The IMF: Immune
From (Frequent) Failure

By STEVE H. HANKE

The International Monetary Fund failed to anticipate the Mexican peso fiasco of 1994-95. This proved to be a huge embarrassment. But never mind. The IMF rode to the rescue. Indeed, Managing Director Michael Camdessus asserted at the time that "we must think big." And that he did. With the support of Treasury Secretary Robert Rubin, he engineered the biggest bailout in history.

But that was not enough. As has been the case with each financial crisis since the collapse of the Bretton Woods system in 1971, the ever-opportunistic IMF invented yet another raison d'être: It installed an early-warning system designed to supply more extensive and timely information to policy makers and market participants. The IMF was then supposed to keep a hawk-like lookout for signs of financial weakness and sound the alarm whenever a crisis was brewing.

***

How well has the IMF's new system worked? On July 2, Thailand devalued the baht, and shortly thereafter other currencies in Southeast Asia got hammered. Predictably, the IMF's early warning system remained deafeningly silent prior to the crisis. Indeed, the IMF's World Economic Outlook, issued in May, failed to flash red. The IMF flunked, again.

What has been the IMF's reaction? In addition to delivering yet another bailout, the IMF said that it knew the right answer all along and that it had the internal documents to prove it. Fine. But is secrecy the hallmark of an early-warning system?

The IMF's failing grade is inexcusable. The Thai baht was pegged to a basket of currencies heavily weighted in U.S. dollars. Unlike a devaluation-immune fixed exchange rate, such as Argentina's or Hong Kong's, a pegged exchange rate is not a free-market mechanism for international payments. A pegged regime is an interventionist system. It requires a central bank to manage its currency's exchange rate, the domestic liquidity and its capital account all at once. This is a tricky, if not impossible, task. Indeed, a pegged rate inevitably results in contradictory policies that invite a speculative attack.

When under siege, a peg cannot last unless interest rates are raised sky-high or foreign exchange controls are imposed. The landscape is littered with pegged regimes that have blown up and been followed by devaluations: the European Rate Mechanism (1992 and 1993), Mexico (1994), the Czech Republic (1997) and most recently Thailand, Indonesia, Malaysia and the Philippines. The IMF should have known that, as a matter of principle, the pegged systems in Southeast Asia were fatally flawed and vulnerable.

The new warning system should have set off loud warning bells in Thailand early in 1997: Its growth in domestic liquidity and inflation had been exceeding the rates of the countries to whose currencies the baht was linked. And the private sector was highly leveraged and burdened with mountains of debt denominated, in large part, in unhedged foreign currencies. The imbalances and contradictory policies were there for all to see, including the ever-vigilant speculators.

By February 1997, the speculators were starting to place large one-way bets against the baht. Consequently, Thai interest rates shot up. The punishing interest rates hit the overblown property market, causing prices to slump. This had an immediate effect on the asset quality of the notoriously fragile banking system.

The speculators knew the game was up. The only ways the Thais could hold the baht peg were to allow interest rates to climb even higher, or to impose even tougher exchange controls. But this would have caused further damage to the property market, decimated the banking system and further squeezed the highly leveraged private sector.

All this information was available in the markets. But it failed to show up on the IMF's radar screens. This should not surprise anyone. Since the breakup of the Bretton Woods system, the IMF has been busy touting the glories of central banking and fine-tuning for less-developed countries. Just look at the results: Average annual inflation in less-developed countries has been 8.6 times higher and the variability of that inflation has been 106.8 times higher than the comparable figures in the developed countries.

Not surprisingly, the IMF's record for promoting sound banking is not any better. Since 1980, more than 50 developing countries have witnessed the complete loss of their banking systems' capital, and in some this has occurred more than once. In a dozen of these countries, more than 10% of annual gross domestic product has been used to clean up the accompanying banking crises. Just since 1980, the total cleanup cost in developing countries has been a staggering $250 billion.

***

With a record like this, it's time to pull the plug on the IMF. In today's global economy, the private sector is able and willing to supply capital to the developing countries. Just five years ago, official capital flows to such countries exceeded private flows, but now private flows dwarf official flows. Who needs the IMF?

At the very least we should stop the policy of rewarding the IMF for failure by providing it with new jobs and more money after every crisis. Each time the IMF's early warning system fails, each member country should have its required capital contribution to the IMF cut by a fixed percentage, let's say 20%. With these incentives, either the early-warning system would start to function properly or the IMF would wither away.

Mr. Hanke is a professor of applied economics at The Johns Hopkins University.