To the Editor:

Martin Feldstein ("Argentina's Fall," March/April 2002) made a number of mistakes in his analysis of what went wrong in Argentina.

First, he incorrectly terms Argentina's "convertibility" monetary system a "currency board." It was actually an awkward combination of currency board and central banking features. An orthodox currency board holds net foreign reserves equal to or slightly above 100 percent of its monetary liabilities. In contrast, the Argentine central bank's reserve ratio ranged between 193 percent and 82 percent in 2001. An orthodox currency board also maintains a truly fixed exchange rate. Upon becoming economy minister in March 2001, however, Domingo Cavallo announced his intention to change the anchor currency of the peso from the dollar to a basket of the dollar and the euro. Interbank interest rates in pesos immediately doubled, reflecting distrust of Cavallo's tinkering, and they never returned to their former levels for any sustained period.

Second, Feldstein says the peso was overvalued, even though he offers no supporting evidence. One relevant measure for judging his claim is the producer price index. Using March 1991 (the month before the convertibility system began) as the base, Argentina's cumulative inflation in producer prices was never more than 15 percentage points higher than that of the United States. From December 2000 until the end of the convertibility system in January 2001, cumulative inflation was actually lower in Argentina than in the United States -- indicating an undervalued peso, if one places credence in this sort of calculation. The Economist's "Big Mac index" likewise indicated in April 2001 that the peso was undervalued against the dollar, by two percent.

Third, Feldstein says Argentina exported too little and imported too much. In fact, Argentina had trade surpluses in 2000 and 2001. Argentina's exports increased every full year of the convertibility system except in 1999, when Brazil, its largest trading partner, suffered a currency crisis and recession. (As Feldstein mentions, Argentina did have deficits in its current account, a broader measure than the trade balance. However, current-account deficits are not very useful as warning signals for countries that have fully convertible currencies. The United States has had current-account deficits every year since 1982, except for a small surplus in 1991, and does not seem to have suffered.)

Finally, Feldstein apparently confuses government debt with foreign debt. Argentina did not have a problem of foreign debt in general, but it did have a problem of government debt resulting from a combination of deficits, tax increases in 2000 and 2001 that killed an economic recovery and thereby made tax revenue lower than expected, and interest rates on the debt that rose as deficits and a shrinking economy persisted. The proper solution would have been to quarantine the government's debt problem, probably defaulting on the foreign debt (as the government did), but avoiding a currency devaluation. Instead, the government chose to contaminate the rest of the economy with its problem by devaluing the peso, freezing bank deposits, and forcibly converting dollar deposits into pesos.

Feldstein has long been a supporter of floating exchange rates. In Argentina, a floating rate and related economic measures have created so much chaos that in late March this year the retail selling rate of the peso briefly touched four pesos per dollar, versus one peso per dollar at the start of the year. A move away from floating to something else is now highly likely.

Kurt Schuler

Senior Economist, Joint Economic Committee, U.S. Congress

Martin Feldstein replies:

Kurt Schuler is a passionate supporter of the concept of a currency board. Rather than accept that Argentina's experiment with a currency board failed, he declares that what Argentina had was not an "orthodox" currency board. I am reminded of those who used to say that communism did not fail because it was never really tried.

Argentina's policy of linking the peso to the dollar at a one-to-one ratio was in trouble long before Cavallo announced that the link would shift to a dollar-euro peg when the euro reached parity with the dollar, an event that never happened.

The difference in inflation rates between Argentina and the United States is basically irrelevant to whether the peso was overvalued, since Argentina's main trading partners are Brazil and Europe. The Argentine peso was overvalued when it was frozen at one peso per dollar and became substantially more overvalued when the Brazilian real declined by 30 percent in 1999. The situation deteriorated further as the euro declined by 25 percent. Argentina's accumulated foreign debt of more than $150 billion reflects the large current-account deficits that could have been reduced or eliminated if the peso had not been overvalued. Further evidence of its overvaluation was its quick 50 percent fall when the official peg was broken.

Argentina's current-account deficit is the best measure of its current imbalance. The U.S. ability to sustain a large current-account imbalance is irrelevant because the United States borrows in its own currency. Argentina's foreign debt, like that of the crisis countries of Asia in the late 1990s, is a serious problem because it is denominated in dollars.

Contrary to Schuler's assertion, I do not confuse government debt with foreign debt. Argentina's government debt to foreign lenders exceeds $90 billion, or 90 percent of its current GDP. Large government debt would have been a problem if it had been financed domestically (because it would have crowded out domestic investment), but it would not have caused the crisis that resulted from massive foreign borrowing.

The chaos since Argentina floated its currency was due not to floating but to the conditions that had been created before the floating began. The same was true in Thailand, Mexico, Brazil, and other countries that were forced by events to abandon pegged exchange rates and shift to a floating rate policy.