Impact of Dollarisation

This paper can be downloaded from the ECB’s website (http://www. ecb. int). 1 This paper has benefited from comments and suggestions by Oscar Calvo-Gonzalez, Arnaud Mehl, Francesco Mongelli, Georges Pineau, Adam Posen, Francisco Ramon-Ballester, Antonio Sainz de Vicuna, Michael Sturm, Pierre van der Haegen, members of the EBOPS committee and an anonymous referee. The authors wish to thank Andre Geis and Stefan Wredenborg for excellent research assistance. The views expressed in this paper do not necessarily reflect those of the European Central Bank. European Central Bank, 2004 Address Kaiserstrasse 29 60311 Frankfurt am Main Germany Postal address Postfach 16 03 19 60066 Frankfurt am Main Germany Telephone +49 69 1344 0 Website http://www. ecb. int Fax +49 69 1344 6000 Telex 411 144 ecb d All rights reserved. Reproduction for educational and non-commercial purposes is permitted provided that the source is acknowledged. The cut-off date for the statistics in this paper was 1 October 2003. ISSN 1607-1484 (print) ISSN 1725-6534 (online) CONTENTS 2 INTRODUCTION DOLLARISATION/EUROISATION – A REVIEW OF THE LITERATURE 4 4 CASES WHERE DOLLARISATION/EUROISATION WAS ABANDONED 36 36 36 36 36 37 39 1. Colonial regimes 7 7 2. Liberia 2. 1 Basic facts 2. 2 Economic developments over the long run 2. 3 Optimum currency area properties 3. Main lessons from cases of abandonment 5 15 RECENT CASES OF DOLLARISATION/ EUROISATION 1. Benefits and costs of dollarisation/ euroisation 2. Dollarisation/euroisation and the bipolar view on sustainable exchange rate regimes 3. Dollarisation/euroisation and optimum currency areas 4.

Summary and outlook: stability and integration as key issues in analysing the potential costs and benefits of dollarisation/euroisation 3 CASES OF SUSTAINED DOLLARISATION/EUROISATION 10 12 40 40 40 40 41 41 43 43 44 1. Overview of cases 2. Kosovo and Montenegro 2. 1 Basic facts 2. 2 Pre-euroisation economic developments 2. 3 Optimum currency area properties 3. Ecuador and El Salvador 3. 1 Basic facts 3. 2 Pre-dollarisation economic developments 4. Common elements and main lessons from recent cases 6 CONCLUSIONS 17 17 18 20 20 20 21 21 22 1. Overview of cases 2.

The endogeneity thesis and the Rose evidence 3. The European microstates 3. 1 Basic facts 3. 2 Economic developments over the long run 4. Panama 4. 1 Basic facts 4. 2 Economic developments over the long run 5. Remaining cases of sustained dollarisation/euroisation 5. 1 Basic facts 5. 2 Economic developments over the long run 6. Common elements and main lessons from sustained cases 3. 3 Optimum currency area properties 20 3. 3 Optimum currency area properties 45 47 50 52 4. 3 Optimum currency area properties 22 24 24 24 REFERENCES 5. 3 Optimum currency area properties 25 33

ECB Occasional Paper No. 11 February 2004 3 1 INTRODUCTION The adoption by one country of another country’s currency has its place in the history of the world economy, but had for a long time fallen out of fashion. This has changed in recent years, mainly owing to the currency crises faced by several emerging market economies in the second half of the 1990s. As eliminating the exchange rate through the outright adoption of another currency would eradicate the potential for a currency crisis, official and unilateral dollarisation/euroisation has become a common policy advice. In the meantime, five countries or territories have officially and unilaterally adopted other countries’ currencies in recent years 2 : in the western Balkans, Montenegro and Kosovo introduced the euro, while in Latin America, Ecuador and El Salvador set out to abandon their currencies in favour of the US dollar. Finally, East Timor effectively dollarised after gaining independence. Other countries have considered the adoption of a foreign currency, with Argentina the most prominent example. In Europe, unilateral euroisation was recommended to some countries in the western Balkans and in central and eastern Europe (Gligorov 2001; Bratkowski and Rostowski 2001; Begg; Eichengreen; Halpern; von Hagen; Wyplosz 2001). 4 Finally, the pros and cons of a unilateral adoption of the US dollar or the euro have been discussed prominently in international financial institutions and academic circles, mostly as a result of the emerging “bipolar” or “corner solution” view on exchange rate regimes. 5 The experience of countries and territories that have officially and unilaterally adopted a oreign currency remains under-researched. While there is a vast literature on policy implications for countries characterised by a high degree of unofficial dollarisation/ euroisation, 6 the costs and benefits of official dollarisation/euroisation have mainly been explored on theoretical grounds (see for example, the overview provided by Berg and Borensztein 2000). Empirically, the analysis of official dollarisation has been largely confined to the case of Panama (Edwards 2001), as most countries that have adopted another country’s currency are small and/or dependencies of the respective anchor countries. Cases of dollarisation/euroisation only became the focus of empirical research when Rose (2000) provided econometric evidence on the trade effects of a common currency, suggesting that two countries sharing a common currency trade far more with each other than comparable countries with different currencies. This research has had a strong impact on the dollarisation/euroisation debate, as it has been seen as evidence in favour of the endogeneity hypothesis of optimum currency areas (OCAs).

According to this hypothesis, the criteria stressed by the traditional OCA theory are endogenous, rather than exogenous to the exchange rate arrangement (Frankel and Rose 1998). Thus, supporters of the endogeneity hypothesis argue that countries considering dollarisation/euroisation do not have to meet the 1 For simplicity and in line with standard practice, the term “dollarisation/euroisation” is used throughout this paper as a general term for the adoption of a foreign currency.

Thus, it is not only used to characterise cases where the currency adopted is the US dollar or the euro, but also those cases where other foreign currencies are involved, e. g. the British pound or the Australian dollar. For simplicity, we refer to all cases of dollarisation/euroisation as “countries”. This does not imply the expression of any opinion whatsoever on the part of the authors or the European Central Bank concerning the legal status of any country, area or territory or of its authorities, or concerning the delimitation of its frontiers.

The abolishment of the peso in favour of the US dollar was openly discussed in 1999/2000 and in the wake of the currency board crisis at the end of 2001 (see Edwards 2002). See also the overview in Backe and Wojcik (2003). The Journal of Policy Modelling and the Journal of Money, Credit and Banking devoted special issues to this subject in 2001. The legal aspects of this exchange rate regime were discussed at the BIS/ CEMLA/MOCOMILA seminar on the “Legal Implications of Currency Boards, “Dollarization” and Similar Arrangements”, held in Mexico City in February 2002.

See, for example, the survey provided by Balino et al. (1999). Unofficial dollarisation/euroisation is usually measured as the share of foreign currency deposits in broad money. Bergsten (1999) uses the terms policy vs. de facto dollarisation instead of official vs. unofficial dollarisation. Moreover, it is rather difficult to collect relevant data for proper analysis as many of the cases of dollarisation/euroisation are not members of the IMF or World Bank, or have only recently dollarised/euroised. 2 3 4 5 6 7 4 ECB Occasional Paper No. 11 February 2004 1 INTRODUCTION Box 1

THE ECB’S POSITION ON UNILATERAL EUROISATION IN ACCESSION COUNTRIES Although this paper does not deal with hypothetical cases of euroisation, it may be worthwhile to recall the ECB’s position with regard to potential unilateral euroisation in EU accession countries. The ECB considers that the euro area represents a multilateral currency union formed by Member States of the European Union with common and shared responsibilities among its members. When forming Monetary Union, the EU Treaty specified certain economic and institutional criteria that have to be fulfilled by future Member States of he common currency area in order to safeguard its sustainability. Moreover, the Treaty provides that there has to be a Community assessment of the fulfilment of these criteria and mutual agreement on the appropriate exchange rates. This is why with regard to current and future EU accession countries, the ECB does not welcome unilateral euroisation, as such an adoption of the euro outside the Treaty process would run counter to the underlying economic reasoning of European Monetary Union.

In particular, it would undermine the process of convergence prior to the adoption of the euro. Unilateral euroisation would also imply circumventing the process of multilateral assessment of new members by current EU Member States and as such would be difficult to reconcile with the cooperative spirit of a community of fellow members (ECOFIN 2000; Duisenberg 2001; European Commission 2002). OCA criteria ex ante, following a lengthy process of convergence or integration with the anchor country before adopting this country’s currency.

Rather, these criteria would endogenously be fulfilled ex post once a common currency has been adopted. This paper adds to the literature in two ways: first, it provides a comprehensive review of all the main cases of dollarisation/euroisation. In particular, it includes all the European cases, which have so far largely been neglected. Second, it provides a systematic, theory-based analysis for each of the main cases. In doing so, it is guided by three questions: – Why did countries opt for dollarisation/ euroisation, and did this exchange rate regime meet their economic needs? Do these countries have certain economic characteristics that distinguish them from otherwise similar countries which have maintained currencies of their own? – Have dollarised/euroised countries pursued a certain set of policies that may explain their success in sustaining this exchange rate regime? The paper is structured as follows: Section 2 reviews the literature on dollarisation/ euroisation and provides an overview of the key criteria to which that literature makes reference in its discussion of the costs and benefits of this exchange rate regime.

The case study review has three parts: initially, the focus is on sustained cases of dollarisation/euroisation (Section 3), which is followed by an analysis of those cases in which dollarisation/euroisation was abandoned (Section 4). Finally, recent ECB Occasional Paper No. 11 February 2004 5 cases of dollarisation/euroisation are discussed (Section 5). 8 The main results, which are summarised in Section 6, are that policies and attributes – mainly exogenous to monetary policy – that foster integration with the anchor country have been crucial in supporting the exchange rate regime.

To this end, most dollarised/euroised countries have exploited advantages that are largely prior to the choice of exchange rate regime, namely their small size, geographic proximity to the anchor country, and politically dependent status. The findings of the paper suggest that recommending dollarisation/euroisation irrespective of countries’ ex ante degree of integration with the potential anchor country seems to bear considerable risks, as dollarisation/euroisation does not seem to be a straightforward substitute for integration.

Hence, despite its alleged merits as a device for achieving macroeconomic stability, countries should carefully consider the option of relying on a suitable domestic anchor for monetary policy before opting for unilateral dollarisation/ euroisation (Berg, Borenszstein and Mauro 2002; Detken and Gaspar 2003; Posen 2004). 8 To this end, substantial efforts have been undertaken to collect data and make them comparable across countries, but question marks over their reliability remain. 6 ECB Occasional Paper No. 11 February 2004 D O L L A R I S AT I O N / E U R O I S AT I O N – A R E V I E W O F T H E L I T E R AT U R E The prospective advantages and disadvantages of unilateral dollarisation/euroisation are mainly derived from two major theories, the “bipolar” or “corner solution view” of sustainable exchange rate regimes, and the theory of optimum currency areas, which exists in an “old” and a “new” version. As explained below, there is significant overlap between the latter two theories, as the arguments put forward by the “new” OCA theory, focusing on the issue of monetary policy credibility, are similar to the ones stressed in the bipolar view.

These approaches emphasise the response to externally-induced currency crises, the credibility of monetary commitments, as well as assuming limited virtues of monetary policy discretion, in their evaluation of monetary regimes largely based on market expectations. In contrast, the “old” or “traditional” OCA theory has a rather different focus, as it emphasises the need for an adjustment mechanism in case of asymmetric shocks and an unsatisfactory level of economic integration between the dollarised/euroised country and its anchor. In other words, it emphasises more the real side of integration.

The review of the literature is complemented by empirical evidence on indicators derived from these theories, assessing whether countries would actually qualify as potential candidates for dollarisation/euroisation. The evidence indicates that under the bipolar view and the new OCA theory there might be a substantial number of countries that could benefit on stability grounds from unilaterally adopting a foreign currency. In contrast, the indicators of the old OCA theory largely suggest that countries should keep their own domestic currency, as their level of integration with the potential anchor country is too low.

States and the euro area as legal tender and official currency, implying that the country chooses to abandon its own currency and the central bank to forego the monetary policy instrument. 9 There is widespread agreement on the main benefits and costs of dollarisation/ euroisation, which can be grouped as follows: A) BENEFITS Fostering macroeconomic stability: Dollarisation/euroisation is expected to foster macroeconomic stability by solving the credibility problem that arises when a domestic central bank is unable to pre-commit itself to a low rate of inflation (Barro and Gordon 1983; Goldfajn and Olivares 2000).

By explicitly adopting the monetary policy of the issuing country, which enjoys a high degree of credibility, inflation and interest rates in the dollarised/euroised economy are assumed to converge towards the level of the issuing country. If this credibility channel works, this convergence should largely avoid the output costs associated with disinflation in a low credibility environment. Moreover, dollarisation/euroisation is seen as one way to accept the logic of the “inconsistent quartet” concept, by relinquishing an independent monetary policy. 0 Finally, dollarisation/ euroisation is supposed to enhance a country’s fiscal discipline by eliminating the possibility of printing money to finance fiscal deficits (Fischer 1982; Eichengreen 2000). Lower risk premia: If risk premia are owing to currency and not to country risk, dollarisation/ euroisation should lead to lower risk premia, because a sharp and sudden devaluation of the 2 DOLLARISATION/ EUROISATION – A REVIEW OF THE LITERATURE 9 BENEFITS AND COSTS OF DOLLARISATION/ EUROISATION Dollarisation/euroisation is defined as the adoption of the US dollar or the euro by the authorities of a country outside the United Institutional specifications of dollarisation/euroisation regimes can differ in detail. For example, some countries have maintained their central bank after dollarising/euroising, while others have abolished it. Some countries, in particular those that had already dollarised/euroised back in the 19th or early 20th century, have never had a central bank.

Moreover, some dollarised/euroised countries still issue domestic coins. However, this merely has a subsidiary or symbolic role. 10 The concept states that the combination of unrestricted capital flows, openness to trade, a fixed exchange rate and monetary policy autonomy is incompatible. ECB Occasional Paper No. 11 February 2004 7 domestic currency against the anchor currency is ruled out by definition. 11 In particular, if a country is confronted with a currency mismatch (foreign currency borrowing by public or private sector entities without major oreign currency revenues), dollarisation/euroisation could eliminate this problem and therefore reduce the sovereign risk. 12 Moreover, assuming well-functioning international capital markets, dollarisation/euroisation is expected to improve a country’s access to those markets as a result of lower currency risks, higher financial sector stability, lower risk of sudden introduction of capital controls (Berg and Borensztein 2000) and lower information costs (Calvo 1999).

Domestic financial sector development: Dollarisation/euroisation is expected to support the development of a country’s financial sector, because a stable currency is a precondition for financial development (Hausmann et al. 1999; Berg and Borensztein 2000), leading to strong and steady economic growth. 13 Elimination of transaction costs: The elimination of costs of exchanging the domestic currency into the currency of the anchor currency is the most visible effect of dollarisation/euroisation (Fischer 1982; De Grauwe 2000), with the cost savings being proportional to the number of transactions conducted in foreign currency.

Stronger economic and financial integration: Dollarisation/euroisation is expected to foster a country’s economic integration with the economy of the issuing country (Frankel and Rose 1998; Rose and Engel 2000a; Dallas and Tavlas 2001). In particular, trade integration is likely to deepen owing to lower transaction costs and the elimination of exchange rate uncertainty, and under the assumption that trade is fairly liberalised. As a consequence of higher economic integration, dollarisation/euroisation might lead to real convergence in terms of GDP levels and convergence of business cycles with the issuing country.

Finally, shocks might become more symmetric between the dollarised/ euroised country and the respective anchor country, while business cycles might become more synchronised, in turn further fostering integration. B ) COSTS Loss of an adjustment mechanism: With dollarisation/euroisation, a country loses the use of the monetary policy instrument as a mechanism enabling it to adjust in the wake of asymmetric shocks and to react to fluctuations in the business cycle that are not in line with those in the anchor country.

Accordingly, the dollarised/euroised economy has to rely on other adjustment mechanisms to avoid substantial output swings owing to asymmetric shocks or unsynchronised business cycles with the issuing country. Loss of the lender of last resort function: With dollarisation/euroisation, the domestic authorities lose the ability to respond to a sudden run on bank deposits by acting as a lender of last resort.

In particular, the authorities are unable to inject an unlimited amount of liquidity into the payment system to prevent a default on deposits (Berg and Borensztein 2000), as the amount available to purchase bank assets and to recapitalise distressed financial institutions is restricted to the country’s stock of foreign reserves. Loss of seigniorage: The most direct cost of unilateral dollarisation/euroisation is the loss of seigniorage revenues from issuing a domestic currency, as these revenues will shift from the domestic monetary authority to the monetary 1 However, if the loss of the exchange rate instrument diminishes a country’s adjustment capacity to asymmetric shocks, dollarisation/ euroisation could increase default risk, thus contributing to higher risk premia (Goldfajn and Olivares 2000). 12 See Calvo (1999, 2001) and Hausmann (1999). Of course, dollarisation/euroisation does not preclude sovereign defaults resulting from an unsustainable fiscal position, unsound financial systems or political turmoil. 13 The positive correlation between financial development and economic growth has been studied intensively over the last few years.

For a comprehensive survey of the evidence, see World Bank (2001). 8 ECB Occasional Paper No. 11 February 2004 Box 2 COSTS AND BENEFITS OF DOLLARISATION/EUROISATION FOR COUNTRIES ISSUING A CURRENCY ADOPTED BY OTHER COUNTRIES Most of the debate on dollarisation/euroisation has focused on the question of whether it is desirable for the country adopting the foreign currency, whereas less attention has been paid to the question of dollarisation/euroisation from the perspective of the issuing country.

This may be explained by the fact that most costs and benefits for countries adopting the foreign currency appear to be readily transferable to the case of issuing countries (Altig 2002). For example, if dollarisation/euroisation facilitates trade and this leads to a higher degree of monetary and financial stability in the newly dollarised/euroised economy, then this might also benefit the issuing country.

Conversely, if the loss of the monetary policy instrument in the dollarised/ euroised economy has negative effects on stability and growth in the dollarised/euroised economy, there might be negative repercussions on the anchor country as well. A readily measurable benefit of dollarisation/euroisation for the issuing country would be seigniorage gains that correspond to the loss of seigniorage revenues for the dollarised/euroised economy, although these gains are likely to be relatively small. ) In cases of widespread unofficial dollarisation/euroisation in the form of currency substitution, the marginal gains would be even lower (Altig 2002). The main costs for countries issuing the foreign currency to be adopted are related to the possibility that in the event of a crisis in the dollarised/euroised economy, there will be pressures on the issuing central bank to accommodate shocks, even if there are no formal obligations to do so (Altig 2002, Alesina and Barro 2001a, Summers 1999). ) Finally, dollarisation/euroisation of third countries is less appealing to the issuing country when countries considering this exchange rate regime have a history of high monetary instability (Bayoumi and Mauro 2001), as this might have negative reputation effects on the issuing country and its central bank. a) Altig (2002) estimates the seigniorage gains (“flow” gains) for the United States to be between 0. 2 and 0. 8% of GDP per year if Mexico and the countries of South America were to adopt the US dollar as their domestic currency. ) The same pressure may arise with regard to the lender of last resort function (Schuler and Stein 2000). authority of the issuing country. 14 The loss of seigniorage includes both one-off “stock” costs arising from replacing the national currency in circulation with foreign banknotes and coins, and “flow” costs arising from the loss of the future earnings stemming from the flow of new currency printed every year. C) FACTORS AFFECTING THE BALANCE BETWEEN BENEFITS AND COSTS The most readily quantifiable benefits and costs of dollarisation/euroisation are the potential transaction cost savings and the loss of seigniorage revenues. 4 In principle, countries could agree to share seigniorage revenues. However, the only such arrangement that currently exists is between Lesotho, Namibia and South Africa. 2 DOLLARISATION/ EUROISATION – A REVIEW OF THE LITERATURE ECB Occasional Paper No. 11 February 2004 9 The potential for transaction cost savings as a result of the adoption of a foreign currency depends on the degree of trade openness of the respective economy and the share of trade that is invoiced in the foreign currency adopted. 5 It is a well-documented stylised fact that in general small countries exhibit a higher level of openness – as measured by the ratio of the sum of exports and imports to GDP – than larger economies, 16 indicating that transaction cost savings may benefit them the most. In general, however, transaction cost savings per se are not perceived as a major point in favour of dollarisation/euroisation. 17 Rather, the argument relies on trade-enhancing effects triggered by a reduction in transaction costs and the elimination of exchange rate uncertainty.

Costs related to the loss of seigniorage revenues are considered to be significant for many emerging market economies, particularly for those with large informal sectors, an underdeveloped tax system and/or high inflation. 18 However, a developed tax system and monetary stability are regarded as general policy goals irrespective of the exchange rate regime. Thus, the proper benchmark for dollarisation/euroisation costs related to the loss of seigniorage revenues seems to be a situation of monetary stability.

Most empirical studies focusing on countries that issue a stable currency put these costs at around or below 1% of GDP (Klein and Neumann 1990, De Grauwe 2000, Schobert 2002). In summary, neither transaction cost savings nor the loss of seigniorage revenues are at the centre of the dollarisation/euroisation debate. Instead, the discussion on the benefits and costs of dollarisation/euroisation focuses on the usefulness of domestic monetary policy as an adjustment instrument in the face of asymmetric shocks. DOLLARISATION/EUROISATION AND THE BIPOLAR VIEW ON SUSTAINABLE EXCHANGE RATE REGIMES The increasing incidence and intensity of currency crises in the 1990s gave new impetus to the debate on the sustainability of exchange rate regimes. At the end of the decade consensus seemed to emerge 19 whereby, under conditions of an increasingly open capital account, soft pegs – the exchange rate regime that was most widely used by emerging market economies – are seen as unsustainable.

Hence, countries that have an open capital account or plan to liberalise capital flows should choose in favour of one of the two “corner solutions” – either flexible exchange rates or hard pegs, such as currency boards or dollarisation/ euroisation. 20 15 In most cases, data on the use of currencies for international trade settlement and invoicing in goods and services are not available. However, there are estimates putting the share of world trade invoiced in US dollars at 50% (McDonough 1997), which is much higher than the United States’ share of world exports.

One reason for this discrepancy is related to the fact that energy and raw materials are internationally priced in US dollars. While the existence of international pricing standards does not preclude, by itself, the use of another currency for invoicing and settlement, trade in energy and materials is almost exclusively invoiced and settled in US dollars. 16 See, for example, Easterly and Kray (2000). From a regional point of view, Asian countries are the most open economies, whereas trade, expressed as a percentage of GDP, is lowest in Latin American countries. 7 In the case of the highly integrated EU economies, the transaction cost savings of the single currency were estimated to represent a quarter to half a percent of EU GDP per annum (De Grauwe 2000; EC 1990). 18 There are different ways of calculating seigniorage revenues (Klein and Neumann 1990). Estimates for seven Latin American countries put the potential stock costs of dollarisation at around 4. 5% of GDP (in the period from 1991 to 1997), whereas potential flow costs were estimated on average at 2. 3% of GDP.

Costs varied substantially from country to country depending on the country’s inflation rate, and its degree of financial development and unofficial dollarisation (Bogetic 1999). Moreover, in many high inflation economies, monetary instability encouraged widespread currency and asset substitution, thereby severely limiting the opportunities for exploiting seigniorage on a sustainable basis (Schobert 2002). 19 See Frankel (1999), the International Financial Institution Advisory Commission (2000) and Fischer (2001a). 0 The collapse of Argentina’s currency board has reopened this debate, as it raises questions about the practical relevance of the “two corner” approach to exchange rate policy (Edwards 2002; ECB 2003). 10 ECB Occasional Paper No. 11 February 2004 Table 2. 1 Unofficial dollarisation/euroisation in selected countries Degree of unofficial dollarisation/euroisation 1) High (> 70%) Middle (> 20 %, < 70%) 2 DOLLARISATION/ EUROISATION – A REVIEW OF THE LITERATURE Lebanon Belarus, Russia, Vietnam, Ecuador, Tajikistan, Egypt, Mongolia, Moldova, Honduras, Ukraine, Yemen

Countries 2) Bolivia, Cambodia, Angola, Lao PDR Nicaragua, Croatia, Peru, Argentina, Turkey, Paraguay, Romania, Guinea-Bissau, Armenia Uruguay, Bulgaria, Costa Rica, Philippines, Zambia, Georgia, Mozambique, Lithuania, Sao Tome & Principe, Azerbaijan, Latvia, Congo, DR Source: Honohan and Shi (2002). 1) Expressed as a share of foreign currency deposits in broad money. Honohan and Shi present data for the years 1990-2000, whenever they are available. Note that Ecuador switched to unilateral, official dollarisation in 2000 (see Section 5).

The breakdown into high and middle unofficially dollarised countries is taken from the World Bank’s ”dollarisation” webpage. 2) In most countries, the US dollar is the main currency in which foreign exchange deposits are denominated. Only in some countries neighbouring the EU, in particular in the countries of former Yugoslavia, such as Croatia, are foreign exchange deposits widely denominated in euro (ECB 2002). The choice of the proper exchange rate regime is perceived to depend largely on the credibility of domestic monetary policy. 1 This is because countries with a low degree of monetary policy credibility are severely limited in their capacity to use monetary policy so as to reduce output fluctuations (Summers 2000; Calvo and Reinhart 2001) 22 or to act as an effective lender of last resort. 23 Thus, they cannot benefit from the gains a floating exchange rate is associated with. The experience of many emerging market economies even suggests that the use of domestic monetary policy might have had destabilising effects in terms of interest and exchange rate volatility (Del Negro et al. 001), amplifying rather than mitigating macroeconomic fluctuations and financial instability. 24 Conversely, the adoption of a foreign currency offers countries with low credibility a strong “commitment device” (Del Negro et al. 2001) that they can use to stabilise monetary and economic conditions, allowing for lower interest rates, supporting non-inflationary growth and fostering financial development. Thus, dollarisation/euroisation is regarded as a key policy tool that can put emerging markets on the road to monetary and financial stability (Calvo 2001). 5 The recent popularity of dollarisation/ euroisation can be partly explained by the fact that many developing countries and emerging market economies can be, or have been, 21 A different point against floating exchange rates is that emerging market countries tend to export commodities and/or light manufactures, making them more vulnerable to high exchange rate volatility (Edwards 2002). 22 For example, in a country with a high degree of unofficial dollarisation/euroisation indicating a lack of monetary policy credibility, adjusting the exchange rate may be of little use, as many wages and prices are tied to a foreign currency.

In this case, they will rise or fall by the same amount as the exchange rate (Frankel 1999). See, however, Cordon (2000) and Broda (2001), who argue that the real effects of exchange rate devaluations on output and employment are empirically significant, at least for larger countries. 23 The limits of monetary policy in acting as a lender of last resort in the case of a high degree of unofficial dollarisation/euroisation are spelled out in Jeanne and Wyplosz (2001). 24 This holds in particular under conditions of increasing international capital market integration. 5 This does not mean, however, that dollarisation/euroisation is seen as a tool for stabilising business cycle fluctuations. On the contrary, it ensures that countries cannot pursue classical stabilisation policies they would be incapable of implementing effectively owing to credibility problems. ECB Occasional Paper No. 11 February 2004 11 characterised by a low degree of monetary policy credibility owing to long periods of high inflation. Moreover, many of them seem to have failed in implementing more orthodox measures regaining credibility.

A particularly relevant measure of the lack of monetary policy credibility is the degree of unofficial dollarisation/euroisation, defined as the share of foreign currency deposits in broad money, as it reflects the response of private agents to currency instability, macroeconomic mismanagement and financial crises. Table 2. 1 provides evidence, collected by Honohan and Shi (2002), that shares of foreign currency deposits in broad money of more than 20% is not a rare phenomenon, but can be observed in many countries.

Thus, the evidence seems to suggest that under the bipolar or corner solution view, there would be a substantial number of countries that meet a key prerequisite for becoming a candidate for dollarisation/ euroisation. terms of credibility and stability gains are potentially high. Hence, as in the sustainable exchange rate regime literature, countries with low monetary policy credibility are considered prime candidates for joining another currency area by unilaterally switching to dollarisation/ euroisation.

Whereas the new view on OCAs stresses the benefits to stability, the old view highlights the integration prerequisites of dollarisation/ euroisation. This is because the cost-benefit analysis of the old OCA theory does not focus on whether a country can use monetary policy effectively to cope with asymmetric shocks and unsynchronised business cycles. Rather, it raises the question whether such a policy is needed. Thus, under the old view, the pros and cons of dollarisation/euroisation are not assessed by analysing the effectiveness of domestic monetary policy, but rather by evaluating the ffectiveness of other adjustment mechanisms and the level of economic integration (see Table 2. 2). The most important adjustment mechanisms that could substitute for monetary policy are price and wage flexibility as well as mobility of factors of production, whereas the degree of diversification in production and consumption is a good indicator for assessing the adjustment capacities and needs of the respective economies. Fiscal and political integration with the anchor country could represent an alternative as well, acting as a shock absorber.

Finally, a high degree of trade and financial 3 DOLLARISATION/EUROISATION AND OPTIMUM CURRENCY AREAS Dollarisation/euroisation not only constitutes an exchange rate regime, but also implies a unilateral link to an existing currency area. Thus, the dollarisation/euroisation debate echoes arguments put forward in the OCA literature. Indeed, the view that dollarisation/ euroisation represents a sustainable and appropriate exchange rate regime for countries with low domestic monetary policy credibility has much in common with the “new” view on OCAs.

This view, developed in the 1980s and early 1990s, is based on (1) the proposition of the long-run ineffectiveness of monetary policy, (2) the credibility issue, and (3) doubts as to the effectiveness of exchange rate adjustments (Mongelli 2002). Since low credibility implies low effectiveness of monetary policy and nominal exchange rate changes even in the short run, the cost of losing the monetary policy instrument is negligible in this view. At the same time, the benefits in Table 2. Properties of optimum currency areas – the “old” view OCA properties allowing countries to forego monetary policy as an adjustment instrument • • • • • • • Price and wage flexibility Mobility of factors of production Diversification in production and consumption Fiscal and political integration Openness and trade integration Financial market integration Business cycle correlation Source: Authors’ compilation, based on Mongelli (2002). 12 ECB Occasional Paper No. 11 February 2004 ntegration would lessen the need for any adjustment mechanism by lowering the probability of asymmetric shocks, leading to a highly correlated business cycle. In the following, these issues will be briefly discussed, providing some empirical evidence on indicators that try to capture the availability of and need for these alternative adjustment mechanisms. The main finding is that most indicators suggest that only a few countries might afford to surrender domestic monetary policy.

PRICE AND WAGE RIGIDITIES Price and wage rigidities, in industrialised as well as in developing countries, seem to be widespread. Indeed, if prices and wages were to behave in a textbook-like flexible price world, the issue of an optimum currency area, either in its old or new version, would be of little relevance (Obstfeld and Rogoff 1996). MOBILITY OF FACTORS OF PRODUCTION In general, inter-regional mobility of labour is rather limited (Mongelli 2002, Bayoumi and Mauro 2001, World Bank 1995). However, there are exceptions.

Data on migration and remittance flows suggest that a rather large share of the population of some developing and emerging market countries is working abroad, mainly in the US, the EU 26 and oil-producing countries in the Gulf region. Indeed, remittances have become an important source of external funding (World Bank 2003) for some small countries in the geographical vicinity of these three economic areas. Evidence from those countries for which data are available suggests that there are several countries financing more than one-fifth of total imports by means of remittances (see Chart 2. 1).

DIVERSIFICATION OF PRODUCTION AND CONSUMPTION A high concentration of exports in specific categories of goods and services is considered to be an argument in favour of flexible exchange rates, i. e. against dollarisation/euroisation, assuming that the potential anchor country is Chart 2. 1 Net remittances (as a percent of imports, average 1995-2000) 60 50 40 30 20 10 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 2 DOLLARISATION/ EUROISATION – A REVIEW OF THE LITERATURE 60 50 40 30 20 10 0 1 Yemen, Rep. 2 Albania 3 El Salvador 4 Samoa 5 Jordan 6 Cape Verde 7 Sudan 8 Dominican Rep. Egypt 10 Bangladesh 11 India 12 Ecuador 13 Morocco 14 Jamaica 15 Sri Lanka 16 Nigeria 17 Nicaragua 18 Pakistan 19 Guatemala 20 Turkey 21 Honduras 22 Portugal 23 Tunisia 24 Greece 25 Peru 26 Croatia 27 Colombia 28 Paraguay 29 Myanmar 30 Mexico Sources: IMF, authors’ calculations. Note: The chart shows all countries where data are available and have a value > 4%. not specialised in the same kind of production. Available evidence suggests that many countries are characterised by a high degree of concentration in production and exports (Cashin, Cespedes and Sahay 2002).

For example, a modified version of the HerfindahlHirschmann index suggests that the concentration of exports is highest for resource-rich countries, such as the oilproducing states in Africa, the Middle East and the CIS, and for some small countries. 27 FISCAL AND POLITICAL INTEGRATION Political integration with sizeable fiscal implications is a rare phenomenon, with the exception of political dependencies (see Section 26 For a recent analysis of migration flows in Europe, see Brucker (2002). 7 The index, calculated by UNCTAD, has a value of 1 when a single export product produces all export revenues. Conversely, when export revenues are evenly distributed over a large number of products, the index approaches 0. In total, 65 of the 150 countries in the UNCTAD sample feature an index value larger than 0. 3, while 36 countries have a value larger than 0. 5, indicating that the export structure is rather concentrated. ECB Occasional Paper No. 11 February 2004 13 3. 5. 3). In the case of the European Union, fiscal transfers among Member States amount to less than 1% of GDP.

However, many developing countries receive bilateral and multilateral aid that can be interpreted as a kind of fiscal transfer from the international community. Although development aid is not granted with the aim of acting as a shock absorber, it might represent a potential link with the anchor country that, in some circumstances, is available to offset the impact of shocks and can therefore add to the sustainability of the exchange rate regime. For some countries, these flows of official development assistance (ODA) constitute a significant share of Gross National Income (GNI).

World Bank data reveal that 75 countries received ODA flows amounting on average to more than 5% of GNI in the period 1990-1999. TRADE AND FINANCIAL INTEGRATION Available evidence suggests that with regard to trade in goods, pre-EMU Europe achieved the highest level of regional integration among independent countries (see Table 2. 3). Only in Europe does intra-regional trade in goods account for the largest share in total trade and, at the same time, feature strongly in terms of intra-regional GDP.

Intra-regional trade of ASEAN countries, which are more open in general, expressed as a percentage of GDP, is slightly higher than in the euro area. However, the share of ASEAN intra-regional trade in total trade is rather low (Bayoumi and Mauro 2001). The opposite is true for NAFTA countries: while these are rather closed economies, Table 2. 3 Intra-regional trade in goods, 1998 1 ) NAFTA intra-regional trade accounted for about half of total trade. On both counts, trade integration is much lower among Mercosur and GCC countries. 8 Turning to trade integration with the US and the EU, the evidence suggests that on average countries show a higher degree of trade integration with the EU than with the US. However, there is a strong regional pattern. Among the countries exhibiting a large degree of trade integration with the EU are all the acceding and accession countries, the non-EU countries of western Europe, Norway, Iceland and Switzerland, and most other countries belonging to the “Euro Time Zone” (Mazzaferro et al. 2002). Countries characterised by a high degree of trade integration with the US are mainly located in the

Western Hemisphere. Turning to financial integration, the 1990s saw a steep rise in international capital flows. Moreover, there is evidence suggesting that long-run co-movements of financial prices have become quite significant across countries. 29 Foreign direct investment (FDI) flows increased as well, increasing globally from about USD 200 billion in 1990 to roughly USD 1,500 billion in 2000. The bulk of financial cross-border flows have remained within industrialised countries (IMF 2001d), 28 For a general overview of trade integration across regions, see IMF (2002). 9 See the review of the evidence on financial integration in developed countries provided by Mongelli (2002). Measures on intra-regional trade in goods Region Euro area NAFTA ASEAN-5 Mercosur Gulf Cooperation Council (GCC) Exports % of total trade 45. 9 55. 3 19. 3 25. 0 8. 0 Imports % of total trade 45. 0 43. 5 19. 3 21. 4 8. 1 Exports % of regional GDP 12. 0 5. 4 14. 3 1. 8 2. 9 Imports % of regional GDP 11. 0 5. 5 11. 7 2. 0 2. 4 Sources: IMF, own calculations. 1) Intra-regional trade: exports/imports of the member countries to/from each other. 14 ECB Occasional Paper No. 1 February 2004 reflecting the relative stability of the institutional environment in these countries (IMF 2002). Moreover, developed economies are still characterised by a home bias, as financial market integration continues to be significantly larger within countries than between countries (IMF 2002). BUSINESS CYCLE CORRELATION Empirical evidence on world business cycles suggests that business cycle correlation for industrial countries is significantly higher than for developing countries or between developing and industrial countries (Calderon, Chong and Stein 2002; Barro 2001).

This result seems to reflect evidence on other OCA properties. Recent research indicates that a higher degree of trade and financial openness corresponds to a higher degree of business cycle correlation. Trade specialisation may also be an important factor, as business cycles of economies with similar structures are – ceteris paribus – more correlated than business cycles of economies with different economic structures (Imbs 2003).

Thus, to a large extent business cycle correlation seems to reflect developments in trade and financial integration as well as trade diversification. the anchor country. Here, some call for ex ante integration to ensure sustainability, while others argue for ex post integration, relying on endogenous integration tendencies related to the adoption of the foreign currency. These considerations explain why, despite broad agreement on the potential costs and benefits of dollarisation/euroisation, the literature is far from conclusive. 0 This relates in particular to cases where – from a stability perspective – countries face difficulties in making efficient use of domestic monetary policy owing to a lack of credibility, but – according to the integration point of view –still need domestic monetary policy as an adjustment tool to use in the wake of asymmetric shocks (see Overview 2. 1). Empirically, these cases are quite relevant as there are a number of countries that seem to be suffering from a lack of monetary policy credibility owing to a history of monetary instability and a high degree of unofficial dollarisation/euroisation.

But, more often than not, the very same countries seem to need domestic monetary policy since other adjustment mechanisms are not available and because they lack a sufficient degree of integration with the potential anchor country. Rather, they are characterised by a low degree of price and wage flexibility and feature a high degree of export product concentration. Factor mobility might have increased for some countries, as indicated by a rise in remittances.

Most countries do not at any rate have the perspective of a political integration process with potential anchor countries, such as the US or the EU. And although some countries can count on substantial transfers from the international community in terms of official development assistance, this cannot necessarily be interpreted as an adjustment tool. Finally, most countries, in particular larger ones, are rather closed economies and lack a high degree of real and financial integration with the potential anchor countries. 2 DOLLARISATION/ EUROISATION – A REVIEW OF THE LITERATURE SUMMARY AND OUTLOOK: STABILITY AND INTEGRATION AS KEY ISSUES IN ANALYSING THE POTENTIAL COSTS AND BENEFITS OF DOLLARISATION/EUROISATION Stability and integration are the two main approaches adopted in this paper to analyse the current cases of dollarisation/euroisation and their policy implications. These two concepts capture the essence of the policy issues included in this specific monetary regime: on the one hand, dollarisation/euroisation is seen, from a stability point of view, as a tool to eliminate a credibility problem and therefore to enhance macroeconomic and financial stability.

On the other hand, from an integration point of view, dollarisation/euroisation is importantly linked to the degree of economic and financial (and often also institutional) integration with 30 See also Del Negro et al. (2001), who in a concluding paragraph summarise several papers on the pros and cons of dollarisation. ECB Occasional Paper No. 11 February 2004 15 Overview 2. 1 The dollarisation/euroisation debate

Sustainable exchange rate regime debate and ”new” view on OCAs Value of having domestic monetary policy at one’s disposal is low if • Monetary policy lacks credibility “Old” view on OCAs • Other adjustment mechanisms are available • There is no need for an adjustment mechanism (symmetric shocks, synchronised business cycles) Indicators used to assess the value of monetary policy • Inflation rate • Degree of unofficial dollarisation/euroisation • Price and wage flexibility • Factor mobility • Diversification of production structure • Fiscal and political integration • Openness and trade integration • Financial integration Support for dollarisation/ euroisation Strong, as many countries seem to be characterised by a low degree of monetary policy credibility and a high ability to converge to an anchor country’s price movements Weak, as many countries seem to need an adjustment mechanism and there is only limited ability to substitute convergence for domestic monetary policy as an adjustment tool Source: Authors’ compilation.

Thus, applying economic criteria may often lead to contradictory conclusions regarding the cost and benefits of a possible move towards dollarisation/euroisation for individual countries. 31 The following sections examine in detail cases where individual countries opted for dollarisation/euroisation. 31 Argentina is a classic example of this (Frankel 1999). An analysis for Latin America as a whole can be found in Berg, Borensztein and Mauro (2002). 16 ECB Occasional Paper No. 11 February 2004 3 C A S E S O F S U S TA I N E D D O L L A R I S AT I O N / E U R O I S AT I O N On the basis of the list of territories published by the United Nations 32 it is possible to identify 51 cases of sustained official dollarisation/ euroisation. Most of them are small, many even involving a population of less than 100,000.

Moreover, almost two-thirds of the territories are politically dependent. While the degree of dependency and the status under international law varies, the term “countries” is, as already mentioned, used to cover both dependent and independent territories for the sake of simplicity. 33 Only where the distinction between independent countries and political dependencies appears to have an economic bearing or implications related to the OCA criteria will we distinguish between countries and other territories. Information on the economic performance of dollarisation/euroisation cases has, owing to a lack of data, been limited. However, when Rose and various co-authors (e. g.

Rose 2000 and 2001, Rose and Engel 2000a, 2000b, Glick and Rose 2001) provided econometric evidence on the large trade effects of a common currency, the sample of countries with a common currency included many of these cases. The feature that has attracted the most attention is that the results of these studies suggest that dollarisation/euroisation could be a sustainable exchange rate regime even if integration with the potential anchor country is rather low ex ante, as it might endogenously increase ex post. This claim, which depends on the specific attributes of small and microstates, merits further scrutiny. The following sections present an overview of the current cases of dollarisation/euroisation and the evidence in Rose et al.

The subsequent sections review in detail their economic performance, searching for common, economically relevant features that might have contributed to the sustainability of the exchange rate regime. The focus is initially on six independent countries that have been dollarised/ euroised for quite some time, namely the five European microstates and Panama. After a broad analysis of the remaining cases, a final section summarises the results and derives some policy lessons. 3 CASES OF SUSTAINED DOLLARISATION/ EUROISATION 1 OVERVIEW OF CASES Of the 51 identified cases of sustained official dollarisation/euroisation (see Table 3. 1), 15 use the US dollar, and 13 the euro. Other currencies used are the South African rand, the Australian dollar, the Danish krone, the New Zealand dollar, the Swiss franc, the British pound and the Indian rupee. 4 Some countries do formally have their own currency, such as the Bahamas, Barbados, Bermuda, Belize, Bhutan, Lesotho or Namibia, 35 but as they have been pegged to the respective anchor currency for such a long time and/or are used interchangeably with the anchor currency, they can be safely considered as being dollarised/euroised. 36 The sample of dollarised/euroised countries has several special characteristics: – Most of the countries are very small. In total, the 51 countries account for a population of roughly 18. 5 million people. Only seven of them have a population larger than 1 million. – In the sample there are 31 politicallydependent territories, usually dependencies or overseas territories of the respective anchor countries.

Of the 20 independent countries in the sample, 13 only became independent between the late 1960s and the early 1990s. 32 See www. un. org/Depts/Cartographic/english/geoname. pdf. 33 Again, this does not imply the expression of any opinion whatsoever on the part of the authors or the European Central Bank on the legal status of any country, area or territory or its authorities, or concerning the delimitation of its frontiers. 34 See also Rose (2000). 35 British dependencies officially use the pound with a reference to the dependencies’ name, i. e. St. Helena pound, whereas French Polynesia, New Caledonia and Wallis and Futuna use the CFP franc. 36 Schuler and Stein (2000) characterise these cases as “semiofficially dollarised”.

On the other hand, Hawkins (2003) calls the monetary arrangement in Bermuda, the Faeroes, the Falklands Islands, Gibraltar, Guernsey, the Isle of Man, Jersey, Lesotho, Namibia and St. Helena a currency board. ECB Occasional Paper No. 11 February 2004 17 Table 3. 1 Sustained cases of dollarisation/euroisation Euroised countries (dependent territories) Reunion Guadeloupe Martinique French Polynesia New Caledonia French Guiana Mayotte Wallis and Futuna St. Pierre and Miquelon Population Dollarised countries (dependent territories) Puerto Rico Guam Virgin Islands (U. S. ) Northern Mariana Is. American Samoa Bermuda British Virgin Is. Turks and Caicos Is. Population

Other cases of official foreign currency adoption (dependent territories) Jersey (GBP) Isle of Man (GBP) Guernsey (GBP) Greenland (DKK) The Faroes (DKK) Gibraltar (GBP) Cook Islands (NZD) St. Helena (GBP) Falkland Is. (GBP) Christmas Is. (AUD) Norfolk Is. (AUD) Tokelau (NZD) Cocos Is. (AUD) Pitcairn Is. (NZD) Total Population 732,570 431,170 418,454 253,506 204,863 177,562 163,366 15,435 6,928 3,937,316 157,557 122,211 74,612 67,084 63,503 20,812 18,122 89,361 73,489 64,342 56,352 45,661 27,649 20,611 7,266 2,895 2,771 1,879 1,445 633 47 394,401 Total 2,403,854 Total 4,443,095 Independent euroised countries (date of independence) Population Independent dollarised countries (date of independence) Population

Independent countries that have adopted another foreign currency (date of independence, currency adopted) Lesotho (1966, ZAR) Bhutan (1949, INR) Namibia (1990, ZAR) Swaziland (1968, ZAR) Kiribati (1979, AUD) Liechtenstein (1806, CHF) Nauru (1968, AUD) Tuvalu (1978, AUD) Niue (1974, NZD) Total Population Andorra (1278) Monaco (1419) San Marino (301) Vatican City (1929) 67,627 31,842 27,336 890 Panama (1903) Bahamas (1973) Barbados (1966) Belize (1981) Micronesia, Fed. States (1986) Marshall Islands (1986) Palau (1994) 2,845,647 297,852 275,330 256,062 134,597 70,882 19,092 2,177,062 2,049,412 1,797,677 1,104,343 94,149 32,528 12,088 10,991 2,124 7,280,374 Total 127,695 Total 3,899,462

Sources: CIA World Fact Book, authors’ compilation. – With the exception of the island countries in the North Pacific, all independent countries use the currency of an anchor country either in their geographical vicinity or with whom they share a common border. Hence, for these countries, dollarisation/euroisation has a regional dimension. Only dependent territories use the currencies of their home countries however distant they are. endogenous trade effects of a common currency. Using a sample of currency unions that includes 22 cases listed in Table 3. 1, 37 Rose and co-authors estimate a standard gravity model of bilateral trade based on extensive cross-country panel data.

The model includes income and distance variables, controlling for other variables potentially affecting trade flows, such as a free trade area, a common border, language, colonial past, size, landlocked countries, islands, etc. Finally, a 2 THE ENDOGENEITY THESIS AND THE ROSE EVIDENCE The case of dollarisation/euroisation attracted attention with the publication of a series of papers by Rose and various co-authors on the 37 The CFA franc zone countries, the countries of the East Caribbean Currency Area (ECCA) and the respective anchor countries are among the 56 countries that constitute the group of countries sharing a common currency in Rose’s analysis. 18 ECB Occasional Paper No. 1 February 2004 dummy variable is included in the regressions, which is equal to unity when a country pair is a currency union, and zero otherwise. The result suggests that two countries sharing a common currency trade far more, perhaps over three times as much, than comparable countries with different currencies. 38 This finding has proven to be statistically significant and robust with regard to other variables potentially affecting trade flows. The impact on the dollarisation/euroisation debate has been strong, as the level of trade integration needed to secure the benefits of stability in the medium and long term might be an endogenous variable. 9 Thus, the result supports the “endogeneity thesis” (Frankel and Rose 1998; Dallas and Tavlas 2001), turning the old OCA view on its head: the traditional OCA properties should thus be seen as a consequence, and not as a prerequisite, of adopting a common currency. The Rose results triggered a debate on several aspects. Methodological issues and questions of data reliability were raised by Nitsch (2002) and by Persson (2002). 40 Other observers wondered whether the results can be generalised (Quah 2000; Obstfeld 2000; Masson 2000), given the special characteristics of many dollarised/euroised countries. 41 Studies on other economic characteristics of sustained cases of dollarisation/euroisation followed.

Most prominently, Edwards (2001) and Edwards and Magendzo (2001, 2002) – analysing the macroeconomic performance of Panama and 13 other cases listed in Table 3. 1 42 – found that dollarised/euroised countries have experienced significantly lower inflation, but have also grown at a significantly lower rate than countries with their own currencies. 43 Moreover, fiscal records of the two groups of countries are rather similar. Overall, this line of research seems to suggest that the adoption of a foreign currency does not automatically ensure a good, let alone superior, macroeconomic performance (Backe and Wojcik 2003). The debate suggests that the policy issues raised by dollarisation/euroisation cannot be nswered by only referring to the possible trade effects of a common currency. Even if they are large, they might not be sufficient to ensure the benefits of an exchange rate regime. 44 In addition, a broader analysis might reveal that other adjustment mechanisms – not captured in the trade regressions – had come into play, supporting the sustainability of the exchange rate regime in those instances where dollarisation/euroisation was maintained. 45 It remains possible that a high degree of trade integration between members of a common currency area might emerge as a result of other policies or country-specific factors beyond those controlled for in the Rose regressions, rather than the common currency itself.

For these reasons, we examine case studies of the examples of sustained dollarisation/euroisation as well as of those few countries which have exited such a regime (see Section 4). Thus, the following analysis does not pretend at drawing final conclusions, but rather aims at providing an overview of the key economic characteristics and economic policies of this set of countries. 3 CASES OF SUSTAINED DOLLARISATION/ EUROISATION 38 Earlier research had largely failed to identify a significant and positive effect of exchange rate stability on trade. 39 For example, Glick and Rose (2001) claim that joining a currency union may cause bilateral trade to almost double.

Frankel and Rose (2002), explicitly referring to dollarisation/euroisation, suggest that this exchange rate regime might be associated with both enhanced economic integration and also higher economic growth. 40 See also the response by Rose (2002). 41 Some countries in the Rose sample represent members of a regional, multilateral currency arrangement, such as the countries of the CFA franc zone and the ECCA. Klein (2002) argues that by confining the analysis to cases of unilateral dollarisation/ euroisation, for example examining the United States’ bilateral trade with countries that have entered into a currency union with the United States, it is much more difficult to find significant trade effects. 2 Andorra, Kiribati, Liechtenstein, the Marshall Islands, Micronesia, Monaco, Nauru, Panama, San Marino, Tuvalu, the Cook Islands, Greenland and Puerto Rico. In addition, Liberia and Palau were included. 43 Edwards and Magendzo (2002) claim that it is the performance of the ECCA countries which led to the econometric result in Frankel and Rose (2002) that currency union members grow at a faster rate than countries with a domestic currency. 44 Glick and Rose (2001) point out that the trade effects of a common currency may take some time. 45 Lockwood (2000) raises the question whether the Rose results actually support the old view of OCAs, indicating that the formation of currency unions may be endogenous to trade flows. ECB Occasional Paper No. 1 February 2004 19 3 THE EUROPEAN MICROSTATES 3. 1 BASIC FACTS In Europe, five countries, which can be subsumed under the heading “European microstates”, adopted a foreign currency as early as the 19 th or early 20 th century: Andorra, Liechtenstein, Monaco, San Marino and the Vatican City. On 1 January 1999, three of these (San Marino, the Vatican City and Monaco) became officially euroised. 46 Moreover, as the EU Treaty legally entitles these countries to use the euro as their official currency, based on agreements between these three countries and France and Italy respectively, these cases do not reflect unilateral euroisation in the literal sense.

By contrast, there is no legal arrangement between the European Community and Andorra regarding the use of the euro in Andorra, nor was there one in the past between Andorra and France or Spain regarding the use of the French franc and the Spanish peseta. Nevertheless, in October 2000 Andorra adopted a law officially introducing the euro in Andorra. Finally, Liechtenstein unilaterally adopted the Swiss franc as legal tender in 1924, after its economy had suffered considerable damage from the erosion of Austria’s one-time currency, the crown, and the failures of several attempts to create a national currency (Klauser 2001). 47 For many years, the European microstates had no institution performing central bank functions.

This changed in the late 1980s and early 1990s, when some established institutions for treasury and tax collection purposes, to supervise the banking system, to administer required reserves, and to manage financial relations and transactions with foreign counterparties. 3. 2 ECONOMIC DEVELOPMENTS OVER THE LONG RUN 48 The European microstates have reached income levels as high as or higher than the countries whose currency they have adopted. This is a major achievement given that economic conditions in these countries were rather poor before World War II. 49 Their rapid post-war development has been almost exclusively based on tourism and banking. Tourism and the hotel sector account for almost 35% of total employment in Andorra, while the financial centre accounts for approximately one-third of real net output in Liechtenstein (Klauser 2001) and for roughly 20% of global turnover in Monaco. 3. OPTIMUM CURRENCY AREA PROPERTIES THE NEW OPTIMUM CURRENCY AREA THEORY The new OCA theory suggests several indicators for evaluating the appropriateness of dollarisation/euroisation for individual countries. In the following sections on sustained, abandoned and recent cases of dollarisation/euroisation, the focus will be on whether the alleged increase in credibility has been accompanied by low(er) inflation, interest rates and bond spreads as well as by smaller fiscal and current account imbalances. Moreover, for those cases where data are available, i. e. mainly for those that have only recently dollarised/euroised, exchange rate developments and the degree of unofficial dollarisation/euroisation before adopting the foreign currency are taken into account.

For the European microstates, the available evidence suggests that inflation rates have been in line with those observed in their anchor countries. Moreover, in general, public finances have been sound. Finally, European microstates have registered overall balance of payments 46 As the Vatican City has a very small population size and owing to a lack of data, the remaining part of this section focuses purely on Andorra, Liechtenstein, Monaco and San Marino. 47 Almost sixty years later Liechtenstein formally became part of the Swiss monetary area within the “Currency Agreement” concluded in 1981. According to this Agreement, all Swiss regulations concerning monetary policy are automatically applicable in Liechtenstein. 8 With the exception of the IMF member country San Marino, the analysis of economic developments in the European microstates is severely hampered by a lack of reliable data. In many cases the information provided in the following section is taken from country websites and the CIA World Factbook. Hence, any conclusions should be interpreted with care. 49 Some sources even categorise their economic status before World War II as “impoverished”. 20 ECB Occasional Paper No. 11 February 2004 surpluses, resulting in a build-up of net foreign assets. THE OLD OPTIMUM CURRENCY AREA THEORY – Factor mobility High growth in the European microstates has been accompanied by unemployment rates that have been much lower than in the respective anchor countries.

Thus, the labour market of several microstates has been characterised by excess demand and labour migration, with commuting from the surrounding European countries forming an important adjustment mechanism to cope with this excess demand (Hitzelsberger, Reuter and Steinle 2001). – Integration via fiscal transfers The European microstates are independent countries and have not received any fiscal transfers from their anchor countries. Moreover, as public finances have in general remained sound, there has also been no need for any fiscal support. – Real and financial integration All microstates are highly open economies. Openness ratios, i. e. the sum of exports and imports divided by GDP, range from 100% in Andorra to 400% in San Marino.

Liechtenstein’s exports are in the range of 200% of GDP. With the exception of Liechtenstein, where the EU and the US are major trading partners, trade relations of the microstates have been dominated by their respective anchor countries and the EU, with whom they also form a customs union. In addition, tourism has been a major source of integration. Andorra and Monaco are among the most tourism-intensive economies in the world, measured by the number of overnight visitors as a percentage of the total population (see Chart 3. 3). Moreover, most of the tourists visiting the European microstates are from neighbouring European countries (World Tourism Organization 2002).

Banking constitutes the second pillar that the European microstates have relied on in developing their economies, benefiting from regulatory and statutory advantages. 50 Thus, the banking sectors of the European microstates have become attractive offshore financial centres (OFC),51 serving a much broader public than domestic households, enterprises and institutions. In Monaco, which represents the most developed financial centre of the four countries, 90% of banking customers are said to be non-residents, while non-residents make up 54% of deposits in San Marino (IMF 2001a). The main activity of the banking system in the microstates is the reshuffling of funds provided by non-residents to international capital markets. Liquid assets by far dominate the balance sheet of the respective banking sectors.

By contrast, loans to customers only play a negligible role, with less than 20% of total assets. 3 CASES OF SUSTAINED DOLLARISATION/ EUROISATION 4 PANAMA 4. 1 BASIC FACTS With a population of 2. 9 million, Panama is by far the largest politically independent country among the cases of sustained dollarisation. The introduction of the US dollar as the domestic currency followed the country’s secession from Colombia in 1903, which was backed by the United States. After having gained independence, the country signed a treaty with the US allowing for the construction of the Panama Canal, and establishing the Panama Canal Zone under US sovereignty. In 1904, the US dollar became legal tender. 52 0 For a detailed account of the advantages for banks and nonresident investors given by tax privileges in San Marino, see IMF (1999). 51 Countries are considered to be OFCs if they are classified as tax havens by the OECD (2000), including the six “advance commitment jurisdictions” that were not named because they had publicly committed themselves to eliminating their harmful tax practices prior to the publication of the report, or were reported as OFCs by the Financial Stability Forum (FSF 2000). 52 The US dollar is used as a paper currency while the local balboa exists as coins. Thus, Panama does have a separate currency, but in practice, most transactions are conducted in US dollars. ECB Occasional Paper No. 11 February 2004 21

The Banco Nacional de Panama (BNP), a stateowned credit institution created in 1904, acts as the financial agent for the central government, is the official clearinghouse for the banking system, and ensures an adequate supply of US banknotes to the banking system. Occasionally, it has also supplied emergency liquidity to banks. Only recently, in 1998, the “Superintendencia de Bancos” was created, marking the start of regulatory and supervisory reform of the banking system, which had not been subject to proper public regulation for decades. 4. 2 ECONOMIC DEVELOPMENTS OVER THE LONG RUN Panama’s modern economic development is intimately linked to the Canal and the financial sector. In 1925, Panama opened its shipping registry to international companies, offering an attractive regulatory and tax system.

This unprecedented act established the system of “flags of convenience”, making it attractive for ships from other countries to be registered in Panama. The Canal is also the basis of the “Colon Free Zone”, located on its Atlantic side and established in 1948. It represents the largest free trading zone in the Western Hemisphere. Since the 1970s, the financial sector has constituted the second pillar on which Panama’s economic development has been built. The capital account is entirely open, and banks have been free to invest excess funds in Panama or abroad. Until the mid-1980s, gross foreign assets increased from basically zero to more than USD 25 billion.

Today, the banking sector, including a well-developed offshore centre, accounts for approximately 14% of Panama’s output (IMF 2001c). The importance of the Canal and banking is also reflected in the structure of Panama’s economy, which is overwhelmingly based on the service sector, accounting for more than 75% of GDP. Services include commerce, restaurants and hotels, transport and communications, financial intermediation, the Colon Free Zone, and the Panama Canal (IMF 2001c). Income per capita is about the average of Latin American and Caribbean countries, and the country’s growth rate has not been statistically different from that of non-dollarised countries (Edwards 2001; Mussa and Loser 1999).

The economy is characterised by the co-existence of two distinct environments. On the one hand, there is a relatively affluent, urban, servicebased economy in Panama City, the Canal Zone and the international banking centre. On the other hand, the country has a relatively depressed rural-based economy which, while representing over half of the population, accounts for only a small share of GDP (EC 2002). 4. 3 OPTIMUM CURRENCY AREA PROPERTIES THE NEW OPTIMUM CURRENCY AREA THEORY Monetary stability in Panama has been remarkable compared with most other countries in the region. Indeed, inflation rates have often been lower than in the United States.

Short-term interest rates have been relatively low as well, with differentials over the United States limited to around 100 basis points over Libor rates for deposits and to 160-200 basis points on commercial loans. Fiscal policy, however, has been rather expansionary for most of the last three decades (see Table 3. 2 and Edwards 2001), largely financed by extensive foreign borrowing. Moreover, with the exception of the late 1980s, when US sanctions hit the economy, Panama has consistently run current account deficits. Thus, the external debt/GDP ratio has been close to 100%. International long-term bonds are traded with a risk premium of 400-500 basis points in terms of bond spreads over US Treasuries. This is significantly higher than bonds of other sovereign Latin American borrowers, e. g. Chile, with similar maturities (Edwards 2001). 22

ECB Occasional Paper No. 11 February 2004 Table 3. 2 Selected macroeconomic indicators – Panama Average 1970-79 GDP growth GDP per capita growth Inflation Government balance (as a % of GDP) Current account balance (as a % of GDP) External debt (as a % of GDP) Sources: IMF, World Bank, authors’ calculations. 1) 1980-1985: -3. 9. 4. 7 2. 4 6. 0 -7. 3 n. a. 57. 1 Average 1980-89 0. 9 -1. 0 3. 1 -5. 1 -0. 11) 99. 7 Average 1990-99 5. 1 2. 9 1. 1 -4. 3 -3. 3 94. 9 3 CASES OF SUSTAINED DOLLARISATION/ EUROISATION Moreover, spreads over US Treasuries seem to be largely driven by factors common to emerging markets (Mussa and Loser 1999; Goldfajn and Olivares 2000).

THE OLD OPTIMUM CURRENCY AREA THEORY – Factor mobility In comparison to other Latin American countries, like El Salvador, Nicaragua or Ecuador, migration flows from Panama to the United States have been rather limited. Estimates put the number of Panamanians living in the United States between 100,000 and 200,000. Remittances are said to reach between USD 20 and 170 million (Orozco 2003) and are not a major source of financing for the country’s imports. – Integration via fiscal transfers There have been no direct fiscal transfers from the US budget to Panama. However, the economy has greatly benefited from US expenditures related to the military presence of the US.

Since the mid-1960s, the country has continuously relied on substantial support from the IMF (Edwards 2001; Goldfajn and Olivares 2002). Moreover, Panama went through several rounds of foreign debt restructuring. Development aid, provided by the US, multilateral development banks and other governments on bilateral terms, has been comparatively modest (see Table 3. 3). – Real and financial integration Panama’s openness ratio stands at around 75%. As per Table 3. 3, roughly 35% of trade is with the anchor country, the United States. Foreign direct investment, in particular from the US, has been strong, with the exception of periods when the Panamanian government took a rather hostile attitude vis-a-vis the US.

The financial sector is the key sector targeted by foreign investors. Finance, insurance and real estate account for more than 95% of all US investments in Panama (Borga and Yorgason 2002). Table 3. 3 Selected indicators of integration – Panama Average 1970-79 Trade (as a % of GDP) Trade with the US (as a % of total trade) FDI (as a % of GDP) Aid (as a % of GDP) Sources: IMF, World Bank, authors’ calculations. n. a. n. a. 1. 3 1. 8 Average 1980-89 76. 1 38. 4 0. 1 1. 0 Average 1990-99 75. 3 35. 7 5. 0 1. 0 ECB Occasional Paper No. 11 February 2004 23 Panama has achieved a high degree of financial integration by attracting non-residents through favourable secrecy and control regulations. 3 More than a third of the roughly 80 banks operating in the country are offshore institutions, more than half of the liabilities of the banking system are foreign liabilities, and a large part of banking activity consists of re-channelling foreign funds to onshore financial markets. 54 Loans to the domestic economy, on the other hand, account for only 50% of total loans. Even this is a rather new phenomenon, reflecting the credit boom Panama has experienced over the last decade, during which time business, housing and consumer loans to the domestic economy have almost tripled, growing much faster than GDP. To substitute for the missing local lender of last resort, domestic banks have established lines of credit with foreign banks with branches in Panama, and have usually been able to draw on them during liquidity crunches (Mussa and Loser 1999). 55 dministered UN trusteeships in the Pacific 57 continue to use the US dollar as their domestic currency even though they gained independence in the mid-1980s/1990s. 5. 2 ECONOMIC DEVELOPMENTS OVER THE LONG RUN Dollarised/euroised countries have on average a higher per capita income than countries with their own currencies (see Table 3. 4). Twelve of them, namely Bermuda, the European microstates, Jersey, Guam, Guernsey, the Faeroes, Greenland, the Isle of Man and Gibraltar, belong to the 40 richest countries worldwide, expressed by per capita income levels according to purchasing power parity. Only five dollarised/euroised countries – Bhutan, Tuvalu, Tokelau, Kiribati and Mayotte – have a per capita income per annum of less than 1,500 USD.

The difference in per capita income levels between dollarised/euroised countries and countries with their own currencies is even more striking when comparing the median of the respective country groupings. 5 REMAINING CASES OF SUSTAINED DOLLARISATION/EUROISATION 5. 1 BASIC FACTS The majority of the remaining 45 cases of sustained dollarisation/euroisation are small, dependent territories. French, US and Danish dependencies have always used the French franc (now the euro), the US dollar and the Danish krone respectively, whereas Bermuda, the British Virgin Islands, and the Turks and Caicos Islands followed other Caribbean countries in the early 1970s and switched from the British pound to the US dollar (Box 3). 6 Turning to independent countries, the British pound played a significant role in the monetary history of several countries in the Caribbean that are currently using the US dollar (see Box 3. 1), and countries in southern Africa that are now part of the Rand zone. The US- 53 Hausmann and Eichengreen (1999) note that “the financial depth and stability of the Panamanian financial system is not associated with the transparency and good practices that dollarization is supposed to bring, but precisely with the country’s lack of transparency. ” Only in the 1990s, when the international campaign against money laundering activities intensified, did the authorities tighten regulations and strengthen the role of the Banking Supervisory Authority (IMF 2001c). 4 The analysis of the Panamanian banking sector mainly relies on information provided in IMF (2001c). 55 Of the 54 banks with a general license, 26 are banks with headquarters in Panama, accounting for 55% of the banking sector’s assets. 18 are branches of foreign banks, and 10 operate as (separately capitalised) subsidiaries of foreign banks. European and North American banks hold around 30% of total banking sector assets. 56 The Pitcairn islands use the New Zealand dollar. Former UK dependencies that became dependencies of Australia and New Zealand implicitly experienced a change in the currency regime as well, since they now use the Australian/New Zealand dollar rather than the British pound. 7 The Federal States of Micronesia, the Marshall Islands and Palau. 24 ECB Occasional Paper No. 11 February 2004 Box 3 3 CASES OF SUSTAINED DOLLARISATION/ EUROISATION FROM STERLING TO US DOLLAR – A SHORT REVIEW OF THE EVOLUTION OF CURRENCY ARRANGEMENTS IN THE CARIBBEAN 1 Up to the end of the 1950s/early 1960s English-speaking countries in the Caribbean operated sterling-based currency boards as a legacy of their colonial past. Trade and financial links were already stronger with the US than with the UK, but under the Bretton Woods system of fixed parities, this mismatch of monetary and real/financial links was not perceived as an economic challenge.

This changed after the 1967 devaluation of sterling against the US dollar, causing an inflationary shock in the region. Thus, when the Bretton Woods regime collapsed, most countries in the region switched to a US dollar peg, including the Bahamas, Barbados and Belize. 2 Since then they have managed their currencies in a quasi-currency board manner, maintaining levels of foreign reserves so that they were always able to intervene successfully in support of the peg. Only Guyana, Jamaica and Trinidad and Tobago abandoned the quasicurrency board strategy in the 1970s and 1980s. 1 This box relies on information provided by Worrell (2003) and Beek et al. (2000). The member countries of the Eastern Caribbean Currency Union, namely Anguilla, Antigua and Barbuda, Dominica, Grenada, Montserrat, St. Kitts, St. Lucia and St. Vincent, decided to establish a common currency, the East Caribbean dollar, based on a currency board arrangement with a US dollar peg. Restricting the comparison to the sample of non-industrialised countries reveals that, both on average and in median terms, dollarised/ euroised countries are significantly richer than developing countries and emerging market economies with their own currencies. Moreover, politically-dependent dollarised/ euroised jurisdictions have on average reached higher income levels than independent dollarised/euroised economies (see Table 3. 4). 8 As in the European microstates, the high level of income of several dollarised/euroised countries is a post-World War II phenomenon, in particular in the Caribbean. Again, tourism and banking constitute core sectors contributing to the rapid development of the respective economies. 5. 3 OPTIMUM CURRENCY AREA PROPERTIES THE NEW OPTIMUM CURRENCY AREA THEORY Available evidence suggests that, in general, the inflation rates of dollarised/euroised countries have generally moved in line with those of the respective anchor countries. In particular, inflation rates in the Bahamas and Barbados are highly and positively correlated with the inflation rate in the United States.

In Belize, correlation with US inflation increased substantially after switching to the US dollar in 1981. Somewhat lower are the correlation coefficients between the inflation rates of the three countries in the Rand zone and the South African inflation rate. Finally, the example of Kiribati relative to Australia suggests that inflation rate movements of the dollarised country do not 58 Among the dependent dollarised/euroised jurisdictions only Reunion, Saint Helena, Wallis and Futuna, Tokelau and Mayotte have a per capita income of less than USD 5,000 (in PPP terms). ECB Occasional Paper No. 11 February 2004 25 Table 3. 4 Per capita income in selected country groups 1 ) in USD, in PPP terms) Groups (political status, sample size) World (all countries and jurisdictions, 228) Countries with their own currencies (all countries and jurisdictions, 182) Dollarised/euroised countries (all countries and jurisdictions, 46) 2) Dollarised/euroised countries (independent countries, 19) Dollarised/euroised countries (dependencies, 27) Non-industrialised countries (205) Sources: CIA World Factbook, authors’ calculations. 1) Most recent data available. 2) Excluding Pitcairn, the Cocos, Christmas and Norfolk Islands, the Vatican City. Average 8,800 8,200 11,120 9,400 12,340 6,840 Median 4,780 4,200 8,850 4,500 11,150 4,090 have to be closely aligned with those in the anchor country. Kiribati has a narrow-based economy, with 80% of Kiribati households making a living by fishing. Hence, the economy is highly vulnerable to external, asymmetric shocks that affect price developments. 59 Thus, inflation developments in Kiribati are only mildly correlated with those in Australia.

As has already been mentioned, fiscal policies in dollarised/euroised countries seem to show similar results as those of countries with their own currencies (Edwards and Magendzo 2001). However, as will be shown below in more detail, many dollarised/euroised countries and jurisdictions have received substantial fiscal transfers from their respective anchor countries. Many countries with a long history of dollarisation/euroisation can be characterised as countries with large trade deficits. 60 Since neither GDP nor current account data are available for many cases of dollarisation/ euroisation, the trade balance/total exports ratio is used as an indicator for the purpose of crosscountry comparison.