The Accidental Theorist (13 page)

Currency-crisis connoisseurs cherish the memory of George Brown, Britain’s Secretary of Economic Affairs in the mid-1960s—the man who blamed his troubles on the “gnomes of Zurich.” (He was misinformed: the relevant gnomes are actually in Basel.) But we may have to remove Brown from his pedestal, to make room for Malaysian Prime Minister Mahathir Mohamad. Last month Malaysia’s neighbor Thailand, after months of promising that it wouldn’t, devalued the baht, and spooked investors began selling Malyasian ringgit (and Philippine pesos, Indonesian rupiahs, and so on) as well. This provoked an outburst on Mahathir’s part that surely counts as an instant classic. Where Brown was vague about both the identity of the villains and their motives, Mahathir had a full-fledged conspiracy theory: The U.S. government had prompted palindromic speculator George Soros to undermine Asia’s economies, because it wants to impose Western values (like democracy and civil rights) on them. And Mahathir’s ministers expanded on his remarks with a rhetoric that was unusual for a government with a long-term interest in maintaining the goodwill of international investors: Currency fluctuations are caused by “hostile elements bent on…unholy actions” that constitute “villainous acts of sabotage” and “the height of international criminality.”

These remarks were entertaining both because, as far as we can tell, Soros was
not
a major player in the crisis (indeed, he seems to have taken a bit of a bath by failing to anticipate this one), and because in the early 1990s one of the world’s most ambitious and reckless currency speculators was…Malaysia’s government-controlled central bank, which only got out of the business after losing nearly $6 billion.

Currency crises often provoke hysterical reactions in government officials. One day your country’s economy is humming along nicely, your bonds are AAA, you have billions of dollars in foreign exchange reserves socked away. Then all of a sudden the reserves are depleted, nobody will buy your paper, and you can only keep money in the country by raising interest rates to recession-inducing levels. How can things go wrong so fast?

The standard response of economists is that to blame the financial markets in such a situation is to shoot the messenger, that a crisis is simply the market’s way of telling a government that its policies aren’t sustainable. You may wonder at the abruptness with which that message is delivered. But that, says the canonical model, is simply part of the logic of the situation.

To see why, forget about currencies for a minute, and imagine a government trying to stabilize the price of some commodity, such as gold. The government can do this, at least for a while, if it starts with a sufficiently large stockpile of the stuff: All it has to do is sell some of its hoard whenever the price threatens to rise about the target level.

Now suppose that this stockpile is gradually dwindling over time, so that far-sighted speculators can foresee the day—perhaps many years distant—when it will be exhausted. They will realize that this offers them an opportunity. Once the government has exhausted its stockpile, it can no longer stabilize the price—which will therefore shoot up. All they have to do, then, is to buy some of the stuff a little while before the reserves are gone, then resell it at a large capital gain.

But these speculative purchases of gold or whatever will accelerate the exhaustion of the stockpile, bringing the day of reckoning closer. So the smart speculators will try to get ahead of the crowd, buying earlier—and thereby running down the stocks even sooner, leading to still earlier purchases…. The result is while the government’s stockpile may decline only gradually for a long time, when it falls below some critical point all hell suddenly—and predictably—breaks loose (as actually happened in the gold market during 1969).

With a bit of imagination this same story can be applied to currency crises. Imagine a government that is trying to support the dollar value of the ringgit—or, what is the same thing, to keep a lid on the price of a dollar measured in ringgit—through foreign exchange market “intervention,” which basically means selling dollars to keep their ringgit price down. And suppose the government’s policies are, for whatever reason, inconsistent with keeping the exchange rate fixed forever. Then there is a complete parallel with the previous story, with foreign exchange reserves taking on the role of the gold stockpile. And by the same logic as before, we can conclude that speculators will not wait for events to take their course: At some critical moment they will all move in at once—and billions of dollars in reserves may vanish in days, even hours.

The abruptness of a currency crisis, then, does not mean that it strikes out of a clear blue sky. In the standard economic model, the real villain is the inconsistency of the government’s own policies.

Is Mahathir’s complaint therefore unadulterated nonsense? No: As Art Buchwald once said of his own writing, it is adulterated nonsense. The truth is that speculators may not always be quite as blameless as the standard model would have it.

For one thing, markets aren’t always cool, calm, and collected. There is abundant evidence that financial markets are subject to occasional bouts of what is known technically as “herding”: Everyone sells because everyone else is selling. This may happen because individual investors are irrational; it may also happen because so much of the world’s money is controlled by fund managers who will not be blamed if they do what everyone else is doing. One consequence of herding, however, is that a country’s currency may be subjected to an unjustified selling frenzy.

It is also true that the long-run sustainability of a country’s policies is to some extent a matter of opinion—and that policies that might have worked out given time may be abandoned in the face of market pressures. This leads to the possibility of self-fulfilling prophecies—for example, a competent finance minister may be fired because of a currency crisis, and the irresponsible policies of his successor end up ratifying the market’s bad opinion of the country.

All of this, in turn, creates a
possible
way for private investors with big enough resources to play a nefarious financial game. Here’s how it would work, in theory: Suppose that a country’s currency is in a somewhat ambiguous situation—its current value might be sustainable, or it might not. A big investor quietly takes a short position in that country’s currency—that is, he borrows money in pounds, or baht, or ringgit, and invests the money in some other country. Once he has a big enough position, he begins ostentatiously selling the target currency, gives interviews to the
Financial Times
about how he thinks it is vulnerable, and so on; and with luck provokes a run on the currency by other investors, forcing a devaluation that immediately reduces the value of those carefully acquired debts, but not the value of the matching assets, leaving him hundreds of millions of dollars richer.

Speculative sharp practice, in short,
can
play a role in destabilizing currencies. But how important is that role in reality?

Well, George Soros pulled the trick off in Britain in 1992; but as far as anyone knows even he has done it only once. True, it was an amazing coup: He is supposed to have made more than a billion dollars. It’s also true, however, that there were good reasons for the pound’s devaluation, and it is unclear whether Soros really caused the crisis or was merely smart enough to anticipate it. Maybe he brought it on a few weeks early.

The other currency crises of the nineties—and it has been a great decade for such crises—have taken place without the help of sinister financial masterminds. This is no accident; opportunities like the one Soros discovered in 1992 are rare. They require that a country’s currency be clearly vulnerable, but not yet under attack—a narrow window at best, since there is a sort of Murphy’s Law in these things: If something can go wrong with a currency, it usually will. Financial markets are not in the habit of giving countries the benefit of the doubt.

Does this mean that there is no defense against speculative attack? Not at all. In fact, there are two very effective ways to prevent runs on your currency. One—call it the “benign neglect” strategy—is simply to deny speculators a fixed target. Speculators can’t make an easy profit betting against the U.S. dollar, because the U.S. government doesn’t try to defend any particular exchange rate—which means that any obvious downside risk is already reflected in the price, and on any given day the dollar is as likely to go up as down. The other—call it the “Caesar’s wife” strategy—is to make very sure that your commitment to a particular exchange rate is credible. Nobody attacks the guilder, because the Dutch clearly have both the capability and the intention of keeping it pegged to the German mark.

Oh yes—there is also a third option. You can erect elaborate regulations to keep people from moving money out of your country. Of course, if investors know that it will be hard to get money out, they will be reluctant to put it in to begin with. There is a case to be made—an unfashionable case, but not a totally crazy one—that it is worth forgoing the benefits of capital inflows in order to avoid the risk of capital outflows. But Asian leaders uttered not a word of complaint when they were receiving huge inflows of money, much of it going to dubious real estate ventures; only when irrational exuberance turned into probably rational skittishness did the accusations begin.

So Mahathir’s claims that he is the victim of an American conspiracy are just plain silly. He has nobody but himself to blame for his difficulties. Or at least that’s what George, Bob, and Madeleine told me to say.

A Note on Currency Crises

I often run into people who assert confidently that massive speculative attacks on currencies like the British pound in 1992, the Mexican peso in 1994–1995, and the Thai baht in 1997 prove that we are in a new world in which computerized trading, satellite hookups, and all that, mean that old economic rules, and conventional economic theory, no longer apply. (One physicist insisted that the economy has “gone nonlinear,” and is now governed by chaos theory.) But the truth is that currency crises are old hat; the travails of the French franc in the twenties were thoroughly modern, and the speculative attacks that brought down the Bretton Woods system of exchange rates in the early seventies were almost as big compared with the size of the economies involved as the biggest recent blowouts. And currency crises have been a favorite topic of international financial economists ever since the 1970s. In fact, it is one of
my
favorite topics—after all, I helped found the field.

The standard economic model of currency crises had its genesis in a brilliant mid-seventies analysis of the gold market by Dale Henderson and Steve Salant, two economists at the Federal Reserve. They showed that abrupt speculative attacks, which would almost instantly wipe out the government’s stockpile, were a natural consequence of typical price stabilization schemes. In 1977 I was an intern at the Fed, and realized that Salant and Henderson’s story could, with some reinterpretation, be applied to currency crises that suddenly wipe out the government’s reserves. A bit later Robert Flood, now at the IMF, and Peter Garber of Brown produced the canonical version of that conventional story.

The key lesson from that conventional model is that abrupt runs on a currency, which move billions of dollars in a very short time, are not necessarily the result either of irrational investor stampedes or of evil financial manipulation. On the contrary, they are the normal result when rational investors contemplate the implications of unsustainable policies.

Some economists, however—notably Berkeley’s Maurice Obstfeld and Barry Eichengreen—argue that the standard model is too mechanical in its representation of government policy; and that the more complex motives of actual governments make speculation a more uncertain and perhaps more pernicious affair. Self-fulfilling crises, in which a currency that could have survived is nonetheless brought down, are a hot topic these days. But everyone agrees that a sufficiently credible currency will never be attacked, and a sufficiently incredible one will always come under fire.

The very first real paper I ever wrote in economics was a piece entitled “A Model of Balance of Payments Crises,” written in 1977. It was a theoretical analysis of the reasons why attempts to maintain a fixed exchange rate typically end in abrupt speculative attacks, with billions of dollars of foreign exchange reserves lost in a matter of days or even hours. What I had in mind at the time were the attacks that brought down the Bretton Woods system in 1971 and its short-lived successor, the Smithsonian agreement, a year and a half later. It seemed to me then that the main interest of the paper would be historical; I did not expect to see attacks of that scale and drama again.

Based on a talk at the Group of 30, London, April 1997.

Luckily, I was wrong. I say “luckily,” because as the founding father of what has long since become the academic industry of speculative attack theory, my citation index goes up every time another currency crisis materializes, trailing its tail of economic analyses and rationalizations. And the decade of the 1990s has so far produced a bumper crop of crises, which is fun for me if not for the ministers involved.

But why are there so many crises? Why haven’t finance ministers, central bank governors, and so on learned to avoid them? To understand the durability of the speculative attack problem, you need to be aware of an underlying dilemma in exchange rate policy. To describe this dilemma, I find it helpful to think in terms of a little matrix of opinion, defined by the different answers to two questions.

The first question is whether flexibility on the exchange rate is useful. A country that fixes its exchange rate, in a world in which investors are free to move their money wherever they like, essentially gives up the opportunity to have its own monetary policy: Interest rates must be set at whatever level makes foreign exchange traders willing to keep the currency close to the target rate. A country that allows its exchange rate to float, on the other hand, can reduce interest rates to fight recessions, raise them to fight inflation. Is this extra freedom of policy useful, or is it merely illusory?

The second question is whether, having decided to float the currency, you can trust the foreign exchange market not to do anything crazy. Will the market set the currency at a value more or less consistent with the economy’s fundamental strength and the soundness of the government’s policies? Or will the market be subject to alternating bouts of irrational exuberance (to borrow Alan Greenspan’s famous phrase) and unjustified pessimism?

The answers one might give to these questions define four boxes, all of which have their adherents. Here is the matrix:

Suppose that you believe that the policy freedom a country gains from a floating exchange rate is actually worth very little, but you also trust the foreign exchange market not to do anything silly. Then you will be a very relaxed guy: You will not much care what regime is chosen for the exchange rate. You may have a small preference for a fixed rate or better yet a common currency, on the grounds that stable exchange rates reduce the costs of doing business; but you will not lose sleep over the choice. This is, I believe, the position of so-called “real business cycle” theorists, who believe as a matter of faith in the efficiency of private markets, and don’t believe that monetary policies of any kind make much difference to how those markets function.

Suppose, on the other hand, that you believe that the freedom gained by floating is very valuable, but that financial markets can be trusted. Then you will be a serene floater: You will believe that freeing your currency from the shackles of a specific exchange rate target, so that you can get on with the business of pursuing full employment, is unambiguously a good thing. This was the view held by many economists in the late 1960s and early 1970s; indeed, I remember as an undergraduate picking up from my teachers a definite sense that they regarded the whole Bretton Woods system as a barbarous relic, a needless straitjacket on macroeconomic policy.

You will be equally sure of yourself if you believe the opposite: that foreign exchange markets are deeply unreliable, dominated by irrational bouts of optimism and pessimism, while the monetary freedom that comes with floating is of little value. You will then be a determined fixer, seeking to lash your currency immovably to the mast—best of all, by creating a common currency shared by as many countries as possible. This is, as I understand it, the position of most of the central bankers of continental Europe.

But what if you believe both that the freedom that comes from floating is valuable and that the markets that will determine your currency’s value under floating are unreliable? Then you will be a nervous wreck, subject to stress-related disorders. You will regard any choice of currency regime as a choice between evils, and will always worry that you have chosen wrong.

Well, the last decade or so has given us a lot of evidence that bears on the two questions. True, the world rarely gives us clean natural experiments—although in some cases it comes pretty close. (For example, when Ireland decided in 1987 to stop pegging the punt to the pound sterling and start pegging it to the Deutsche mark, the prices of Irish goods abruptly stopped tracking the U.K. price level and starting following the German index instead—a fairly dramatic demonstration that money matters and therefore that monetary policy can matter, too, even in a small country.) I believe, however, that there is enough evidence to make a clear pronouncement: The nervous wrecks have it. Yes, the monetary freedom of a floating rate is valuable; no, the foreign exchange market cannot be trusted.

Let me start with the value of a floating rate. The classic case against there being any such value is that any attempt to make use of monetary autonomy will quickly backfire. Imagine, in particular, a country that drops its commitment to a fixed exchange rate, and uses that freedom to cut interest rates (which of course leads to a decline in the value of its currency). Ardent defenders of fixed rates insist that instead of an increase in employment, the result will very soon be a surge in inflation, wiping out both any gain in competiveness vis-à-vis foreign producers and any stimulus to real domestic demand. This was a position that seemed to be supported by a reading of the evidence in such cases as the repeated depreciations of the pound in the 1970s, or the devaluation of the Swedish krona in 1982.

To be honest, I never accepted that interpretation even of those events. But in any case a succession of more recent events has made it harder and harder to sustain this view of the way the world works. In the mid 1980s the U.S. dollar quickly dropped from 240 to 140 yen, from 3 marks to 1.8; there was not a hint of the surge in inflation that some had predicted. Many Europeans discounted this experience—after all, America, with its huge economy and relatively small reliance on international trade, is a special case; matters would be different if a European country tried the same thing. Then came the crises of 1992. I was assured by many French economists that Britain’s departure from the Exchange Rate Mechanism would swiftly be punished; and Sweden’s finance minister personally assured me that a depreciation of the krona would be an inflation disaster. Yet Britain got off scot-free—and even Sweden, when it depreciated a few weeks after my conversation with Ms. Wibble, suffered none of the predicted pain.

Nobody would claim that devaluing your currency is always and everywhere a good thing. What we can say, based on experience, is that the freedom to pursue an independent monetary policy that comes with a flexible exchange rate is indeed valuable as long as you start from a position of low inflation, and as long as domestic price increases are restrained by the presence of a lot of excess capacity. These may sound like restrictive conditions, but they aren’t: They are exactly the conditions under which you would want the freedom to cut interest rates in the first place.

So far so good—but then there is the problem of the foreign exchange market.

Economists are, for understandable reasons, attached to the “efficient markets” theory of financial markets—a theory that attributes all fluctuations in financial prices to news about current or expected future fundamentals. As a sort of benchmark, an explanation of first resort for how asset prices behave, this theory has been enormously useful and productive—not to say lucrative, since much of the modern risk-management industry is based on that theory. But in many markets, very much including the foreign exchange market, the theory has become more or less impossible to believe as a literal description. Part of the argument involves technical tests—the anomalies just keep piling up, and the efforts to rationalize those anomalies within an efficient market framework have become an ever more desperate matter of adding epicycles.

But beyond these technical assessments, there is the simple question of plausibility—what it sometimes known as the “smell test.” Can anyone come up with a good reason—which is to say real changes in what the markets knew about fundamentals—that justified a shift of the yen/dollar exchange rate from 120 plus in 1993, to 80 in 1995, then back to 120 plus in 1997? Isn’t it far more reasonable to view the whole thing as a case of exchange traders following a trend they themselves had created? And of course these were not small swings: For Japanese industry the absurdly strong yen was a body blow.

So we are, as I said, firmly placed in the “nervous wreck” box of my little matrix. And that is the reason why the speculative attack industry—or perhaps I should say industries, since it comprises both those like me who write about such attacks, and, far more important, those like George Soros, who actually engage in them—continues to thrive.

After all, if ministers were serene floaters, they would simply treat the foreign exchange market with benign neglect; and they would therefore offer no target for speculative attack. If they were utterly determined fixers, they would do whatever was necessary to beat back speculative attacks—and the speculators, knowing this, would rarely attack in the first place. What creates an environment in which Soros can make money and I can write papers is the prevalence of finance ministers who decide to fix their currencies, but are suspected of being less than total in their commitment to that policy.

Let me say a bit more about the kind of speculative attacks that flourish in this environment; for there have been some major developments in the theory since I and a few other people started it twenty years ago.

The original models of such attacks imagined a country that was known to be following policies ultimately inconsistent with keeping its exchange rate fixed—for example, printing money to cover the budget deficit. That is, the eventual doom of the fixed rate was not in question to an informed observer. The speculative element came from the incentive of investors to anticipate the inevitable. Knowing that eventually the currency would drop in value, investors would try to get out of it in advance—but their very effort to get out of a weak currency would itself precipitate the collapse of the fixed rate. Understanding this, sophisticated investors would try to get out still earlier…the result was that a massive run on a currency could be expected at a time when it might still seem to have enough reserves to go on for many months or even years.

This analysis is still the canonical one, but recent discussion has emphasized three further concerns.

First, some economists have argued for the importance of self-fulfilling currency crises. They imagine a country whose government is prepared to pay the cost of sticking to an exchange rate indefinitely under ordinary circmstances, but which is not willing and/or able to put up with the pain of keeping interest rates high enough to support the currency in the face of speculators guessing that it might be devalued. In that case the fixed rate will survive if investors think it will; but it will also collapse if they think it will.

Second, there is the obvious point that if markets are subject to irrational shifts in opinion, to running with the herd, this applies as much to speculative attacks on fixed exchange rates as to gyrations in the value of flexible exchange rates. One remarkable fact is that there is not a sign in the markets that the great currency crises of the nineties were anticipated—that is, until only a few weeks before Black Wednesday in Britain, or the Mexican crisis of 1994, investors were cheerfully putting their money into pounds or pesos without demanding any exceptional risk premium. (This in spite of the fact that quite a few economists were actually warning in each case that a crisis might be in prospect). Then quite suddenly everyone wanted out. Was this because of some real news—other than the observation that everyone else suddenly wanted out?

Finally, there is the return of the gnomes of Zurich. Finance ministers whose currencies are under attack invariably blame their problems on the nefarious schemes of foreign market manipulators. Economists usually treat such claims with derision—after all, the British politician who coined the phrase was so ignorant that he didn’t even know that the gnomes actually live in Basel. But nobody who has read a business magazine in the last few years can be unaware that these days there really are investors who not only move money in anticipation of a currency crisis, but actually do their best to trigger that crisis for fun and profit. These new actors on the scene do not yet have a standard name; my proposed term is “Soroi.”