a dangerous bet. On the one hand, it needs a fiscal straightjacket that is credible, which is not easy to achieve. And yet, even when the fiscal stance does not seem to be the critical problem, the typical emerging economy remains too volatile to fix its exchange rate to a given level.

Despite the difficulties of the CFA countries to cash in the benefits of their integration to the Franc, could one make a better case when the issue becomes one of locking in African currencies to the euro, or as an alternative to the dollar? Table 2 (from Cohen, Kristensen and Verner, 1999, hereafter CKV) shows the diversity of exchange rate regimes in African. On the face of it, it seems that a good case towards unifying these exchange rate regimes into one monetary agreement, say euroization or dollarization, could be made. It would boost, one would hope, intra-African trade and reduce the volatility which is inherent to these piecemeal agreements.

In favor of euroization, one can point to the fact that some 43 percent of merchandise imports to Sub-Saharan Africa originate from the European Union. EU countries account for 24 percent of all merchandise exports from Sub-Saharan Africa. For the CFA countries the share is already much larger, at 40 percent, one-quarter of it going to France (Table 3, from CKV).

These numbers however are misleading, if one fails to take account of the fact that most Sub-Saharan African exports to EMU countries are almost exclusively primary commodities, such as cotton, fruits, nuts, fish, coffee, pearls, silver, platinum, and crude petroleum. Petroleum alone accounts for 35 percent of the value of Sub-Saharan African export to OECD countries. Since primary commodities, which generally have a low income elasticity, account for the greatest share of exports from Sub-Saharan Africa, any increase in exports to EMU countries induced, say, by economic growth in Europe will be limited. The medium-term impact on CFA members of a 1 percent increase in euro-area GDP has been estimated by IMF staff as a 0.6 percent increase in exports and a 0.2 percent increase in GDP (Feldman et al. 1999). In addition, most of these commodities are priced in dollars, which further reduces the stability benefits of a euro peg.

In other words, despite the significance of the trade relationship between African countries and Europe, the degree of economic integration is lower that it appears on the face of it. Pegging to the euro would make those countries vulnerable to fluctuations of the real exchange rate of the euro to the other currencies, in particular to the dollar. The same could be said of the opposite choice in favour of the dollar.. This is the very lesson of the CFA countries, which benefited first of the depreciation of the franc to the dollar, and then suffered from the opposite swing. If the euro were now to appreciate relative to the U.S. dollar, CFA countries with a relatively low share of exports going to the European Union would be the most affected. Among these are Benin (16.9 %), Togo (15.8 %), Senegal (14.7 %), Guinea-Bissau (14.4 %), and Gabon (12.8 %). Cameroon is in fact the only CFA country with more than half its exports going to the European Union (73 percent).

Trade is one dimension of the problem. Financial integration is another one. If African countries were to lock in their exchange rate regime to any one of the two leading currencies, what would be the likely financial implications? Could it raise the credit ranking of African countries, better their access to world financial markets, raise foreign direct investments (FDI)?

Let us start by the latter. The distribution of FDI is very unequal: South Africa and Nigeria alone accounted for 68 percent of FDI in 1997. A very large share of the FDI in Sub-Saharan Africa is either in natural resources or concentrated in South Africa where there is a significant domestic market. In case of euroization or dollarization, would a portfolio shift occur in favor of Africa? For the most part, Sub-Saharan African countries lack the basic features needed to attract foreign private investors (especially pension funds), such as a long history of macroeconomic stability, good credit ratings (by Moody’s, Standard & Poors, and others), and reasonably well developed domestic security markets and stock exchanges. Few Sub-Saharan African countries are rated by international agencies. Euromoney and The Institutional Investor which measure a broader range of countries, including those in Sub-Saharan Africa, rank them against each other in terms of risk (Table 4, from CKV). All Sub-Saharan African countries are ranked very low, except for Botswana, Mauritius, and South Africa. The general picture that emerges is of severely underdeveloped domestic capital markets, with limited access to private foreign finance, according to Euromoney, except for the same three countries that are ranked relatively better in terms of risk ratings. In particular, one sees that CFA countries are not treated in more favorable terms either in terms of credit rating or in terms of access to the world financial markets. This confirms the view that monetary agreement do little to improve the credit rating of a country or its access to capital markets. Good policies seem to work better, as exemplified by Mauritius ratings, for instance, against Cote d’Ivoire or Senegal.

In conclusion of this section, one then sees that African countries have no easy solution at hand. Their trade pattern are too diverse to make a one fits all formula attractive in terms of currency agreements. Even the CFA countries, despite their historical link to France, are not a clear case of strong trade integration with Europe. Furthermore, when analyzing the credit rating of these countries, or the level of financial integration to the world financial markets, there is no clear sign that these countries have performed better than the other African countries.

Currency boards, once designed as a monetary arrangement for British colonies and then disused as countries gained political independence, have been back in fashion (until?) recently. Currency boards now exist in Argentina, Bosnia, Bulgaria, Estonia, Hong Kong, and Lithuania. They consist of exchange rates which are strictly fixed, not just by policy but by law. Domestic money can only be issued when it is fully backed by foreign exchange, removing monetary policy discretion from the government and the central bank.

Supporters of currency board arrangements have stressed that the regime provides credibility, transparency, low inflation, and financial stability in countries where the central bank is unable to pre-commit to a low rate of monetary growth. While the traditional reason for the time inconsistency problem of monetary policy has been an employment creation motive, the desire to inflate away nominal debt and to strengthen external competitiveness have been considerations of greater importance to developing and emerging-market economies. Currency board supporters have argued that a particularly important feature of that regime is to lower and stabilise domestic interest rates, by reducing devaluation and default risk and by reducing countries’ exposure to speculative attacks. Low and stable interst rates, in turn, should encourage investment and growth.

Such claims have been validated by the historical track record of currency boards. Atish Ghosh, et al. (2000) find that countries operating under a currency board arrangement have experienced lower inflation than either those with a floating or simple-peg regime, reflecting both a discipline effect (lower rate of money growth) and a credibility effect (higher money demand growth). The authors find also that better inflation performance has not been bought at the expense of lower output growth, although they concede that this might be due to a rebound effect from the typically depressed output levels before a currency board was adopted.

Argentina provides one of the most-debated cases of a currency board regime. In April 1991, after a long history of macroeconomic mismanagement and two episodes of hyperinflation, the currency board started to operate, with the peso pegged to the US dollar at one. The regime is based on the Convertibility Law passed in March 1991 by Congress, which grants the dollar legal tender status, and was subsequently supported by comprehensive deregulation of the economy and the full liberalisation of the current and capital accounts of the balance of payments.

Argentina’s regime features some notable design elements that represent a deviation from a strict currency board. These elements were introduced to accommodate the loss of a lender of last resort which a currency board entails and which exposes the country to financial crises with unsufficient provision of liquidity; this in turn requires strong and liquid domestic banks. First, the currency board is integrated into the central bank, there are no designated currency board accounts. Second, currently 20% of the money-base cover can be provided in the form of dollar short-term Argentinean public debt, rather than through international reserves. Third, the Argentine system is characterised by demanding capital requirements and a series of liquidity provisions. Banks are obliged to hold 21% of all deposits in liquid international reserves at the Central Bank or at Deutsche Bank New York. The Central Bank has also a contingent line of credit with a dozen international banks covering 10% of deposits in the banking system..

As seen from Figures 2 and 3, Argentina’s economic performance has been mixed. The short-term contribution of the currency board to Argentina’s economic performance was undoubtedly positive. The board arrangement provided a linchpin for deep reform of a very distorted economy and helped bring inflation down quickly. Inflation and interest rates came down quickly, supporting rapid GDP growth (which was helped by idle capacity). Argentina’s currency board has survived two major financial crises, with contagion from Mexico in 1994-95 and from Brazil in1999, not least because of its strong bank regulatory system. Bank regulatory policy promoted privatisation, financial liberalisation, free entry and proper risk management by banks (Calomiris and Powell, 2001).

Source: Grandes (2001)

Source: Grandes (2001)

Ultimately, however, the currency board system has neither delivered a sustained reduction in devaluation risk nor in sovereign risk. Growth, apart from recovery episodes after the adoption of the currency board and after Mexico’s 1994-95 crisis, has been low and volatile; investment and employment creation have remained anemic. The failure of the currency board system in Argentina to deliver further reductions in risk premia and to stimulate investment, growth and employment can be traced to insufficient fiscal discipline, an overvalued real effective exchange rate, and to the disincentives for savings promotion due to heavy liquidity requirements in the banking system. Let us discuss in turn:

Argentina’s currency board arrangement has ceased to confer sufficient fiscal discipline and the consolidated public-sector deficit has been gradually rising from 1995 on, culminating at 4.1% of GDP in 1999. This has gradually set in motion a vicious cycle of rising country risk premia and depressed growth, in turn fuelling the public deficit through lower tax receipts and higher debt service cost.

In a simulation exercise for Argentina, Martin Grandes (2001) demonstrates the strong endogeneity of these variables. In a forecast variance decomposition, he finds that 40 to 60% of the variance of the seasonally-adjusted fiscal deficit, seasonally-adjusted output growth and the sovereign risk premium can be explained by a shock to these very variables. While these findings may confirm the hypothesis of hard-peg supporters that in theory super-fixed exchange-rate regimes can trigger off a virtuous cycle of lower deficits, lower yield spreads and higher growth, the cycle has in practice turned very vicious indeed. Deutsche Bank (2000), in a thorough study on debt sustainability in Latin America, recently concluded that "Argentina will have to close an underlying fiscal gap that extends beyond improved tax collections associated with growth (p.3)". The bank, which calculated the non-interest (primary) budget balance at 0.63% of GDP for 2000, saw a need for a further 2% of GDP improvement in the primary balance in order to stabilise the public debt to GDP ratio. By comparison, in 2000 both Brazil and Mexico, countries on an exchange rate float, showed sufficient fiscal discipline to ensure debt sustainability according to Deutsche Bank indicators.

Initial inflation inertia and ongoing nominal wage rigidity have implied real appreciation of the Argentinean peso, with attendant current account deficits and a recessionary impact on the economy. While disinflation undid much of the initial overvaluation during the 1990s, Brazil’s devaluation early 1999 has strongly impacted Argentina’s real effective exchange rate, an indicator for external competitiveness. In early 2000, real overvaluation of the Peso was estimated at between 7 to 17% according to estimates by Deutsche Bank, Goldman Sachs and JP Morgan (Edwards, 2000). Even if estimates of exchange-rate disequilibria have to be consumed with caution, the fact that influential investment banks issue such estimates cannot fail to damage the credibility of Argentina’s currency board regime. More importantly, the US dollar is an anchor currency for Argentina that is certain to destabilise her real effective exchange rate: just 8% of Argentina’s exports are directed towards the US. Business cycles in the US and Argentina have not been synchronised over the 1990s and, given the different structures of the two economies, are not likely to coincide for long.

Indeed, Domingo Cavallo - the architect of Argentina’s convertibility law - suggested that, based on growing confidence in the Argentine economy, an eventual currency union between Mercosul members might be based on a currency basket also including the euro (Financial Times of 17^th March 1999, "Cavallo says Argentina could float its currency"). The reaction of the markets to the proposal was negative, with a sudden increase in the currency premium (measured by the spread of local peso time deposit rates over local US dollar interbank deposit rates with maturities up to 2 months, see Schmukler and Serven, 2001).

Finally, to make up for the lack of the lender-of-last-resort function in a currency board (or in a fully dollarised system), high liquidity requirements are needed for the domestic banking system to withstand a drawdown of deposits in times of crisis. Just like any minimum reserve requirement, high liquidity needs drive an important wedge between lending rates, which are increased, and saving rates, which are lowered (McKinnon and Mathieson, 1991). Such a wedge obviously discourages both savings and investment. This again may support a vicious cycle of growth being constrained by low investment and foreign debt fuelled by the lack of local savings. This may go up to a point where exploding debt dynamics (driven by the differnce of debt cost over the growth rate) and rising default risk leave the country with just three options: exit from the currency board, default, or new foreign finance.

In early 2001, the IMF board approved a loan agreement to cover Argentina’s borrowing needs for 2001 (and beyond), with the Fund offering $13.7 bn , $6 bn from the IDB and Spain, and some $20 bn from private, including domestic, sources. The currency board arrangement had got another, perhaps the last chance, to prove itself right.

In March 2001, Domingo Cavallo was called to defend the regime he introduced a decade earlier. In mid June, the congress approved Cavallo’s plan to add the euro alongside the dollar in the peso’s peg . Interestingly, the initial reaction of the markets was again a sudden increase in the currency premium. Furthermore, the onerous liquidity requirements imposed on the country’s banking system were alleviated, while a financial transaction tax that had proved an effective tax raiser in Brazil allowed corporate taxes to be reduced and public accounts to be rebalanced. These policy measures rectified the essential elements that caused the vicious cycle of Argentina’ rigid currency board scheme. They should help stabilise effective exchange rates (hence halt the decline in competitiveness), and lower the wedge between saving and borrowing rates (hence stimulate savings and investment).

Meanwhile the need to restore fiscal balance was made even more acute by the rising spreads on Argentine debt instruments. A policy of zero borrowing in 2001, to be confirmed by the senate in late July, was deemed "impressive" by the US treasury - in spite of recurrent market concerns about whether long term debt sustainability was insured. Even under stable debt dynamics, a repetition of the run on bank deposits observed after Mexico’s crisis would interact perversely with the social and political unrest associated with strenuous fiscal adjustment. The well designed banking supervision scheme that has been put in place provides for greater transparency than usual in emerging markets but it does not, by itself, add more commitment to discipline banks than already exists (Diamond, 2001). In any event, if the zero borrowing fiscal policy does not succeed in stabilizing debt dynamics, a financial crisis might occur at a time when no new payments on the existing debt are due. This would be the cruel revenge of the fiscal straitjacket.

In short, one cannot properly addressing the benefits of a hard peg without answering first the question: where does the relevant financial instability come from? It may stem from the lack of credibility of governments with respect to fiscal sustainability and underlying inflation. Or from the extrinsic noise that arises from the booms and bust of financial euphoria. The question then is whether dollarization can fit the bill and protect the emerging countries from these risks.

As we saw above, the answer is not obvious. The local banks operating in dollar and making loans either to the public or the private sector. In case of a financial crisis, the risk that the local banks will lose access to the international inter- bank market remains intact. The (expected) optimal response of the authority in front of such risk is easily derived from the textbook: it is the suspension of convertibility. The risk of a credit crunch and of generalized default is therefore still alive.

Dollarization in itself does not protect the country against the risk of distrust geared by the threat of default and suspension of convertibility. In that sense dollarization may be a fast lane to import credibility, but long-term policy credibility cannot be imported, it has to be earned. As the domestic reforms that earn credibility abroad are often unpopular, the exchange rate regime is but one of the difficult choices to be made by emerging markets and developing countries.

If one thinks that the critical driving force behind such risk is the lack of fiscal discipline, dollarization can only be an answer if one believes that governments debt will be held in check, once they are denominated in dollar. The massive defaults on international bonds in the past show that this is not a generalised outcome.

One can argue that -to the least- the rest of the economy would be better insulated from government default. But this is not obvious either: regional banks and the private sectors are likely to hold government bonds, and it can actually become harder to differentiate default to foreigners and nationals, raising systemic risk rather than lowering it. One cannot avoid thinking that dollarization can only go well if some fiscal straightjacket is also provided. The question then becomes, taking for granted that semi-constitutional commitments to low deficit are feasible, whether it remains optimal to go all the way towards dollarization.

A second important policy lesson that we derive from both the CFA and the Argentine experience is the choice of a numeraire. Whenever the anchor currency reflects a shock, the endogeneity of structural variables is not high enough to prevent peggers becoming victims of asymmetric shocks. This suggestion of the optimal currency area literature carries the proviso, due to Jeff Frankel and Andy Rose (1996), that the criteria for the choice between floating and pegging the currency are in fact endogenous (we return to this in part II). The criteria suggest that countries which are seen to be small (with the nontradable sector negligible) and open (in terms of trade shares in GDP) would be advised to peg. Such countries will exhibit a high degree of regional concentration of trade towards the country that could provide a potential monetary anchor and they would also face shocks similar to the country with the numeraire currency. Varieties of the Barro-Gordon model in open economies would suggest a peg for countries with a bad record of abusing monetary discretion. These countries can reach a lower inflation equilibrium if they can credibly commit to sustain the peg with the anchor currency that enjoys a better reputation than the local currency. A peg becomes hard to the extent it makes exit hard.

The foregoing discussion of the CFA and Argentine experiences showed that a hard peg, or dollarisation, does not automatically bring about long term policy credibility. The importance of the choice of the numeraire was also noted, and indeed reference was made to euroisation and to basket pegs. This might suggest that the creation of a European single currency was equivalent to a hard peg. While the euro certainly reflects the "don’t fix, don’t float" conundrum, we believe it must be interpreted as the outcome of a process of convergence among EU members sharing a common MSF. In this regard, the euro is rather a case of "float in order to fix" and more portable outside the euro zone than generally thought. This is what we proceed to show in Part II. First we describe the experience of the EMS, then we embed it in a view of European integration based on "peer pressure" and suggest that this is a way to earn credibility that is accessible to emerging markets world-wide, interested in acquiring financial reputation.

After the demise of the Bretton Woods system of fixed but adjustable exchange rates in 1973, various continental arrangements to stabilize exchange rates were tried, the last of which was the EMS, created in 1978 by a Resolution of the Council of Finance Ministers of the Union (EcoFin) and supported by an agreement among participating central banks.

The primary objective of the 1986 Single European Act was achieving free trade in goods and services and assets, as well as free movements of people among twelve nation-states. In turn the abolition of internal borders created market pressure for stable exchange rates in terms of the European Currency Union (ECU) basket.

Most member states changed their economic regime towards fiscal discipline and stable prices after realizing that they could no longer improve export competitiveness by engineering exchange rate devaluations. This followed the creation of the ERM, with the Netherlands foregoing devaluation after 1982 and France after 1983. Poorer states took longer to be convinced but - as discussed next - the economic regime did change in Ireland after 1987, in Spain after 1989 and in Portugal after 1992.

The EMS functioned without any realignments after January 1987 and the progress towards the single currency accelerated. At the Madrid European Council in 1989, the report of a Committee of Central Bank Governors chaired by the President of the European Commission (EC) was accepted as a basis for EMU. The single currency was to be achieved in three stages, beginning July 1, 1990. Rather than relying on national reserve currencies, a new currency was chosen, the ECU (it was renamed euro at the Madrid European Council of 1995).

Over the years, a code of conduct built up as the ERM developed from a mere exchange rate arrangement into a powerful convergence instrument. In addition to compulsory intervention, for unlimited amounts, at the agreed bilateral limits and to the need to reach a consensus for modifying a parity; the Basle-Nyborg Agreement called for convergence to establish and maintain stable exchanges rates. The rules also refer to the creation of ECUs through swap operations, to the provision of currencies for intervention purposes, to the settlement of claims arisen from intervention, and so on.

The ERM code of conduct implied the acceptance of the DMark as the anchor of the system and thus the recognition of the leadership of the Bundesbank. It also involved a consensus on crisis management.

Shortly after the first stage of EMU began, the United Kingdom joined the ERM. Sterling appeared to trade its past allegiance to the broad Atlantic standard for a narrower continental bloc. This first experience lasted less than two years but it involved the United Kingdom in the design of a MSF which turned out to be decisive for the sustainability of the system.

With respect to the MSF, the rules of the game changed for all the member states including the UK. Without such a broad base, the MSF would have been less credible and might consequently had hindered the governance of the eurosystem. This emphasis on the MSF also reflects the finding, going back to Cooper (1968), that the costs and benefits of cooperative responses to growing interdependence depend on the coordination of domestic institutions.

The plans for the single currency were agreed upon at the Maastricht European Council in late 1991. They were conditional upon convergence and cohesion, as explained in Macedo (2001b). The second stage of EMU was set to begin on January 1, 1994. The third and final stage of EMU was to begin after the 1996 revision of the Treaty signed at Maastricht if convergence was sufficiently high, and in 1999 if not.

A medium term orientation of macroeconomic policy, coupled with measures designed to improve the functioning of factor markets and of the public sector, is favored in principle. In transition and developing economies, though, the institutional framework for such an orientation is lacking, so that the rules for monetary stability are not credible. The expectation of EU membership, under conditions of convergence and cohesion, provides this credibility. In contrast with the Asian and Latin American experiences mentioned at the outset, candidates for membership have not been willing to set up a MSF among themselves, even when there exists an institutional vehicle like CEFTA (Macedo, 2000a).

Yet this is the way in which geographical peripheries can acquire global reputation. In a sense, they overcome the cost of physical distance through financial proximity. Of course initial and terminal conditions matter as much as the capacity to transform. Doctrinal controversies often reflect different assumptions about each one of these three factors. The principle of a stability oriented policy based on the respect of property rights and open markets goes back to the gold standard, and reflects "rules of good housekeeping" valid at the core and at the periphery (Bordo and Rockoff ,1995; Macedo, Eichengreen and Reis, 1996).

For EMU, "sustained regime change" was identified in EC (1990, chapter 9) as a condition for benefits to accrue to peripheral nations or regions. This argument was especially strong under the limited labour mobility and flexibility, coupled with low fiscal redistribution among states, which prevails in the European economy. In these circumstances, exchange rate adjustments may become necessary to eliminate declines in competitiveness but they may not succeed in changing relative prices. The greater the underlying capital mobility and the more likely the repetition of exchange rate adjustments, the less effective a devaluation will be.

EC (1990) also used survey data to suggest that firms did not expect devaluation to solve their problems but rather thought that credit constraints were a more severe hindrance to expansion at the peripheries than at the center. The fear that restrictions on fiscal policy called for by the excessive deficit procedure (EDP) contained in the treaty and later by the Stability and Growth Pact (SGP) would hurt growth and prosperity was addressed in Buti et al (1997), who showed that the retroactive application of the SGP would not have exacerbated recessions over the 1961-97 period.

Yet enhancing price and exchange rate stability and buttressing the soundness of public finances is a formidable task in countries with histories of high inflation, where neither the social partners nor public employees automatically appreciate the benefits of the regime change that the policymakers are attempting to engineer.

Errors in policy appraisal can unduly raise the costs of reform, when information about the change in regime is not readily available to international financial markets. Repeated market tests of the authorities' commitment to exchange rate stability may result from this imperfect information. If these tests of the authorities' resolve greatly increase the cost of defending the exchange rate, they can lead to policy reversals. Conversely, if the volatility of the exchange rate is a direct consequence of system turbulence, market tests will be short-lived and the threat of a reversal will become less and less credible, both abroad and at home.

Since its meeting in Brussels in late 1993, the European Council has been issuing "broad guidelines" against which policy and performance in the member states are to be gauged in what has become a regular test of the MSF for all EU member states. The progress of policy reform stands on how effective this MSF might be among union officials whose interaction with national officials should be accountable in their respective parliaments and in the European parliament.

The time it takes for a nation to acquire a reputation for financial probity varies but it typically involves several general elections where alternative views of society may confront each other. The number of years most frequently cited in financial circles is 10. This suggests that it may be better to take time and set on foot a self-reinforcing process of reform than to attempt a succession of overly ambitious and excessively drastic measures that will ultimately fail, damaging policy credibility.

To construct a social consensus domestically, credible signals that the authorities are committed to reform may be needed. If stable democratic governments succeed in implementing reforms which help to achieve convergence between poorer and richer nations and regions, they can set off a self-reinforcing virtuous cycle of stability and growth. On the other hand, there will be a vicious cycle if short-lived governments, fearing the social conflicts associated with reforms, delay implementation and impair convergence. The cases of Spain, Greece and Portugal discussed in Bliss and Macedo (1990) confirm both the need for the regime change and the difficulty in bringing it about without overcoming the resistance of vested interests.

The initial conditions a government inherits may limit the alternatives at its command. For example, the 1992-93 recession aggravated the plight of Europe's unemployed, making it more difficult to reduce the generosity of unemployment benefits. At the same time, by demonstrating the costs of labour market rigidities and the importance of competitiveness at the firm level, the experience of the 1992-93 recession may have actually encouraged structural adjustment and, ultimately, cohesion.

In the Spring of 1992, all Community currencies except the Greek drachma were in the ECU parity grid. But the Atlantic dimension was very weak. Even in the presence of sterling, the continental bloc continued based on the Dmark. Indeed, it included currencies of countries in the European Free Trade Association which were to become members of the union, like Finland and Sweden. In September 1992, sterling and the Italian lira left the ERM. Until August 1993, political instability and speculative attacks on the grid interacted with a severe recession and with the highest unemployment the Community ever witnessed.

The currency crises threatened the reputation for financial stability in small national markets, to the extent that national policies became less relevant than the proximity to a turbulent large market. Examples of this effect of "geographic" rather than "economic" fundamentals on the value of currencies were provided by Portugal and Ireland, who suffered currency attacks based on what was happening to the Spanish and British currencies. The attacks were short-lived but they nevertheless led Ireland to request a realignment in January 1993 and Portugal had partly to follow several realignments of the peseta.

Using the case of the Portuguese escudo and a technique of analyzing changes in the variance of the exchange rate which was first applied to the US stock market, Macedo, Nunes and Covas (1999, MNC hereafter, updated in Macedo 2001b) report weekly probabilities for the period preceding the widening of the bands in August of 1993. The period begins with the last entry into the ERM, which involved the escudo itself in April 1992. This includes some of the realignments involving the peseta and the escudo and the shifts from high to very high volatility show the incidence of a currency crisis. The few instances of very low volatility in the sample, which show instead massive intervention by the central bank shortly before the November 1992 realignment.

In 1993, on the contrary, there is no instance of this "artificial stability", suggesting that the code of conduct had meanwhile been learned by the Bank of Portugal. Given that the financial reputation of the country was not fully established as the regime change was quite recent, this serves as an illustration of the power of the ERM code of conduct as a convergence instrument. The case of Portugal is one where the regime change was gradual because of the need to combine selling political stability at home and earning credibility abroad (Macedo, 1996).

A related reason is that testing the ERM parity made sense when the real appreciation was perceived as excessive by export-oriented firms and the government may have been sensitive to their pressure.

The reason why the convergence process was not hurt by the decision to widen the band was that external credibility, while necessary for medium term policy credibility of any nation-state, is never sufficient. This was again apparent in the turbulence in early March 1995, which led Spain to ask for a new realignment in spite of fairly sound fundamentals. The lack of political stability was undermining the confidence in the currency.

As it turns out, after this realignment, Italy joined the ERM in late 1996 and Greece followed in early 1998, consolidating their own regime changes. At one time or another, therefore, all of the EU member states followed the ERM code of conduct. Austria joined with accession in 1995 and Finland in October 1996. Sweden shadowed the ERM before the 1991 banking crisis.

The single market for financial services, established in 1993, also built on the operation of the ERM code of conduct. In effect, the gradual acceptance of stability oriented policies is at the heart of such code of conduct. This is why it remained valid after the widening of the bands, even though the obligation of compulsory intervention for unlimited amounts at the agreed bilateral limits was unlikely to be applied.

The need to reach a consensus for modifying a central rate remained, as the parity grid was not changed by the decision to widen the fluctuation bands. It is also noteworthy that the economic priorities of the Treaty (low inflation, sound public finance, medium-term stability framework) remained undisputed among member states and Community institutions, stressing the need for convergence to establish and maintain stable exchanges rates.

During the second stage of EMU, the MSF designed to ensure convergence of national economies towards price stability and sound public finances became binding. The EDP, in particular, determined whether or not a member state could adhere to the single currency. Since convergence was not achieved in a majority of national economies, the EcoFin Council approved the SGP to ensure that the entry conditions for EMU would continue to be met after the euro was created and set the beginning of the third stage for 1999.

With high capital mobility, exchange rate stability requires a speedy real and nominal convergence process. The indicators of budgetary discipline have become signals of regime change sustained by the structural reform of the public sector.

Given that financial markets tend to exaggerate rather than to dampen such signals, apparent reversions during a relatively rapid convergence are also more liable to misinterpretation. The cohesion objective involves a degree of social awareness that may not be required with respect to the convergence of fiscal variables. In any event, whatever the credibility of national policies, it became apparent during the first stage of EMU that fast convergence was more difficult with slower growth. Moreover, during the transition, the main macroeconomic costs arise before the main microeconomic benefits are felt.

The Treaty convergence criterion relating to exchange rate stability requires the observance of the "normal fluctuation margins" during two years, and not having been involved in any realignment during the same period (or at least not having initiated one). Maintaining the currencies within the parity grid is the result of more than intervention by participating central banks. It reflected the credibility of national policies especially in Germany, and also that of the entire EMS relative to the dollar or the yen.

If, in the final analysis, the exchange rate reflects the credibility of national policies over the medium term, it may do so with considerable noise if the entire parity grid is under attack. This is why little indication about the credibility of national policy could be gathered from the realignments which occurred during this period. Speculative attacks on more vulnerable currency parities will have more negative effects on the system if parities are already locked than if they continue to be flexible. Flexibility within a sufficiently wide band allows speculation not to be a one-way bet. That lesson was learned in the twelve months preceding August 2, 1993 when very wide bands of 15% replaced the normal fluctuation margins. The temporary nature of the move notwithstanding, these new "normal fluctuation margins" eliminated the need for exceptional measures, such as exchange controls, designed to deal with a protracted second stage of EMU.

As can be gathered from MNC, foreign exchange market turbulence began in late August, early September 1992 when dollar interest rates fell substantially. In the meantime, German short-term interest rates remained high. Pressures for wage increases increased the reluctance of the Bundesbank in acknowledging that a European wide recession was imminent. The policy conflict led to the exit of some currencies from the ERM and to speculative attacks against others. The attack of July 1993 (pictured in MNC for one bilateral rate) was so massive that an emergency meeting of the EcoFin Council including Central Bank Governors was convened and exchange rate fluctuation margins were broadened to 15% on each side of the parity.

The 15% wide band was not used by any participating central bank and margins of 2.25% were observed between the Dmark and the Dutch guilder. The basic difference relative to the previously normal fluctuation margin was the absence of one-way bets on parities. The external discipline provided by the grid no longer obtained and each central bank decided whether or not to intervene within the old fluctuation bands. Most decided to do just that, so that the convergence process was not hurt by the decision to widen the band.

After the EMS crises, successive emerging markets were similarly hit during 1997-99. The fact that few opportunities for testing the credibility of exchange rate parities were missed by market operators made the exchange rate regime as crucial a determinant of macroeconomic stability as fiscal, debt management and banking policy. The perception that pure floating was beyond most emerging markets, mentioned in part I, rationalised direct policy responses such as exchange controls, perhaps along the lines of the so-called Tobin tax on short-term capital movements. Bartolini and Drazen (1997) stress the credibility effect of capital account liberalisation. See also Dornbusch (1998). A more cautious stance can be found in Eichengreen (1999).

With respect to the exchange rate regime, and except for the Dmark, which was the anchor currency of the ERM, the former national currencies of the eurosystem could neither float nor credibly fix without the well defined institutional framework for multilateral surveillance provided by the 1992 treaty (including the introduction of the EDP and of the ESCB) and by subsequent adjustments like the SGP and the Euro Group.

The pound sterling and the Swedish krona follow an inflation targeting monetary rule which allows the exchange rate to float, but in fact the latter has been fairly stable against the euro. Calvo discusses in chapter 3 of this volume the similarities between hard pegs and inflation targeting in emerging markets. The question of credibility is different in EMU members because they are more used to follow a MSF.

Financial crises have caused hardship in individual countries, but to date they have not threatened globalization. Any tendency to draw back from world financial markets, as in Malaysia, has been isolated and temporary. The debate about Malaysia´s crisis is summarized in Edwards and Frankel (2001), with Dornbusch (2001) taking the position that the general presumption against controls remains.

The same presumption applies to the Chilean experiment, described in Reisen (1999). So as to revive the domestic stock market, Chile lowered its barriers in Spring 1998 to short term capital inflows (a tax called encaje) and set the tax rate to zero in the Fall for a few months. Even a country endowed with a relatively well functioning administration found it difficult to keep an exchange control geared to a long term objective when the environment became turbulent. The objective was to improve the composition of capital inflows towards long term instruments, especially foreign direct investment relative to short term flows which were considered more volatile. In any event, the rationale for the encaje was clear during the boom of the mid 1990s but ceased to apply afterwards, reinforcing the idea that such measures work temporarily and only if they are introduced in good times.

The problem is rather how to tailor macroeconomic and financial arrangements to the imperatives of globalization and in particular to limit the vulnerability of the domestic economy to the crisis problem. The complexity of this problem is evident in the exchange rate dilemma. It has become clear that exchange rate systems involving fixed but adjustable rates, bands and crawls are increasingly crisis prone. The alternatives are to float more freely, as such Latin American countries as Brazil, Chile, Colombia and Mexico and such Asian countries as the Philippines, Korea and Thailand have begun to do, or to adopt a hard currency peg, as in Argentina, Hong Kong and, more recently, Ecuador.

The crisis encouraged responses that would not have been possible in calmer periods. Bank restructuring which took place during the crisis may not have otherwise happened; as a result, debt structures are in better shape now than if countries had postponed reforms.

The lesson of the widening the ERM bands described in the previous section was that a MSF must go beyond exchange rate surveillance. Moreover, it is by allowing responses that would not obtain in calm periods that financial crises serve as co-ordinating devices. In effect, co-ordinated systems like the one implied by the ERM code of conduct are difficult to adapt to a world system without shared values, even when they refer to what is essentially a shared variable, the exchange rate.

The exchange rate regime is just one instance of needed improvements in financial architecture. Nevertheless, it plays a central role in the debate on the reform of the system of international relations and its main institutions, which for the most part were established in the aftermath of World War II. The portability of the European MSF on a broader scale presumes that the exchange regime is well defined.

Intermediate solutions between the pure float and the currency board are sometimes taken to be too complex to be credible but the credibility of an exchange rate regime cannot be judged in general. Since the best indicator of policy credibility is that a MSF exists and is effective, it is the MSF that determines the choice of an exchange rate regime. This is why a way out of the "don’t fix, don’t float" conundrum may be that you may have to float in order to fix.

As a consequence, the MSF that is emerging among Mercosul and ASEAN+3 members does not imply for the moment a single exchange rate regime among them. On the other side, the MSF in the eurosystem should not neglect trends in the current account, as argued by Decressin and Dysiatat (2000).

The adaptation of the ERM code of conduct to improving international financial architecture would also support the creation of new networks including major emerging markets, as long as they manage to enforce financial stability. This applies to the Financial Stability Forum and to the G-20, for example.

Can devaluations and exchange controls be co-ordinated at the regional or global level, to lessen their beggar thy neighbour character? Probably not without co-ordination mechanisms among monetary and fiscal authorities like the ones found in the EU. How precisely the MSF which evolved from the ERM code of conduct may apply to ASEAN or Mercosul remains to be thoroughly investigated. There has been no formal application to relations within CEFTA in part because the candidate countries are already greatly involved with the EU, whereas there have been recent instances of adaptation in Asia and in Latin America. The lessons from the crises in the ERM may also help design a new international financial architecture because they were overcome by the ERM code of conduct.

With the experience gathered during the first two years of the euro, a new code of conduct may be developing, which acknowledges the importance of international banking supervision. The potential costs, stability and magnitude of private capital flows to developing countries are an important criterion to assess current proposals to reform the functioning of the international financial system. In this regard, there have been proposals for regional fora, which could help the IMF improve its performance when exchange rate and banking issues are difficult to disentangle, as is more and more frequently the case. While this is certainly true, institutions of global governance have been essential in preventing the 1997/99 financial crises in emerging markets from becoming a 1930's style global depression and they continue to be needed in the future.

The usefulness of the euro for international financial architecture hinges on recognizing the role of peer pressure and yardstick competition as they have been applied for decades. The proposals for flexible European integration hinge on ensuring that regional economic and financial institutions are complementary to the global ones, IMF, World Bank and WTO.

The immediate effect of the 1997-99 emerging market crises is to underscore a lesson from the inter-war period: liberalization and globalization must be managed in order to face the threat of protectionist pressures, which could conceivably spread from the tariff escalation to non-tariff barriers like exchange controls. Therefore, containing financial instability means avoiding a relapse of protectionism while fostering reform in the international system. This will allow for a more effective regional and global response to threats of contagion of national crises. In short, it is widely acknowledged that globalization calls for better governance at national and global levels. Our interpretation of the European experience suggests that regional governance based on peer pressure may be a crucial ingredient for better governance at national and, indeed, global levels.

Defining traits of this early 21^st century are the move towards global and regional integration of finance, production and trade; the introduction of the euro in the major part of Europe and the corresponding benign neglect of fluctuations between the world’s key currencies; and the high incidence of currency and financial crises in a world of intense capital mobility. This is the background for the choice of an appropriate strategy for monetary integration and exchange-rate regimes in many countries throughout the developing world. Mainstream advice has recently favoured monetarist corner solutions to that choice: the international monetarist variety recommends hard pegs, the domestic monetarist variety pure floating.

We hold against that mainstream view. We show that corner solutions are not as crisis-free as is often maintained. Pure floating has at times led to costly misalignment of exchange rates, devastation of unhedged balance sheets, and inflation import – reasons why it is rarely practised in developing countries. Hard pegs have become more popular, but we visit two important cases – Africa’s CFA experience and Argentina’s currency board – which warn against underestimating the risks involved. In maritime terms, no sensible sailor drops the anchor before the boat stops moving.

While hard pegs often confer initial gains in credibility and hence lower capital cost, these gains can be ephemeral when not supported by a sufficient degree of institutional development and economic flexibility. Both Africa and Argentina became trapped by an inappropiate anchor currency, inappropriate as the anchor did neither reflect their trade directions nor their cyclical needs. As there are few currencies available to borrow credibility from, this lesson will not be unique: it suggests either a basket peg or, if a realistic option, to build rather than borrow credibility.

The prospect of regional integration invalidates corner solutions as non-cooperative (float) and costly to exit (hard pegs), but it revives the intermediate exchange-rate regime. The EMS experience shows that target zones plus effective codes of conduct, wide enough to allow for sufficient flexibility, can indeed confer sustained credibility as to avoid large misalignments and to reduce crisis vulnerability. What they need to achieve these objectives goes beyond the public perception that the central parity is consistent with lon-term fundamentals. Expectations need to be guided by mutually agreed and surveilled governance codes towards intensifying integration, based on visible progress in macroeconomic stability and regulatory reform.

The MSF has to be "owned" by, rather than imposed on, the countries concerned. It must therefore be supported by peer pressure and yardstick competition, both of which are built gradually. The "Eurocentric" approach to earning credibility on the way to monetary integration holds impressive successes in the former European periphery. This is why the practical operation of the EMS provides important lessons for authorities struggling to implement sustainable exchange-rate regimes to support economic convergence. These lessons are beginning to spread beyond the European continent, attracting the attention of reformist governments worldwide.

Bartolini, Luigi and Alan Drazen, "Capital Liberalisation as a Signal", American Economic Review, March 1997

Besley, Tim and Ann Case, "Incumbent Bahavior: Vote Seeking, Tax Setting and Yardstick Competition", American Economic Review, March 1995

Bordo Michael and Hugh Rockoff, "The gold standard as a good housekeeping seal of approval", NBER Working Paper No. 5340, November 1995

Bliss, Christopher and Jorge Braga de Macedo, editors, Unity with Diversity in the European Community, Cambridge: Cambridge University Press, 1990.

Branson, William "German reunification, the breakdown of the EMS, and the path to stage three" in David Cobdham, editor, European Monetary Upheavals, Manchester: Manchester University Press, 1994, pp. 16-28.

Branson, William " Intermediate Exchange Rate Regimes for Groups of Developing Countries", chapter 2 in this volume

Buiter, Willem and Anne Sibert, "Transition Issues for the European Monetary Union" unpublished paper, September 1997

Buti, Marco, D. Franco and H. Ongena , "Budgetary policies during recessions: retrospective application of the ‘Stability and Growth Pact’ to the Post-War period", European Commission, DGII mimeo, Brussels, March 1997

Calomiris, Charles and Andrew Powell, Can Emerging Market Bank Regulators Establish Credible Discipline? The Case of Argentina, 1952-99, in Frederic Mishkin, editor, Prudential Supervision, NBER, University of Chicago Press, 2001, pp. 147-191

Calvo, Guillermo, "The case for hard pegs in the brave new world of global finance", chapter 3 in this volume.

CEPR, Flexible Integration, Towards a more effective and democratic Europe, Monitoring European Integration 6, London, 1996

Cohen D., N. Kristensen and Dorte Verner "Will the Euro Create a Bonanza for Africa?" CEPR Discussion Paper No. 2304

Cohen D., "The Costs and Benefits of a European Currency" in Marcello De Cecco and Alberto Giovannini (eds). A European Central Bank? Perspectives on monetary unification after ten years of the EMS: Cambridge University Press, 1989, pp. 195-209.

Cooper, Richard, The economics of interdependence, New York: Mc Graw Hill, 1968

Decressin, Joerg and Piti Dysiatat, "Capital markets and external current account: what to expect from the Euro", IMF Working Paper, October 2000.

Deutsche Bank, Debt Sustainability in Latin America, DB Global Markets Research, December 2000.

Diamond, Douglas (2001), Comment on Calomiris and Powell in Frederic Mishkin editor, Prudential Supervision, NBER, University of Chicago Press, 2001, pp. 191-194

Dornbusch, Rudi and Douglas Diamond (2001), Comment on Calomiris and Powell (2001) "Capital controls: an idea whose time has gone", MIT, March 1998

Dornbusch, Rudi, "Malaysia's crisis: what was different?", MIT, January 2001

Edwards, Sebastian (2000), "Exchange Rate Systems in Emerging Economies", mimeo., UCLA

Edwards Sebastian and Jeffrey Frankel, editors, Preventing Currency Crises in Emerging Markets, Chicago: University of Chicago Press, forthcoming

Eichengreen, Barry, Towards a new international financial architecture a practical post Asia agenda, Washington IIE, 1999

Eichengreen, Barry (2000), "When to Dollarize", mimeo. UC Berkeley

Eichengreen, Barry and Jorge Braga de Macedo, "The European Payments Union and the evolution of international financial architecture", in Alexandre Lamfalussy, Bernard Snoy, Jérôme Wilson (eds), Fragility of the International Financial System – How can we prevent new crises in emerging markets?, Fondation Robert Triffin, forthcoming

European Commission, "One Market One Money: An evaluation of the benefits and costs of forming an economic and monetary union", European Economy No. 44, October 1990

Fischer, Stanley (2001), "Exchange Rate Regimes: Is the Bipolar View Correct?", www.imf.org

Franch and Japanese Finance Ministries Staff (2001), Exchange Rate Regimes for Emerging Economies, Discussion Paper prepared for ASEM Meeting, January 2001

Frankel, Jeff and Andrew Rose, "The endogeneity of optimum currency areas", NBER Working Paper No. 5700, August 1996

Frankel, Jeffrey (1999), "No Single Currency Regime is Right for All Countries or at All Times, NBER Working Paper No. 7338, September 1999

Frankel, Jeffrey, Sergio Schmuckler and Luis Serven (2000), "Verifiability and the Vanishing Intermediate Exchange Rate Regime", mimeo. Harvard University

Ghosh, Atish, Anne-Marie Gulde and Holger Wolf (2000), "Currency Boards: More than a Quick Fix?", Economic Policy 31, October 2000

Goldstein, Morris (rapporteur), Safeguarding prosperity in a global financial system: the future international financial architecture, Council on Foreign Relations, New York: 1999.

Glick, Reuven and Andrew Rose, "Contagion and Trade: Why are currency crises regional", CEPR Discussion Paper No. 1947, August 1998

Grandes, Martin (2001), External Solvency: Dollarisation and Investment Grade. Towards a Virtuous Circle?, forthcoming, OECD Development Centre Technical Paper

Granville, Brigitte (2001), Exchange Rates in Transition, chapter in this volume

Gros, Daniel et al Quo vadis euro? The cost of muddling through, Brussels: CEPS 2000

IMF (International Monetary Fund). 1998. Exchange Arrangements and Exchange Restrictions Annual Report. Washington, D.C.

IMF "The West African Economic and Monetary Union", Occasional Paper No. 170, 1998.

Jacquet, Pierre and Jean Pisani Ferry, Economic policy coordination in the euro-zone; what has been achieved? What should be done? London: Center for European Reform

Joumard, Isabelle and Helmut Reisen (1992), "Real Exchange Rate Overshooting and Persistent Trade Effects: The Case of New Zealand", The World Economy, 15.3., May 1992

Krugman, Paul Peddling Prosperity, New York: Norton, 1994

Macedo, Jorge Braga de, ‘Portugal and European Monetary Union: Selling Stability at Home, Earning Credibility Abroad’, in Francisco Torres, editor, Monetary Reform in Europe, Lisbon: Catholic University Press, 1996.

Macedo, Jorge Braga de "Portugal", in Kii Fukasaku and Ricardo Haussmann, editors, Democracy, Decentralisation and Deficits in Latin America, OECD Development Centre, pp. 191-200, 1998.

Macedo, Jorge Braga de, "Converging European Transitions", The World Economy, November 2000a

Macedo, Jorge Braga de, "Financial Crises and International Architecture: a 'Eurocentric' Perception," OECD Development Centre Technical Paper No. 162, August 2000b.

Macedo, Jorge Braga de, Statement to the Preparatory Committee of the UN Financing for Development Conference, New York, February 20, 2001 available at prof.fe.unl.pt/~jbmacedo/oecd/un.htm.

Macedo, Jorge Braga de, Luís Catela Nunes and Francisco Covas "Moving the escudo into the euro", CEPR Discussion Paper No. 2248, October 1999

Macedo, Jorge Braga de, Barry Eichengreen and Jaime Reis editors (1996), Currency convertibility: the gold standard and beyond, London: Routledge

Reisen, Helmut (1993a), "Integration with Disinflation: Which Way?", in Richard O’Brien (ed.), Finance and the International Economy:7, The Amex Bank Review Prize Essays in memory of Robert Marjolin, Oxford University Press

Reisen, Helmut (1993b), "South-East Asia and the Impossible Trinity", International Economic Insights, Institute for International Economics, Washington D.C., May/June

Reisen, Helmut (1999), After the Great Asian Slump: Towards a Coherent Approach to Global Capital Flows, OECD Development Centre Policy Brief No. 16, January 1999

Reisen, Helmut and Marcelo Soto (2001), "Which Types of Capital Inflows Foster Developing-Country Growth?’, International Finance 4.1., March

Schmukler, Sergio and Luis Serven, Pricing Currency Risk: Facts and Puzzles from Currency Boards, paper presented at the NBER Interamerican Seminar on Economics, July 20, 2001

Williamson, John (1996), The Crawling Band as an Exchange Rate Regime: Lessons from Chile, Colombia and Israel, Institute for International Economics, Washington D.C.

Williamson, John (2000), Exchange-Rate Regimes for Emerging Markets: Reviving the Intermediate Option, Institute for International Economics, Washington, D.C.

Source: IMF Occasional Paper

Source: IMF 1998 and Cohen et al. (1999).

Source: IMF, Direction of trade data base and Cohen et al. (1999).

a (or should it read "1"??)^. The maximum score is 5.00, which is obtained by most OECD countries.

Source: Euromoney and The Institutional Investor and Cohen et al. (1999)

The final product reflects the current and past work program of the OECD Development Centre, beginning with the invitation by François Bourguignon that I organise a panel on capital flows at the ABCDE in Paris in June 2000. There Daniel Cohen and Helmut Reisen acted as the discussants of the papers by Guillermo Calvo and Brigitte Granville which now appear as chapters 3 and 4 of this volume. I provided the synthesis along what I called "eurocentric" lines in Technical Paper nº 162, of August 2000.

We then decided to consolidate the comments and the synthesis into chapter 1 and to extend with a survey of intermediate exchange rate regimes. Visits to the Centre by William Branson during the first half of 2001 allowed him to complete what became chapter 2. The result reflects numerous exchanges between Paris and Princeton, against the background of the original contributions from London and Washington. During the intervening time, the "don´t float, don´t fix conundrum" lost popularity and the merits of intermediate solutions were increasingly recognised. Nevertheless, the experience of how a largely unwritten code of conduct by the European Monetary System in the early 1990s allowed its members to "float in order to fix" is not restricted to the cohesion countries which were then earning external credibility. Rather it is applicable outside Europe, from Latin America to Asia, and even to Africa.

Jorge Braga de Macedo