Subcommittee on Economic Policy
Subcommittee on International Trade and Finance


Hearing on Official Dollarization in Latin America


Prepared Testimony of Dr. Michael Gavin
Head of Economic Research for Latin America
Warburg Dillon Read, LLC


9:00 a.m., Thursday, July 15, 1999

It is a great pleasure and an honor to appear before this committee to discuss dollarization in Latin America. As important as this issue may become for the United States, it is many orders of magnitude more important for the Latin American economies that are considering, or may consider dollarizing their economy. I congratulate the Committee and its staff for the vision that it has displayed in raising this issue for public debate in the United States.

I will divide my remarks into three main parts. First, I will briefly lay out some reasons why official dollarization, by which we mean the abolition of an existing national currency and its replacement with the US dollar, may make sense for a country in Latin America. I will then discuss the United States' interest in another country's decision to dollarize. Having concluded that dollarization by Latin America is in the enlightened self interest of the United States, I finally discuss the potential US role in the process of dollarization. Here I will focus mainly on the importance of the seigniorage, and in particular, finding a way to compensate countries for the seigniorage that they would lose if they chose to dollarize.

Two key points and five red herrings

1 . Official dollarization is an evolutionary, not a revolutionary change in Latin America's monetary arrangements. This is not to suggest that official dollarization would be a minor change in monetary arrangements, it would be revolutionary. But the fact is that many countries of the region are already heavily dollarized, in the sense that US dollars circulate and are used alongside the national currency, and that much of the banking system already operates in dollars. In Argentina, Peru and Bolivia, for example, most of the private sector deposits in the local banking system are in dollars, and the banks do most of their medium and long-term lending in dollars. Governments often denominate their debt in dollars, whether the debt is issued domestically or internationally, because investors are willing to buy long-term dollar-denominated paper but not long-term peso-denominated paper. In countries where the domestic financial system is not dollarized, governments and corporations have accumulated substantial dollar liabilities because foreign lenders are unwilling to lend in local currency. Even (in fact, especially) in countries that are currently moving toward more flexible exchange rates, there is 'creeping dollarization', as authorities promote the dollarization of their debt (as in Colombia recently) or of bank deposits (which has been announced as a possibility in Venezuela) in order to keep local investors at home. The irony is that this creeping dollarization, prompted by exchange-rate instability, increases the costs and decreases the benefit of exchange-rate flexibility. Today's flexible exchange rate regimes may, therefore, be setting in motion forces that will eventually lead to

2 . Official dollarization DOES mean giving up monetary policy, as well as the income that government's receive from printing money. Partial dollarization of the sort that we have just described can co-exist with floating exchange rates and a discretionary monetary policy, though sharp movements in the exchange rate can be severely destabilizing for partially dollarized economies. But a government that abolishes its own money and replaces it with another country's loses control over the exchange rate, interest rates, and the money supply.

3 . Dollarization does NOT require that the US change its monetary policy objectives or procedures. The US has barely noticed the gradual adoption of the dollar as a means of payment in many Latin American economies, and official dollarization - which implies that the dollar would become the only means of payment - would not change matters much.

4 . Dollarization does NOT require that the US become involved in the supervision or regulation of the dollarizing economy's financial system. This is often raised as an objection to official dollarization, even by otherwise economically literate commentators. But there is no logical connection between dollarization and financial-market regulation. Panama has successfully operated a fully dollarized economy for several decades, and the US has never involved itself in the super-vision and regulation of that country's banks.

5.. Dollarization does NOT require that the US provide lender-of-last-resort facilities to the banks or government of the dollarizing economy. See below.

6 . Dollarization does NOT mean that the local monetary authority cannot act as lender of last resort It is true that the central bank of a dollarized economy cannot print dollars, and therefore cannot print money in order to lend it to banks that experience liquidity problems. However, the central bank can provide liquidity support to local banks if it keeps on hand excess dollar reserves to use for this purpose. Holding these reserves has a cost, of course, which means that operation of a lender of last resort facility becomes a fiscal problem. But it's a fiscal problem under a discretionary policy regime as well, the main difference being that in a dollarized system, the inflation tax cannot be used as a means of financing lender-of-last-resort operations. This is a good thing, not a bad thing, especially if it makes central banks more discriminating in their lender-of-last-resort operations..

7. Dollarization does NOT require as a precondition the creation of a pristine banking system. It is sometimes argued that dollarization cannot or should not be done until weaknesses in the banking system are fully resolved. It's true that you'll have a mess if you try to run a dollarized financial system with a weak banking system. But you'll also have a mess if you try to run a non-dollarized financial rate system with a weak banking system. I would argue that banks are likely to come under more, not less, stress under a weak monetary regime, and that dollarization would thus complement, rather than undermine efforts to strengthen the banks. The problem with waiting until all banking problems are resolved before dollarizing is that the exchange-rate and interest-rate volatility that is inherent in non-dollarized financial systems may itself weaken the banks, forcing indefinite postponement of the creation of a more stable financial environment. Whatever the monetary regime, if the banks need fixing they should be fixed as rapidly as possible, but weak banks should not be used as an excuse for the perpetuation of a weak monetary system.

Why dollarization may make sense for a Latin American economy

Dollarization brings with it a loss of monetary autonomy - once an economy has dollarized, it loses the capacity to control the exchange rate, the domestic money supply, or interest rates, all of which would be determined by the US Federal Reserve. In a few moments I will suggest that this loss of monetary policy autonomy is more apparent than real, but at least in theory, one would not want to dollarize unless there were some offsetting advantages to compensate for the loss of control over monetary policy. I think that dollarization brings with it two key benefits, integration and stability.

Dollarization would promote economic integration. The evidence is compelling that deepening the economic ties between a developing country and its industrial country counterparts can promote economic development in the less developed economy. Outward-oriented trade policies are consistently associated with economic success. No less important is financial integration, which permits abundant industrial country capital to work with, and raise the living standards of, workers in emerging markets where capital is scarce.

However, the existence of many different national. currencies, fluctuating in value from day to day, impedes this economic integration. Switching to a common currency would simply and reduce both the costs and the risks of international trade, thus promoting trade integration and economic development in Latin America. I suspect that simply switching to a common currency would promote hemispheric integration as effectively as would many thousands of man-hours devoted to the ongoing negotiations over a free trade agreement for the Americas.

More important, I think, would be the impact of dollarization on domestic financial systems, and their capacity to support deeper financial integration. Under the status quo, currency risk segregates national banking systems. With national currencies, foreign capital that enters the country and is intermediated through the banking system, is subject to currency risk. This means that foreign lenders have to make their lending decisions on the basis of monetary considerations that have little or nothing to do with the inherent quality of the enterprise that is seeking credit. Industrial country investors do not want to lend very much, even to very sound businesses, in an environment where a monetary accident can suddenly halve the value of even a well-conceived investment. The existence of many national currencies thus curtails the amount and kind of cross-border lending that can be done, to the detriment of capital-hungry businesses in developing economies. And the cost of the resulting shortage of credit and high interest rates is borne, of course, mainly by small and medium-size enterprises whose growth is stunted.

This point is certainly not lost on businessmen in Latin America. I recently spent four days visiting our clients in Guadalajara, Monterrey and Leon, in the industrial heartland of Mexico. Much of the discussion was about dollarization which is, as the Committee knows, is a topic of enormous interest to many Mexicans at the moment. They listened patiently to the list of macroeconomic pro's and con's, but the problem that was really on their mind was the constraint that having to borrow in pesos at high interest rates from the Mexican banking system imposed upon their competitiveness and growth. They understood at a deep intuitive level what it would mean for them if they had the same access to the pool of dollar savings that their counterparts to the North now enjoy. Though they did not use these words, dollarization made sense to these businessmen as a means to promote a deeper integration of credit markets, giving them the opportunity to compete for capital with their industrial-country counterparts on the basis of their business prospects alone, unhampered by barriers to international lending that are created by potentially devastating currency fluctuations.

In joint work with several co-authors from the Inter-American Development Bank, we recently investigated the link between exchange-rate regimes, the depth of local financial markets, and real interest rates in Latin America. This work corroborated these businessmen's insight - we found that fixed exchange rate regimes have in fact been associated with significantly deeper financial markets and lower real interest rates than have flexible exchange rate regimes.

Official dollarization can promote economic stability. This assertion seems a little counter-intuitive, and in fact I'm staking out a contrarian position here. The convention wisdom is that, since monetary policy can in principle be used to stabilize an economy, adoption of a currency board or official dollarization must imply the loss of some capacity to stabilize the economy. I do not think that this logic is valid. The reason is that under any regime that allows for discretion in monetary policy, there is no institutional barrier to devaluation - policymakers retain the option to devalue whether they intend to use the option or not. But this option to devalue is tremendously destabilizing when economies come under stress.

The year 1998 illustrated this point perfectly. During the summer and fall of last year, the entire region came under stress because of commodity price shocks associated with the Asian crisis, and the world financial panic that was associated with the Russian devaluation and default. Did the countries operating on relatively flexible exchange rate regimes - Chile, Colombia, Mexico, and Peru, use that exchange rate flexibility as a 'shock absorber'? No. When their currencies came under attack they defended them with very high interest rates. They had to do this because simply allowing their currencies to collapse would have left the countries with unmanageable financial problems, since many corporations and households have substantial dollar debts, and would have left the country with a very costly inflation problem.

The interesting and important point is that all four of these countries were forced to raise interest rates higher, and keep them high for longer, than Argentina, the country in the region with the most rigid and credible commitment to an exchange rate target. Why? Because even though the authorities in the other countries made clear their intention to defend the currency, and backed up the intention with high interest rates, investors in the country knew that the authorities could change their mind at any time. Much as they might have liked to do so, the central banks could not credibly renounce the option to devalue. Investors, facing the risk that the central bank could at any time change its mind and let the currency go, were willing to keep their money at home only if the central bank compensated them for the risk of devaluation with cripplingly high interest rates. It is these interest rates that lie behind the deep recessions from which Chile, Colombia and Peru are only now beginning to emerge.

Of course, Argentina was also thrown into recession by the Russia crisis and the near-collapse of Brazil, its most important trading partner. This proves that credible commitments to an exchange rate regime are no panacea. Of course they are not. Nothing is a panacea. But without the currency board, it is almost certain that fears of devaluation would have forced the Central Bank to raise interest rates to the level that were reached in Chile, Colombia and Mexico, worsening the Argentine recession.

Thus, while the credible exchange-rate commitment was no panacea, it provided Argentina with some insulation from the world economic and financial shock. Meanwhile, those countries operating under more flexible exchange rates, who had thought that they enjoyed monetary autonomy, found themselves unable to use that autonomy when they needed it.

Of course this is just one episode. But the pattern emerges more generally. In the joint work referred to above, we also looked at the correlation between real (inflation-adjusted) interest rates in Latin American economies and real interest rates in the US over the past 25 years. We found a strong positive correlation, confirming that when US rates rise, Latin American rates need to follow. Contrary to what one might expect, the correlation was somewhat higher under flexible than under fixed rates, suggesting that flexible exchange rates may actually provide less monetary autonomy than is conventionally assumed.

What are US interests?

The direct impact of official dollarization in Latin America is likely to be small. The United States thus benefits mainly to the extent that the policy generates deeper hemispheric integration - an avowed goal of US policy and one that offers much potential for the United States economy - greater economic and financial stability, and more rapid development of our neighbors in the hemisphee .

As integration proceeds, the greater stability of Latin American economies that is likely to be generated by a more accident-proof monetary regime becomes increasingly important. Monetary mishaps that lead to a collapse of wages and living standards in Latin American economies impose important collateral damage on US workers, who may suddenly be forced to compete with Latin American labor that has been impoverished by a currency collapse. Other than workers in Latin America, no group has more at stake in the avoidance of monetary crisis in the region do US workers.

What are the costs?

The main potential cost for the US is the possibility that, if Latin America becomes part of a de-facto dollar zone, the US Federal Reserve will come under political pressure to adapt monetary policy to the requirements of those countries that have dollarized, as well as those of the US. This is not a compelling concern. First, the Fed is an independent entity whose job it is to fend off political pressures. Second, under the status quo Latin American economies are already excruciatingly sensitive to US monetary policy actions, and the Fed is internationalist enough - because it understands the dangers of a financial crisis abroad for the US economy - that is already comes under pressure to avoid monetary policies that would do great damage to the rest of the world, including Latin America. If, as we believe, official dollarization would make Latin American economies more robust, and thus better able to withstand fluctuations in US monetary policy without falling into crisis, the Fed may even come under less pressure to modify policy than it does now.

What policy stance should the US adopt?

With respect to the decision to dollarize, the US should mainly stand aside and allow individual countries to make the decision through their own democratic process. However, given that the US has an enlightened self interest in allowing dollarization to proceed, we should certainly remove unnecessary barriers to dollarization. The most important of these is the issue of seigniorage.

Seigniorage is the fiscal revenue that a government receives from printing money. This is not a trivial consideration. Consider the case of Argentina. In that country, roughly $15 billion of peso notes and coins circulate in the country. Under the miles of the convertibility system, the country holds $15 billion of international reserves to back those notes and coins. (The central bank also holds additional reserves to cover the rest of the monetary base, but that does not concern us here.) The Central Bank doesn't hold those reserves as dollar bills, it holds them in the form of interest-bearing foreign bank deposits and securities. At an interest rate of 5 percent, this represents about $750 million per year.

If Argentina were to dollarize without some cooperative agreement with the US, the central bank would have to sell $15 billion of reserves to obtain the cash that it would need to retire the pesos in circulation by exchanging them for dollars. Since it would no longer have the $15 billion in bank deposits, it would lose $750 million per year in interest payments. Who would gain? The US Treasury. This is because the Fed would in effect be selling non-interest earning dollars to the Argentines in exchange for interest bearing Treasury bills.

Seigniorage does matter. $750 million may not seem like an enormous sum - it's about 0.5 percent of Argentina's national income, which is probably small by comparison with the macroeconomic costs and benefits of dollarization. But in the political context in which the decision whether or not to dollarize would be taken, it's a big number. It is on the order of 20 percent of the education budget. It might very well be hard to resist an appeal to reject dollarization and spend the $750 million thus saved by increasing education spending by 20 percent, even though the harder to quantify economic benefits of devaluation might dwarf the budgetary implications.

Finding a way to compensate dollarizing economies for this lost seigniorage would thus remove a financially marginal but potentially potent political obstacle to dollarization, leaving the decision whether or not to dollarize to be decided on other, more germane issues.

Compensating a country for lost seigniorage would not cost the US taxpayer. It is important to recognize that a decision by the United States to rebate, or share the seigniorage that countries would lose - and the US would otherwise gain - from dollarization, would not involve a true cost to the United States. If the US does decide to share seigniorage, and a country decides to dollarize, the United States' financial position would be the same as it would be if a country decides not to dollarize. True, the US would be worse off financially than it would be if it decided not to share seigniorage and a country were to dollarize anyway. But if the US does not decide to share seigniorage, the odds of official dollarization are very low, so this comparison is not relevant. In any event, in a US context the amounts of money that we are talking about are small by comparison with the other broader economic and financial issues at stake.

How to share seigniorage. Broadly speaking, there are two mechanisms for the sharing of seigniorage. In the first, the dollarizing economy would be required to exchange its international reserves for the US currency that is required to replace the domestic currency that is being extinguished. The US would appropriate an annual amount equal to the estimated loss of seigniorage - equal to the loss of interest on the international reserves that the country's central bank has lost - and transfer it to the central bank or finance ministry of the country that has decided to adopt the dollar. This method has two key disadvantages. First, it injects the seigniorage issue into the annual budgetary process, generating the potential for misunderstanding and political frictions down the road and exposing the dollarizing economy to the risk that some future Congress will change its mind and fail to compensate the country for lost seigniorage. Also, the loss of international reserves associated with the purchase of the required US dollars would leave the central bank illiquid, without the dollar reserves that it might need to satisfy changes in the public's demand for currency.

A second method is cleaner. Rather than transfer money to compensate for lost seigniorage, the US would make a once-and-for-all swap of US currency for the country's local currency, allowing the country retire the old currency without expending its international reserves. (Provision would also have to be made for future increases in currency demand, but this is a relatively minor issue.) The central bank of the dollarizing economy would not have to expend its reserves to obtain the currency that it needs to replace its own, thus leaving the financial system more liquid and less vulnerable. And the US Congress would not have to revisit the seigniorage issue in each and every future year.

What is the risk? The risk is that the economy that has dollarized will, at some point in the future, decide to reverse the dollarization, and re-introduce its own currency. If they did that, the US dollars that have been floating around being use for transaction purposes in the local economy would suddenly become available to be spent on US goods and services. This would represent a multi-billion dollar gift to the economy in question, which is most definitely not appropriate. To avoid this, the contract or treaty that defines the sharing of seigniorage should specify that, if the country were to reverse the dollarization and introduce a local currency of any kind, the US would have a right to demand a reversal of the swap arrangement, forcing the country to pay the US, in dollars, for the pesos that the US has been holding, at the original exchange rate. Of course, there is some risk that the country may be unwilling or unable to comply with this demand, which provides one reason why the US may want to be somewhat selective in its application of the mechanism to share seigniorage, and even, perhaps, to include Maastricht type criteria for eligibility.

Once again, I thank the Committee for the opportunity to testify on this important topic. I congratulate the Committee its foresightedness and vision for taking up this topic, and place myself at its disposal if there are any questions that I might address.


Home | Menu | Links | Info | Chairman's Page