Originally Published in the

March/April issue of Foreign Affairs

Like death and taxes, international economic crises cannot be avoided. They will continue to occur as they have for centuries past. But the alarmingly rapid spread of the 1997 Asian crisis showed how shifting perceptions alone can disrupt even fundamentally stable countries. In the wake of Russia's default, skyrocketing interest rates in emerging markets underlined these economies' vulnerability to investor skittishness. Unfortunately, there is no international "911" that emerging markets can dial when facing economic collapse. Neither the International Monetary Fund (IMF) nor a new global financial architecture will make the world less dangerous. Instead, countries that want to avoid a devastating rerun of the 1997-98 crisis must learn to protect themselves. And self-protection requires more than avoiding bad policies that make a currency crisis inevitable, for the threat of contagion makes even the virtuous vulnerable to currency runs.

Liquidity is the key to financial self-help. A country that has substantial international liquidity -- large foreign currency reserves and a ready source of foreign currency loans -- is less likely to be the object of a currency attack. Substantial liquidity also enables a country already under a speculative siege to defend itself better and make more orderly financial adjustments. The challenge is to find ways to increase liquidity at reasonable cost.

The need for enhanced liquidity stems from the inevitable absence of an international lender of last resort -- a dependable source of credit when a country needs additional foreign currency. This is not the case within a national economy, where the central bank provides liquidity to the banking system. A solvent commercial bank -- one with assets that exceed its liabilities -- does not have enough liquid assets to pay depositors if they were all to demand their funds at once. But a central bank can prevent a run of depositors on the bank by following the advice provided more than a century ago by British economist Walter Bagehot: to lend "freely, at penalty interest rates, on appropriate collateral."

However, there is no international equivalent that lends a stable foreign currency to temporarily illiquid but ultimately solvent countries that could eventually repay loans with export earnings. Although the IMF has organized large credit packages for the crisis-stricken countries, it does not "lend freely" at a time of need. Instead, it disburses its funds gradually as the stricken countries meet a series of detailed conditions. Because the IMF=s conditions are onerous and complex, private creditors will not know whether funds will actually be available. The IMF packages therefore cannot bolster confidence that the country will have enough foreign exchange reserves to defend itself against future speculative attacks. Another fundamental difference between the IMF and national central banks is that a national central bank can create as much domestic currency as its banking system wants, whereas the IMF=s ability to provide dollars or other currencies is constrained by its limited capital.

True, the G-7 announced a plan in October 1998 for an "enhanced" IMF lending facility that would disburse funds quickly to countries before a crisis occurs. But it would make funds available only to "pre-approved" countries who were "pursuing strong IMF-approved policies." None of the ill-fated Asian economies would have passed this test in 1997. In practice, the IMF's new approach might simply try to impose the same kind of intrusive economic restructuring as it did in the post-crisis plans in Asia. (See my "Refocusing the IMF," Foreign Affairs, March/April 1998.) But if it did, few countries would likely participate. The recent Brazilian devaluation also suggests that the new plan would fail. The IMF inaccurately described its November 1998 loan to Brazil, which did not formally belong to the G-7 arrangement, as "preemptive." Although some of the funds were available immediately, the rest were to be parceled out only conditionally as Brazil met a variety of stringent fiscal conditions. This lack of dependable liquidity may explain why Brazil was forced to devalue its currency just two months later.

More generally, the G-7 proposal is fundamentally misconceived because it does not recognize that most emerging market countries need increased liquidity long before the crisis stage to reduce their risk of crisis. Even with its recently increased capital, the IMF does not have the funds to finance such a plan. Nor is there any plan to assure private-sector involvement in rolling over short-term debt, as the G-7 assumes would happen. True, its other proposal -- that national governments force private creditors to continue lending to a crisis-ridden country -- would diminish an investor's tendency to lend irresponsibly in good times, or the so-called "moral hazard" problem. But the fear of being locked into risky and deteriorating loans would also substantially reduce the long-term availability of credit to emerging market economies. Instead of establishing a new global financial architecture, the world needs new ways for emerging markets to increase their own international liquidity.

IT PAYS TO BE FLEXIBLE

"Emerging markets need increased liquidity long before the crisis stage."

First, it is important to understand how exchange-rate policies helped frame the Asian crisis. Before the 1997 turmoil, many of the soon-to-be-stricken countries attempted a fixed-rate policy -- pegging their currency to the dollar to achieve a lower cost of credit and a low rate of inflation. While their intentions were noble, they forgot a basic economics lesson: when it comes to exchange rates, flexibility holds many advantages. Although sharp swings can debilitate national economies, normal fluctuations help by reallocating global capital to better uses. For example, if Mexican business investment becomes more profitable, higher demand for credit will drive up Mexican interest rates and make Mexican stocks more attractive. Foreign investors then buy pesos to invest in Mexican stocks and bonds while foreign companies buy pesos to acquire or build more businesses in Mexico. Rising demand for the peso strengthens the currency and reduces Mexico's exports (by making them relatively more expensive) and increases its imports (by making them relatively cheaper). In short, the stronger peso provides an extra flow of capital to invest in local businesses. By extension, when Mexican investment becomes less profitable, investors sell the peso and direct capital to other countries where it can be used more productively.

Importantly, you do not need to own a currency to sell it. Foreign investors who think the peso will fall against the dollar can borrow pesos, sell them for dollars, and then invest the dollars; they then expect to sell their dollars after the peso's fall and repay the original peso debt with cheaper pesos, a strategy known as "selling short." This means that peso borrowers, not just domestic residents and foreign investors who own the currency, can drive the peso's value down. Hence, expectations often drive exchange rates, causing the fluctuations in currency values to be large and rapid.

The Asian experience shows how much serious damage a sharp currency drop can cause by increasing the relative value of foreign-currency-denominated debt. When the Thai baht fell from 25 baht to the dollar in June 1997 to more than 50 baht to the dollar in January 1998, that meant that the number of baht that a Thai company needed to repay its dollar-denominated debt more than doubled. In addition, a falling currency raises the price of imports and allows domestic producers to raise prices as well. This sparks higher inflation unless the government raises interest rates, which depresses domestic production and employment still further. The larger a country's trading sector, the more acute the problem becomes, especially for small economies with large flows of exports and imports. In the end, many governments try using foreign exchange reserves and monetary policy to avoid what they see as unfounded currency declines or to slow the process of adjustment. This means that if private investors want to sell that currency, the government may use its dollar reserves to buy it or raise interest rates to induce investors to keep the currency.

To complicate matters, many governments control their exchange rates to maintain a noninflationary monetary policy. Since a lax monetary policy that causes inflation also erodes the value of the national currency, a government that commits itself to a fixed exchange rate against the dollar is implicitly committing itself to the same low inflation rate as the United States. If domestic business and labor believe in this commitment, they will be less likely to try to raise prices and wages. Although a government could achieve low inflation in principle without such an external anchor, governments have often found that this is a useful way of reducing inflation and keeping it down. Despite this virtuous intention, however, fixed-exchange-rate policies often precede a currency crisis by causing an explosive trade deficit.

BAHT HUMBUG

There are actually four distinct causes of financial crises: an unsustainable current account deficit, a balance-sheet crisis, a banking implosion, and irrational "contagion" turbulence. Enhanced liquidity can address all of these ills, but the trick is achieving that extra liquidity at an acceptable cost. Thailand, the first country to succumb to the Asian financial flu in 1997, offers a classic example of a current account crisis. Thailand fixed its exchange rate at 25 baht to the dollar in 1987 and promised to use whatever policies were necessary to keep it there. But by the mid-1990s the baht was grossly overvalued, causing a current account deficit -- the trade deficit plus the net interest due on foreign debts -- equal to eight percent of Thailand's GDP, one of the highest anywhere in the world. How did this happen?

A current account deficit must be financed with an equally large inflow of funds from the rest of the world. Because the Thai government promised a fixed baht-dollar exchange rate, Thai banks and businesses felt comfortable borrowing dollars to finance activities in Thailand. Similarly, foreign creditors were willing to lend in order to reap the yields of the slightly higher Thai interest rates. The banks and the foreign lenders did recognize the risk they were taking: the huge current account deficit could be sustained only as long as foreign investors and lenders kept sending money to Thailand and increasing their total Thai exposure. If the day came when they were no longer willing to do so, the current account would have to shrink to whatever could be financed with available capital inflows. In turn, reducing the current account deficit would require a fall in the baht to make Thai exports cheaper and imports more expensive. The cheaper baht would hurt those Thai banks and businesses that had borrowed in dollars. Foreigners who lent baht would also lose, as would those who lent dollars to Thai banks and businesses now unable to repay their dollar obligations.

In 1997 this potential risk became a reality. Eventually fear began to outweigh greed, and foreigners became too nervous to keep lending. As the baht began to fall, other skittish investors with baht assets sold baht as well. Others borrowed baht in order to sell it on the assumption that the baht would continue to slide. The government tried to support the currency but eventually exhausted its foreign reserves. By the time the baht stopped falling, it had plummeted by more than half.

What Thailand forgot was that a fixed exchange rate set too high inevitably causes a ballooning trade deficit, followed by an attack on the currency, a loss of foreign exchange reserves, and a corrective fall in the exchange rate. Thailand could have avoided exhausting its reserves and eased its painful adjustment had it abandoned the fixed exchange rate earlier and allowed the baht to settle at a more appropriate level in the market. (By the end of 1998, the market panic had indeed subsided and the baht had settled at 36 baht per dollar.) But while it is easy to criticize Thailand in hindsight, many other countries still refuse to abandon fixed rates or opt for a devaluation, fearing the resulting inflation, steeper interest rates, and greater foreign debt burdens. These countries cling to the hope that an overvaluation might correct itself without a change in the nominal exchange rate if domestic prices fall or if the dollar slips. This strategy worked for Argentina, which maintained its "overvalued" fixed dollar-peso exchange rate during the past decade and stayed competitive because its domestic producers lowered the cost of Argentine goods by increasing productivity. And an overvalued currency and current account deficit can temporarily help keep import prices down and export prices up while providing a capital inflow to finance profitable domestic investments. But Thailand's time had run out. After more than a decade of current account deficits, Thailand mistakenly hoped that it would be able to go on financing them and boost its productive capital rather than sustaining them with domestic savings alone. Ultimately, financial realities made a baht devaluation inevitable, especially after Japan's economic slump and declining yen sharply reduced Japanese demand for Thai exports.

However, a country with sufficient liquidity -- namely, large foreign exchange reserves and access to additional dollar or yen credits -- can try to sustain the high value of its currency by buying it, maintaining or even raising its value temporarily to "punish" speculators and others who bet that the currency will fall. Speculation against a currency is expensive because the interest rates on the currency being sold are often substantially higher than the rate on the dollar or yen. Since speculators have much to lose if the currency does not weaken, they bet only against currencies that look particularly vulnerable. Large reserves and access to substantial additional foreign loans can reduce that vulnerability and increase confidence in the currency, thereby reducing the likelihood of speculation -- but this requires much larger reserves than most emerging market countries have traditionally held. In the end, the decision whether to devalue is more than a purely economic issue with a technically correct, IMF-prescribed solution. It is a balancing act of risks that is the essential responsibility of a sovereign government.

WON OVER

A country with a floating exchange rate and a current account surplus could still come under attack if its short-term foreign currency debt exceeds its foreign currency reserves. It may be solvent in that it can service that debt with future export earnings, but it remains temporarily illiquid because it lacks the funds to meet its immediate obligations. If foreign creditors have confidence in the country and its currency, they will continue to roll over the short-term debts or extend their maturity. But if enough foreign creditors worry that the country will not meet its obligations because other foreign creditors will not continue to lend, the government is forced to default, reschedule its loans, and devalue its currency to increase export earnings enough to repay its debts more quickly.

South Korea met this fate in 1997. After it deregulated part of its banking system in the early 1990s, some investment banks used their new freedom to borrow foreign currencies with short-term maturities (less than one year) and invest those funds in Korean won assets and high-risk, high-yielding foreign securities like Russian bonds. By late 1997, the value of these short-term obligations far exceeded the government's foreign currency reserves. Nervous international investors saw this disparity on South Korea's balance sheet and began shedding won assets. South Korea could have stopped its inappropriate foreign borrowing and earned enough foreign currency on strong exports to repay the existing foreign obligations. But it could not have done so quickly enough if all foreign creditors had demanded payment promptly and refused to roll over existing obligations. In short, South Korea was solvent but illiquid -- that is, able to meet its obligations in the longer run but unable to do so immediately.

Had South Korea not allowed its foreign exchange reserves to deteriorate so much, no crisis would have erupted. But global financial markets saw that it lacked liquidity and that a won devaluation was likely to occur. As a result, they sold won assets or borrowed won to sell for dollars. Before the end of the year, the currency collapsed and the economy's precarious health was in the hands of the IMF.

TAKE THE MONEY AND RUN

Another calamity that plagued many Asian countries in 1997 and 1998 was a widespread collapse of the banking system, often in tandem with the other crises described above. Many of these failures stemmed from a rise in the burden of dollar-denominated debt following a decline of the local currency versus the dollar. The plummeting currency also hurt banks with dollar-denominated loans (as in Thailand) because corporate borrowers had dollar debts but only local currency income.

Often a banking failure was waiting to happen even before the currency turmoil. Many Asian banks had accumulated so many bad loans and poor investments that their liabilities exceeded the true value of their loans and other assets. Banks can retain deposits only as long as depositors trust that they can withdraw their funds on demand. When depositors get nervous that a bank will be unable to meet future withdrawal requirements, their fear becomes self-fulfilling. Depositors may also worry that the national central bank will help temporarily illiquid banks but not insolvent banks with liabilities exceeding the true value of assets. But because depositors do not know the true health of individual banks, the safest thing to do is to withdraw their funds from all domestic banks and deposit them in safer investments abroad. If a bank run ensues, fleeing depositors will push the local currency down as they sell their local currency. If the government tries to buy back the local currency, it will deplete its foreign reserves and increase the risk of a complete exchange-rate collapse. (However, a domestic banking crisis need not cause a currency run if depositors can shift their funds from home to local banks owned abroad and backed by strong overseas banks and central banks, as in Argentina today.)

The obvious advice is that governments should prevent commercial banks from favoring particular industries and tighten banking supervision so that widespread insolvencies do not occur. Unfortunately, the first suggestion runs contrary to many national development strategies, while the latter is hard to implement, as banking failures in the United States, Japan, and many European countries have shown. But sufficient liquidity, either through foreign currency reserves or access to foreign credit, would let a government restructure and recapitalize its banks without experiencing a currency crisis. The more international liquidity a government has, the less depositors will feel that they must rush to convert their currency before the reserves are depleted. And preventing a currency decline can be the best way to protect bank solvency.

RAGIN' CONTAGION

Finally, a currency can crash even without current account imbalances, balance sheet mismatches, or banking system failures. Currencies that are not overvalued may still succumb to contagion based on geography, trade patterns, or investor portfolio strategies. Such declines can trigger inflation, an economic downturn caused by an increased debt burden, or both. But again, a country with sufficient foreign exchange can intervene to defend its currency from such "irrational" pressures. The very presence of liquidity can reinforce investor confidence and reduce the likelihood of irrational speculation. The traditional alternative -- defending the currency by raising interest rates -- only depresses the domestic economy and can even exacerbate capital outflows by sparking a local stock-market crash, as it did in Thailand and South Korea in 1997.

THREE STEPS TOWARD SELF-HELP

An emerging market economy can inoculate itself against future attacks by taking the right measures to boost liquidity: reducing short-term foreign debt, accumulating liquid reserves, and organizing a collateralized credit facility. Each of these measures carries a stiff price, which is why many countries are tempted instead to do nothing and to run the risk of a currency crisis. But the Asian experience shows that the tremendous costs of a currency collapse often exceed the price of maintaining enough liquidity to ward off future crises. These three steps are mutually reinforcing, and a cost-effective strategy probably involves doing some of all three.

Avoid High Short-Term Foreign Debt. Banks, companies, and governments in emerging markets often borrow with short-term maturities in dollars or yen because such loans have lower interest rates than longer-term bonds or domestic currency loans. But such borrowing is particularly risky. Foreign lenders may refuse to renew lending if conditions in the borrowing country appear to change or if lenders become more risk averse in general, as they did after the August 1998 Russian default. Just the accumulation of too much foreign debt relative to reserves may trigger a currency crisis, as South Korea learned in 1997. And if a currency does fall, the rise in dollar-denominated debt can ignite corporate bankruptcies and recession.

Chile successfully avoided this pitfall. It discouraged short-term capital inflows by requiring Chilean borrowers to deposit 20 percent of foreign loans with the government for six months without interest -- a substantial penalty for short-term borrowing but not for long-term capital inflows. The policy did not stem the total capital inflow to Chile but caused investors to shift funds to less volatile assets. Although this strategy raised the cost of capital for Chile's private borrowers by forcing them to rely on longer-term loans (with higher interest rates), it successfully reduced the risk of a currency crisis.

High short-term foreign debt is not the only source of risk. In countries without capital controls, like most of Southeast Asia and Latin America, local residents can convert any amount of their cash or bank deposits from the local currency to foreign currencies. A bank can increase the amount of convertible deposits simply by lending more. Hence, limiting short-term debt to less than the sum of foreign exchange reserves is not enough to ensure sufficient international liquidity. Although the unwillingness of foreign creditors to roll over existing loans may be the most probable danger, the magnitude of widespread capital flight by domestic residents and speculation by foreigners may pose the greater risk. In short, reducing short-term foreign debt is a good idea but not enough to prevent a run on the currency.

Accumulate More Foreign Exchange Reserves. The most direct way for a country to achieve liquidity is to accumulate substantial amounts of liquid foreign reserves. China's $140 billion plus in reserves sends a strong signal to international investors and domestic residents that China will not be forced to devalue the yuan. Many countries with floating exchange rates hold substantial reserves to reduce the risk of currency crises -- even though they theoretically have no need for reserves since the market alone determines their currency's course.

Reserves are costly in two ways. First, accumulating reserves by having exports exceed imports requires a cut in domestic consumption and investment. Second, when the government buys foreign currency from its exporters and issues domestic bonds in exchange, it generally pays a much higher rate of interest on its domestic bonds than it receives on its reserves, which it usually invests in low-yielding U.S. Treasury bills. Accumulating reserves by running a trade surplus also takes time. And holding export earnings as reserves rather than investing them in domestic infrastructure or making the funds available for private investment sacrifices a substantial real rate of return for the nation.

"China's $140 billion in reserves sends a strong signal to investors."

However, a country can accumulate reserves more quickly, without reducing real domestic investment, by borrowing abroad with longer-term maturities and investing those funds in liquid international securities. China, with $160 billion in foreign debt, has implicitly adopted this approach. But this strategy can also be quite expensive. Mexico now pays an 11 percent interest rate on a 20-year bond, six percentage points more than it would earn on an investment in U.S. Treasury bills. That spread of six percentage points means that an addition of $30 billion, which would double its current reserves, would cost Mexico $1.8 billion a year -- nearly half of one percent of its GDP. Although this price may seem high, it must be compared to the potential losses from a currency crisis. The price tag may also overstate the true national cost of building up reserves, because a rise in reserves often lowers the cost of private debt and equity capital. China, awash in reserves, pays interest of only about seven percent on its long-term dollar bonds.

A country could substantially lower the net cost of holding reserves by investing them in liquid assets with higher yields than short-term U.S. Treasury bills. U.S. corporate stocks had a real yield of seven percent from 1945 to 1995, more than six percentage points higher than the real yield on U.S. Treasury bills over the same period. Although corporate stocks are riskier than Treasury bills, a country could lower its overall risk by increasing its foreign exchange reserves and investing some of them in stocks rather than solely in Treasury bills.

More generally, a government should not judge the adequacy of its reserves in relation to the value of imports. A common reserve goal of, say, six months of imports ignores the fact that currency crises are about capital flows, not trade financing. What matters is the value of reserves relative to the potential selling of assets by speculators even if the country's fundamental economic conditions do not warrant a currency deterioration.

Create a Collateralized Loan Facility. A country's ability to borrow large sums of foreign currency on short notice is virtually as useful as holding an equal amount of reserves. The challenge is finding sufficient credit from the private sector. The G-7 was right to emphasize the desirability of a precrisis loan fund but wrong to expect that the IMF and the G-7 governments could provide the loans. The IMF has less than $200 billion of total lending capacity, not large in relation to the 1998 package of $42 billion for Brazil alone. Japan's plan to help Asian governments by guaranteeing their debts involved only $15 billion of Japanese funds. In contrast, American banks alone have more than $5 trillion in outstanding loans and investments. The G-7 called for private creditors to continue lending to emerging markets during economic stress. But given the losses of many banks on the heels of the crisis, it is not hard to imagine why banks and bond holders would try to reduce their exposure under those circumstances. Some countries with substantial state assets slated for privatization could use the proceeds of future sales to back new bond issues. (See my "How to Save the Ruble" in the Financial Times, July 8, 1998.) But what can countries without such assets do?

The only reliable way to maintain private lending and increase credit is to create a collateralized credit facility: an agreement that provides collateral to protect the providers of credit. Although there is no substitute for sound economic behavior, creditors value good collateral more than the appearance of virtue. And with solid collateral, there is no need to argue about whether a country is unable to meet its debt service requirements or just unwilling to do so.

Done properly, such a facility could substitute for an international lender of last resort, lending freely at penalty interest rates against good collateral. The problem is to identify assets that banks and bond holders could rely on in the case of default. It cannot be real estate or other domestic assets that would be difficult for creditors to seize and convert into dollars or other hard currencies. Instead, the most suitable collateral would be the export earnings in hard currency of the domestic firms. This is now common in many private credit arrangements between firms in emerging economies and their industrial-country creditors. A collateralized facility was used by the United States and Mexico in 1983 and 1995, with a dollar loan secured by the receipts from the Mexican government's future oil sales. (Argentina now holds an essentially uncollateralized $7 billion line of credit from the foreign banks that operate in that country, but has difficulty in maintaining that arrangement.)

A collateralized credit facility would be a voluntary arrangement between an emerging market government and a group of foreign banks. Participation by the IMF or the World Bank on the same terms would create even greater confidence since countries are very reluctant to default on payments to those institutions. The participants would determine the maximum amount of foreign exchange credit available and agree in advance on the market-based interest rate or spread that the borrowing country would pay. The government would have the option to borrow these funds at any time during a two-year period, subject to withdrawal on two years' notice by any of the parties. The government would also pay each participant for the option of obtaining credit under less favorable circumstances. The potential lenders could participate either on equal terms or with a more elaborate structure that would allow some creditors to take more risk -- for example, being the last to be repaid -- in exchange for higher fees and interest. If it becomes necessary to use the loan facility, the borrowing government would require that all of its export earnings be repatriated immediately and paid to a trustee institution like the Bank of England or the Federal Reserve. The trustee would transfer a fraction of these receipts to the creditors to service the newly incurred debt. The institution would then transfer the remainder of the export receipts to the emerging market government, which would in turn pay the exporters the value of their exports with a combination of the local currency and the currency of the export earnings. To make this work on short notice, the borrowing country must pass laws in advance to allow the diversion of export earnings if it uses the credit facility. It would also need currency controls when the loan takes effect, permitting only importers and others with specific needs to convert domestic funds into foreign exchange. The plan would work even better if the G-7 countries required importers in their own countries to divert payments for imports to the trustee.

A collateralized credit facility is much less expensive than accumulating and holding foreign exchange reserves -- as long as it is not actually used. If it is used, the capital controls and the payment diversions would involve large administrative costs and open up the possibility of waste and corruption. A firm time limit on such borrowing therefore is important. In the end, a country must decide whether deterring irrational speculation and potential contagion -- and giving the country time to handle crises that do occur -- is worth the cost of the potential credit.

ANOTHER WAY OUT?

What about alternatives to self-protection through increased liquidity? I have already commented on the inadequacy of past IMF programs and the newly proposed "enhanced lending facility." Three other alternatives also deserve comment: economic isolation, currency boards, and dollar or yen zones.

Economic Isolation. Malaysia's prime minister, Mahathir bin Mohamad, has imposed complete controls indefinitely on all capital inflows and outflows. This not only denies Malaysia the ability to supplement domestic savings with foreign capital but discourages foreign firms from making direct investments in Malaysia for fear that controls will later prevent repatriation of profits. Efforts to evade controls will only lead to waste and corruption. The argument that isolation allows a country to stimulate its economy through lower interest rates fails to convince. Malaysia could achieve the same stimulus with targeted tax incentives that offset the effect of high interest rates.

A Currency Board. Argentina and Hong Kong operate a modified currency board system in which they hold one dollar for every unit of local currency distributed as cash or as reserves in the commercial banks. Everyone has the right to exchange that currency for dollars. A currency board system, if strictly enforced, would prevent a successful run on the currency because a drop in foreign exchange reserves would force the central bank to reduce the money supply and raise interest rates. The government never has to devalue because interest rates keep rising as long as the country loses reserves -- until interest rates are high enough to get investors to stop selling the currency. In principle, this self-regulating mechanism discourages speculators, so the interest rates need never rise to high levels.

The success of a currency board as a deterrent to speculators depends on market confidence that the government will let interest rates rise as long as foreign exchange reserves dwindle, no matter how much damage those high rates do to the economy. If an ensuing collapse in domestic demand and a sharp rise in unemployment induces the government to bend the rules and permit the central bank to issue bank reserves without full foreign currency backing, the automatic confidence factor disappears. The currency board then becomes an empty promise of the government not to devalue or to pursue an inflationary monetary policy. In short, whatever its virtue as an anti-inflation signal, a currency board still cannot prevent currency crises. Indeed, if the Hong Kong government eventually succumbs to the deepening recession by devaluing the Hong Kong dollar and reducing interest rates, Argentina's currency board would probably lose credibility as well. The theory of the currency board would then return to the textbooks of monetary history.

Dollar or Yen Zone. Another alternative for avoiding currency crises is to abandon the local currency completely and to replace it with the dollar or the yen. Something similar has already happened in Europe, where the euro has replaced the national currencies. Could Mexico adopt the dollar as its national currency or Thailand adopt the yen?

There is a fundamental difference between these emerging market countries and Europe. The eurozone countries still have access to a lender of last resort. If commercial banks in Italy or Portugal are in trouble, their central banks will provide as many euros as the depositors want and a widespread bank run will be averted. Likewise, while Mexican commercial banks hold bad loans amounting to 20 percent of GDP, the Mexican banking system remains intact because the public knows that the Bank of Mexico can generate enough pesos to replace the deposits of any bank that fails. But if Mexico relabeled those peso deposits as dollar deposits, the Mexican central bank could not provide all the dollars that depositors demanded. The result could be widespread bank runs and ultimately a currency crisis if Mexicans took their funds out of the country to safer, foreign banks. The only way for Mexico to adopt the dollar as its currency would be to have American banks take over Mexican banking services and implicitly cede control of Mexican monetary policy to the U.S. Federal Reserve. Even if the Mexicans were willing to go along, a currency union of two very different economies would cause serious economic problems for Mexico.

In the end, emerging market countries must realize that they can control their financial health to a considerable extent. They must first kick the bad economic habits of the past -- bloated current account deficits, excessive short-term foreign currency borrowing, and shaky banking systems -- to reduce the risk of devastating currency crises, painful exchange-rate fluctuation, and externally driven swings in domestic interest rates. There is no substitute for sound economic policies: moderate trade deficits, low short-term foreign debts, and strong banking systems. At the same time, even responsible policies can be undone by the powerful forces of market contagion, shifting risk aversion, and irrational speculation. Because the IMF and other international organizations cannot act as lenders of last resort, emerging market countries must protect themselves through increased liquidity. The measures discussed above -- reducing short-term capital inflows, increasing foreign exchange reserves, and establishing collateralized credit facilities -- can all fill this essential role. Each of these options is expensive but far less painful than the damage of currency crises. Through financial self-help, emerging markets can take their economic fate into their own hands.