12 Currency Mismatches and Domestic Liability Dollarization

Adrián Armas, Eduardo Levy Yeyati, and Alain Ize
Published Date:
July 2006
  • ShareShare
Show Summary Details
Philip Turner1

12.1 Introduction

My views on the policy implications of dollarization – that is, the use of foreign rather than local currency – can be briefly summarized. While some have argued that it is endemic, I shall argue that dollarization in medium-sized countries reflects bad policies. Indeed, better policies during the past decade have diminished reliance on foreign currency and increased borrowing in domestic currency. The rapid development of domestic currency bond markets, together with the growing interest of foreign investors in such markets, represents huge progress. But it does raise new questions that require attention.

All economists working on Latin America have heard of ‘original sin’. The idea is that the inability of developing countries to borrow abroad in their own currencies makes currency mismatches almost endemic. The leading exponents of this idea have been Barry Eichengreen and Ricardo Hausmann, who have presented several versions of this thesis.2

The basic idea of the ‘original sin’ metaphor is that it is innate weaknesses, not past behaviour, which determines how developing countries can borrow. Morris Goldstein and I have argued in a recent book that the statistical measures used by Eichengreen and Hausmann are flawed and that their ‘original sin’ hypothesis is too pessimistic.

It is necessary to consider the logic of a currency mismatch. The notion of currency mismatch refers to the impact of a change in the exchange rate on the present discounted value of future income and expenditure flows. Hence all assets and liabilities must enter the calculation – not just cross-border assets and liabilities. The denomination of contracts between residents in foreign currencies matters because a sharp change in the exchange rate can disrupt such contracts and this can have real economic effects. Foreign currency debts between residents may ‘cancel out’ in normal times, but do not do so in a crisis. This is why the focus of this volume on domestic liability dollarization is so important.

The currency denomination of other income flows is also important: foreign currency borrowing to finance the production of tradables is one thing – to finance non-tradables quite another.

It follows from this that a key ratio is the ratio between the currency denomination of debt and the share of tradables in GDP. To put the point at its most measurable: countries with high export/GDP ratios can sustain higher foreign currency shares in total debt. If this ratio is greater than one – more foreign currency debt than foreign currency earnings to finance – then the country has a problem. How large a problem depends on a country’s net foreign currency position: a large net liability position compounds the difficulty.

The index for aggregate ‘effective’ currency mismatch (termed AECM) that Morris Goldstein and I constructed is the product of these two variables. In symbols, the mismatch ratio is:

where FC%TD=Foreign currency share of total debt
X=Exports of goods and services

Then the aggregate ‘effective’ currency mismatch (termed AECM) is the product of MISM and net foreign currency assets (NFCA) as a percentage of GDP viz:

If foreign currency assets are exactly equal to foreign currency liabilities then AECM is zero – that is, there is no aggregate effective currency mismatch. This would be true even in a dollarized economy – where debts were largely denominated in dollars. If a country has a net liability position in foreign exchange (i.e., NFCA is negative), AECM will also be negative.

This measure can be thought of as a stress test for the economy – combining a mismatch ratio with a measure of a country’s net foreign currency position. What is the interpretation of this ratio for a dollarized economy? In a dollarized economy the pure mismatch ratio – basically the foreign currency share of total debt – is very high. How much of a risk this presents to the country depends on the country’s balance between its foreign currency assets and its foreign currency liabilities, that is the net foreign currency position.

12.2 A domestic policy agenda

Several contributors to this volume echo the view Morris Goldstein and I advocated: policy-makers in medium-sized emerging market economies can do a lot to encourage the denomination of debts in local, not foreign currency. There is obviously a macroeconomic dimension. Countries that pursue inflationary or erratic macroeconomic policies obviously find it hard to borrow in their own currencies. Fiscal deficits must be limited. Monetary policy should aim for low inflation. Fixed exchange rates, which encourage undue reliance on ‘cheaper’ dollar borrowing, should generally be avoided.

The microeconomic policy dimension is also important, and the following five policies deserve attention. Governments should:

  • Ensure accounting rules require the full disclosure of exchange rate-related losses from foreign currency-denominated borrowing. Without this, governments and corporations can report lower interest payments if they borrow in dollars or euros, so that their financial position looks healthier than it really is. A more prudent borrower – borrowing in local currency and paying a high rate of interest – therefore looks worse. The monitoring of foreign currency exposures by banks and regulators needs to be continuous and flexible enough to catch new forms of exposure, particularly through derivatives.3

  • Not weaken their own credit standing by denominating too much of their borrowing in foreign currencies. A simple balance sheet perspective would suggest that borrowing to finance local currency assets should be in local currency. In addition, borrowing in local markets helps fiscal discipline because it exerts immediate upward pressure on domestic interest rates. The unpopularity of higher interest rates can, with luck, create a political constituency for lower budget deficits. Issuing foreign currency debt abroad attempts to evade this key element of market discipline.

  • Encourage the entry of foreign banks. Too many crises have demonstrated all too clearly the dangers of relying on cross-border dollar loans from foreign banks. For this reason, the increase of local currency lending through foreign affiliates has been one of the remarkable developments of the past five or six years.4 It is not often appreciated just how far this process has gone. According to some preliminary Bank for International Settlements (BIS) estimates, for instance, more than 60 per cent of the exposures of banks in major financial centres vis-à-vis emerging markets is now in local currency.

  • Develop a liquid local currency bond market.

  • Make the prudential oversight of financial institutions more mismatch-aware. This does not, of course, mean that public policy should crudely force de-dollarization – but just that it should ensure all risks are correctly assessed. Prudential oversight has both a micro dimension (the oversight of individual institutions) and a macro dimension (oversight of the system as a whole). There is little to add to the excellent chapter in this volume on the micro dimension by Jorge Cayazzo, Socorro Heysen, Antonio Garcia Pascual and Eva Gutierrez.

The better supervision of individual institutions needs to be supported by more effective oversight of the system as a whole. The practical difficulty is that quantifying such exposures is quite complex. The standard balance sheet indicators such as net open positions and gross foreign currency lending to domestic firms tell only part of the story. The IMF has recently done some very interesting work in this area.5 A major virtue of this work is the trouble taken to combine the currency and the maturity dimensions of mismatches.

During the past decade, policies in Latin American countries have improved. Fiscal and monetary policies are better. Most countries have flexible exchange rates. Domestic bond markets are developing – Mexico is a striking case in point. Foreign currency-linked borrowing is being phased out – look at Brazil over the past eighteen months. Supervisory regimes have been overhauled.

What has happened to mismatches? In the mid-1990s, the pure mismatch ratio mentioned earlier – the ratio between the foreign currency proportion of debt and the share of exports in GDP – was well over one in all the larger countries except Chile. By 2004, the ratio had been significantly reduced. This is shown in the top panel of Figure 12.1. At the same time foreign exchange reserves have been built up.

Figure 12.1Currency mismatches and reserves coverage ratio1

1. Simple averages. 2. Brazil, Chile, Colombia, Mexico, Peru and Venezuela. 3. Latin American countries cited plus China, the Czech Republic, Hungary, India, Indonesia, Korea, Malaysia, the Philippines, Poland, Russia, South Africa, Thailand and Turkey.

Sources: BIS; Goldstein and Turner (2004) updated (details available from

According to some preliminary cross-country work on the determinants of sovereign credit spreads over the period 1994 to 2004, this reduction in mismatches has had a statistically significant effect in narrowing credit spreads. The implication of this finding is that lower mismatches do indeed improve a country’s standing in international capital markets.

Another significant variable was forex reserves as a percentage of short-term external debt, shown in the lower panel of Figure 12.1. Increases in this ratio – a proxy for liquidity – also led to lower credit spreads. The implication of this is that large aggregate liquidity buffers can help protect dollarized economies. This seems to explain why Peru, one of the most highly dollarized economies in Latin America, weathered regional financial turmoil rather well. Substantial official reserves and the liquid foreign assets of the banks were in effect the country’s insurance policy.6

12.3 Is there foreign interest in Latin American local currency bonds?

A question for the future is this: what is the likely appetite of foreign investors for Latin American local currency debt? An important distinction is between international bonds and local bonds. Each must be considered in turn.

International bonds

Should governments issue international bonds in local currency? Perhaps not. It is possible that liquidity in international bond markets and the ability to fully exploit new instruments (e.g., CDCs) requires concentration on just a few currencies. Whatever the reason, the great bulk of international bonds are, in fact, issued in just two currencies: 43 per cent in US dollars and 41 per cent in euros.7 These currency percentages are much more concentrated than, say, the US or euro-area’s share of world GDP. This is presumably because investors value the liquidity of these markets. Spreading over additional currencies would almost certainly dilute liquidity.

It is true that recently there have been (or will soon be) some highly publicized issues by Latin American borrowers. But some of these issues seem to be either very special (e.g., related to debt restructuring) or have complete or partial exchange rate mitigants. This issue merits further analysis.

Domestic bonds

What about domestic bonds? Foreign investors may still be interested in local currency bonds issued in local markets by Latin American entities. Bond markets are developing in depth in several countries. Domestic debt securities of Latin American entities amounted to about $270 billion at the end of 1994; by the end of 2004, outstandings had risen to $650 billion.

The development of a long, well-defined yield curve in Mexico has been particularly striking.8 While there has been relatively little research on Latin American bonds, recent work at the BIS suggests that Asian local currency bonds can be very attractive to foreign investors. First, the standard analysis of returns and variance suggests that Asian local currency debt provided foreign investors with a lower risk alternative to dollar debt. The mean return on a portfolio of local currency bonds was only a little lower than dollar-denominated funds but the variance was much lower.9 Second, the covariance between a portfolio of local currency bonds and a standard global portfolio is low. Hence, local currency bonds offer valuable diversification benefits. And there is some evidence that correlation remains low, even in periods of stress – so that diversification benefits tend to be rather resilient and survive even ‘bad’ times. An important question is: what will be the properties of local currency bonds in Latin America? It is too early to say. The experience of Asia is encouraging, even if foreign portfolios are still dominated by a few countries.

There is no comprehensive information on foreign investor holdings of domestic bonds. Nevertheless, sustained increases in the share of domestic bonds in the total volume of trading by international financial intermediaries reported by the Emerging Market Trading Association suggests that non-residents have become more involved in domestic bond markets in emerging economies.

This important development forces policy-makers to pay attention to the quality of their own local bond markets – the infrastructure, tax arrangements, the investor base and so on. Monitoring the vulnerabilities that arise from such debt requires better data on maturity and on interest rate linking than exist in many countries. How vulnerable domestic debt is to changes in the interest rate cycle requires more attention than it has received to date.


    Bank for International Settlements (BIS) (2002) ‘The Development of Bond Markets in Emerging Economies’ BIS Papers No. 11 (Basel).

    BoothJ. (2003) ‘Emerging Market Debt Comes of Age’ The Banker (September).

    Committee on the Global Financial System (2004) ‘Foreign Direct Investment in the Financial Sector of Emerging Market Economies’ CGFS Papers No. 22 (Basel: BIS).

    Committee on the Global Financial System (2005) ‘Foreign Direct Investment in the Financial Sector – Experiences in Asia Central and Eastern Europe and Latin America’ CGFS Papers No. 25 (Basel: BIS).

    EichengreenB.R.Hausmann and U.Panizza (2003) ‘The Mystery of Original Sin’ (University of California, Berkeley, Harvard University and Inter-American Development Bank).

    GoldsteinM. and P.Turner (2004) Controlling Currency Mismatches in Emerging Economies (Washington, DC: Institute for International Economics).

    International Monetary Fund (2004) ‘Debt-related Vulnerabilities and Financial Crises: An Application of the Balance Sheet Approach to Emerging Market Countries’ Policy Development and Review Department1July (Washington, DC).

    McCauleyR.N. (2004) ‘Diversifying with Asian Local Currency Bonds’ BIS Quarterly Review (September).


This chapter draws on Morris Goldstein and Philip Turner (2004). The views expressed here are my own, not necessarily those of the BIS.

A major part of IMF surveillance should be to keep a close watch on forex exposures. They should do this both quantitatively – helping to improve statistical coverage and crosschecking with creditor data – and qualitatively – satisfying themselves that local regulators have the means to monitor mismatches and are indeed doing so.

The BIS Committee on the Global Financial System has recently published a comprehensive assessment of the impact of increased entry of foreign banks in emerging economies. See Committee on the Global Financial System (CGFS) (2004 and 2005).

Christian Keller makes this argument very effectively in IMF (2004), pp. 42-5. He shows that Peru’s official reserves plus banks’ liquid foreign assets have consistently covered two-thirds to three-quarters of the sum of the country’s short-term external debt and domestic dollar deposits.

Another 12 per cent in sterling or in yen. The Swiss franc, Canadian dollar and the Australian dollar have only 1 per cent each. These averages are for 1999 to 2004.

See the papers on debt markets in Brazil, Chile, Colombia, Mexico and Peru in BIS (2002).

Similar findings are reported by Booth (2003) and McCauley (2004).

    Other Resources Citing This Publication