Once upon a time, the world had a single currency, the globo. It was generally well managed: the Global Reserve Bank (popularly known as the Glob), under its chairman Alan Globespan, did a pretty good job of increasing the global money supply when the world threatened to slide into recession, trimming it when there were indications of inflation. Indeed, in later years some would remember the reign of the globo as a golden age. Businessmen in particular liked the system, because they could buy and sell anywhere with a minimum of hassle.

But there was trouble in Paradise. You see, although careful management of the globo could prevent a boom-bust cycle for the world as a whole, it could not do so for each piece of that whole. Indeed, it turned out that there were often conflicts of interest about monetary policy. Sometimes the Glob would be following an easy-money policy because Europe and Asia were on the edge of recession; but that easy money would fuel a wild speculative boom in North America. Other times the Glob would feel obliged to tighten money to head off inflation in North America, intensifying a developing recession in Latin America. And there was nothing regions could do about it.

Over time frustration over this impotence built up; and when the Glob failed, through policy misjudgements, to prevent a serious global recession the system broke up. Each region introduced its own currency: Europe adopted the euro, Latin America the latino, North America the gringo, and so on. But how should these local currencies be managed?

At first officials were afraid to let the new currencies be traded freely: you were only allowed to exchange latinos for euros or gringos if the government granted you a license, and licenses were given only for "legitimate" imports. But over time it became clear that this system both discouraged beneficial trade and offered many opportunities for corruption. One by one, the world's regions moved back to free convertibility of currencies. But they were still afraid of instability, so governments tried to stabilize the rates at which these currencies exchanged by buying and selling on the foreign exchange markets.

Alas, this system too turned out to have serious problems. After all, the whole point of going from a world currency to multiple local currencies was to give governments the ability to have independent monetary policies, so they could fight recessions when necessary. But a country could not simultaneously print money to fight a recession and maintain the value of its currency on the foreign exchange market. It could improve its competitive position by devaluing - say by raising the price of a gringo from 1.2 latinos to 1.5. But once latinos were freely convertible into other currencies, the mere hint that a devaluation might be in the offing would cause massive speculation against the vulnerable currency.

So what was the answer?

One answer was simply to give up the attempt to stabilize exchange rates, and just let the market do the job. The trouble was that experience showed that the market did the job badly. You might have thought that the exchange rate between, say, the euro and the gringo would be determined by the needs of trade: by North Americans trading gringos for euros in order to buy European goods, and conversely. It soon became clear, however, that mainly the market was dominated by investors - people buying and selling currencies in order to purchase stocks and bonds. And since these investment demands were highly variable, including a large component of speculation, currency values also proved unstable. Worse yet, people began to speculate on the values of the currencies themselves. The result was that exchange rates bounced around, creating uncertainty for businesses who could never be sure what their overseas assets and liabilities were really worth.

Some governments decided that this was unfortunate, but a price worth paying; for example, North America was unwilling to sacrifice domestic goals to achieve exchange rate stability, and so it practiced benign neglect toward the gringo's value. This was, of course, easier for big, relatively self-sufficient economies than for smaller currency zones. Nonetheless, some of these also seemed to thrive under freely floating exchange rates. The kangaroo tended to hop around, but Australia's economy did very well. And for a while most economists believed that floating exchange rates, while admittedly imperfect, were a workable system for just about everyone.

But when the world's poorer regions tried to behave like the First World, responding to speculative attacks on their currencies by simply letting them float, disaster struck. When the kilogram was allowed to float against the euro, nothing terrible happened: the currency fell by 15 percent, then stabilized; in effect, investors seem to have thought "Good, that's over" and been willing to start putting money back into the country. Indeed, the central bank found itself able to cut interest rates, and engineer an economic recovery. But when the latino was allowed to float against the gringo, investors panicked: it went into free fall, losing half its value in a matter of weeks. Since many companies had debts denominated in gringos, this was a financial catastrophe. So the government tried to stabilize the latino by raising interest rates to 75 percent, hoping that this would induce investors to keep their money in the country; the effect, however, was a disastrous recession, which in a way ratified the investors' panic.

And the same story played itself out repeatedly, in one country after another. Indeed, after a while the whole thing started to feel like a recurrent nightmare, with the same horrifying events replayed each time. First a country which had seemed to be doing well would find itself the sudden object of speculative attack, sometimes because of some real but probably manageable economic difficulties, sometimes because of psychological "contagion" from an economic crisis halfway around the world, and even on occasion because of a short-selling conspiracy by hedge funds. Then a team from the Global Monetary Fund would arrive, promising to save the country by lending it money, but only if it did things that were guaranteed to produce a severe slump: raising taxes, cutting spending, and increasing interest rates to punitive levels. These measures were supposed to restore market confidence, stabilizing the situation; but by depressing the economy, and often destabilizing its internal politics, they would usually precipitate a new crisis after a few months. Some countries eventually recovered from these crises, and these cases were celebrated as demonstrations of the success of the GMF's recommendations; but after four years of rolling crisis, which had devastated the economies of eight nations and counting, the world economy was starting to look like a very dangerous place.

What could the poor Latins (and Asians, and ...) do?

Some economists argued that many developing countries could still move to floating rates - that the currency collapses were actually the consequence of misguided GMF policy, and could have been avoided. And in some cases they may have had a point. On a Friday in January 1999, for example, Amazonia let its currency float, and the initial results were far better than expected: the currency fell only modestly, and the stock market soared. Nonetheless, over the weekend GMF officials demanded that the country raise its interest rates. (Were they, perhaps unconsciously, unwilling to see a non-GMF-style policy succeed, and thus call into question their previous actions?) When the rate increase was announced on Monday, panic set in, and the fragile sense of optimism was shattered.

But most economists were more pessimistic. It seemed to them that in general developing countries were, for whatever reason, held by financial markets to a different standard than First World nations; and for them floating rates did not work. So something else had to be done.

One possible answer was to achieve credibility by tying oneself to the mast: to adopt a currency board - that is, back every latino with a gringo of reserves, and pledge never ever to change the parity - or, if even this wasn't enough, to give up on having your own currency at all, and "gringoize" (or euroize) the economy. In effect, this would mean going back to a sort of inferior version of the globo standard - inferior because it would have all the weaknesses of that system, plus new problems. You see, while Alan Globespan managed the globo on behalf of the world as a whole, his successors - Mr. Gringspan, who controls the gringo supply, and Mr. Euroberg, who controls the euro supply - have more parochial concerns. (And because those concerns differ, the euro-gringo rate is highly unstable, making it problematic for small countries that trade extensively with both regions to peg to either one). And nationalism is not dead: could independent countries really be expected to swallow their pride and accept monetary subservience on a permanent basis?

Another possible answer was to reimpose exchange controls, so as to limit the vulnerability of economies to speculative attack. The reasons for avoiding such controls were as strong as ever; but countries that had maintained controls on capital movement had been noticeably less savaged by crisis than those that had not. Perhaps, in an imperfect world, the costs of controls were a price worth paying.

The worst thing to do, of course, was to put off making a choice: to try to defend a currency of suspect credibility with high interest rates, producing a recession and budget crisis that inevitably led investors to worry - despite all denials by the government - that capital controls might be the next step. And yet of course, politics and human nature being what they are, that is what most countries did when the crises came. (Less forgivably, it is also what the GMF, time and again, advised them to do).

And so the world lurched from crisis to crisis; and they all lived unhappily ever after.