In an Uncertain World (41 page)

Models can be a useful way of looking at markets and can provide useful input to making decisions. But ultimately traders have to make judgments because reality is always far messier and more complicated than even the most sophisticated models can capture. In fact, LTCM's models may have been valid, over a long enough time frame. As a theoretical proposition, yield spreads probably would have returned to the mean, and I gather that many of LTCM's positions would have worked out in time. But LTCM was essentially betting that a return to normal would come without some prior highly aberrational move. The unusually high degree of leverage LTCM employed meant that the firm lacked the staying power to weather severe temporary aberrations. Creditors would require additional margin as spreads moved against LTCM. LTCM's forecasts might be vindicated long after it had gone broke.

I remembered this kind of situation well from 1986, when Steve Friedman and I had taken over responsibility for the fixed-income division at Goldman Sachs. As in the LTCM case, the problem then wasn't just that one company had a set of bad positions. Traders at other firms had similar kinds of positions, because they all used similar models and similar historical data. When positions began to move against them, they all wanted out at the same time, exacerbating the movement. And since the major players already had these positions, there were no buyers. That meant that traders and investors had to unload other, better investments to obtain cash. This selling skewed the ordinary relationships and patterns that traders expected. Bond spreads that according to historical norms should have contracted instead got wider, and spreads that should have widened got narrower.

Everyone who had similar positions lost money. But LTCM was faced with massive losses that threatened to become much larger than the remaining capital the firm held. The immediate public policy question this raised was what kind of harm a forced liquidation of LTCM's assets could do. In normal circumstances, governments shouldn't worry about the tribulations of any particular firm or corporation. But if a situation threatens the financial system, some kind of government action might be the best among bad choices. No one wanted to rescue LTCM's partners or investors. But there was a concern that liquidating such large positions could lead to a general unraveling of the markets. With the hedge fund's creditors—Chase, Citigroup, Goldman Sachs, Bear Stearns, Morgan Stanley, Merrill Lynch, and many others—all selling into the same decline, the entire financial system could freeze up, with a spillover into the real economy as confidence was damaged and businesses and consumers found credit less available and more expensive.

The ideal solution would have been for LTCM's creditors to agree to extend their loans on terms that might be somewhat harmful to each of them but less harmful to all than a default—as in South Korea. But doing anything to promote this kind of agreement raised tricky problems. The Fed regulates financial institutions and wouldn't want to be seen as coercing the lenders to act. Yet without some outside pressure, LTCM's creditors would almost surely not come together in their own self-interest. The banks collectively would benefit from working out an agreement that would prevent LTCM from going into default. But each of the banks stood to benefit even more by free riding on whatever agreement was reached—that is, by pulling out at full value while the others made some sacrifice to keep LTCM intact.

New York Federal Reserve president Bill McDonough convened the heads of the big investment and commercial banks at the Fed's New York headquarters. He walked a fine line, calling the CEOs of the country's biggest banks in but then leaving the room so they could work out the details on their own. After a lot of jockeying, fourteen different institutions agreed to provide a total of nearly $4 billion in additional credit to LTCM, with strict terms attached. This capital infusion gave the hedge fund breathing room to liquidate its positions in a more orderly fashion. Although I did not share the view that a collapse of LTCM was likely to lead to systemic disruptions, I thought the concerns—and Bill's actions—were sensible and appropriate, given the general market and economic duress at the time.

The broader public policy question arising out of the LTCM mess was whether anything could be done to reduce the probability and severity of this kind of event in the future. This was a frequent topic of discussion among Larry, Alan, and me, with some of the discussion taking place in meetings of the Financial Markets Working Group, which also included Bill McDonough; SEC chairman Arthur Levitt; Brooksley Born, the chair of the Commodity Futures Trading Commission; the heads of the other principal financial market regulatory bodies; and Gene Sperling from the NEC. Some members of this group thought that derivatives—instruments such as options, futures, and forwards whose value depends on the performance of an underlying security, currency, or commodity and whose value can change in complicated ways that is hard for even experienced traders to anticipate—by their nature could pose a systemic risk. Others thought the unrestricted leverage available to hedge funds such as LTCM was a problem. Some thought neither was a problem.

I thought both derivatives and leverage could pose problems. I had been involved with derivatives from the pioneering days of the founding of the Chicago Board Options Exchange. Derivatives serve a useful purpose by providing a means to manage risk more effectively and precisely, but they can create additional problems when the system is stressed. One way to contain those risks is by limiting the permissible leverage of buyers and sellers of derivatives. If you think periodic market excesses are inevitable because human nature is likely to lead to excess, you should try at least to limit the damage to the system. Capital requirements and margin requirements—both leverage limits—help to do that, both by decreasing the size of positions and by increasing the amount of money backing each position.

Larry thought I was overly concerned with the risks of derivatives. His argument was characteristic of many students of markets, who argue that derivatives serve an important purpose in allocating risk by letting each person take as much of whatever kind of risk he wants. That is right in principle, but it is not the whole story. Throughout my career, I had seen situations where derivatives put additional pressure on volatile markets (for example, through the additional selling in the stock market that can occur when portfolio managers sell calls to arbitrageurs, who in turn hedge by shorting stock against the calls for protection as the market falls). I also thought that many people who used derivatives didn't fully understand the risks they were taking—the situation we had found ourselves in at Goldman in 1986. Larry's position held together under normal circumstances but seemed to me not to take into account what might happen under extraordinary circumstances. Of course, Larry thought I just wanted to keep markets the way they were when I'd learned the arbitrage business in the 1960s—his point about “playing tennis with wooden racquets” again.

   

THE ASIA CONTAGION finally hit Latin America in late summer 1998, adding to the sense of gloom during the darkest period of the crisis. In September, people began to talk about Brazil, the world's eighth-largest economy, the way they'd been speaking about Russia a few months earlier. Brazil had a fixed currency, large current account and budget deficits, a great deal of short-term debt coming due, and an impending presidential election. The turmoil in the bond market, exacerbated by Russia and LTCM, was making it difficult and expensive for Brazil to roll over its debt. Underlying worries about debt sustainability and fiscal control in the country's provinces—problems that had been present for some time—now came to the fore. Foreign banks started to reduce their lines of credit, and foreign investment slowed to a trickle. The central bank's once very large foreign currency reserves were being depleted rapidly as the government clung to an exchange rate fixed against the dollar. Market rumors began to fly that Brazil was on the verge of defaulting on its debts, devaluing, or both.

With financial markets under great strain worldwide, we were fearful that either event could have severe consequences for the global economy. The IMF, U.S. Treasury, and Federal Reserve and other major countries began to work intensively on putting together a convincing rescue package. The stakes were high enough to warrant putting U.S. government money at risk alongside that of the IMF, provided the IMF could reach agreement with Brazil on a workable reform program.

Unlike in Russia, we had considerable confidence in Brazil's commitment to reform. Under the leadership of President Fernando Cardoso, Brazil's historic hyperinflation problem—the subject of my senior thesis in college—seemed finally to have been solved. The nation, which accounted for around 45 percent of South America's GDP, had taken important steps toward implementing sound macroeconomic policies since recovering from the debt crisis of the 1980s and had a political leader prepared to call upon his people to support measures that, while difficult in the short term, were necessary for stability and growth. Cardoso's highly capable and experienced economic team—headed by Finance Minister Pedro Malan—recognized that the alternative would have been even worse.

But there was one enormous issue in dispute: the country's exchange rate policy. President Cardoso adopted some significant reforms that made sense in light of Brazil's fiscal deficit, such as cutting spending and raising taxes. But the Brazilians were unwilling to devalue the real. Cardoso had been elected in 1994 largely on the promise of his Plan Real, which had fixed Brazil's currency to the U.S. dollar and, by so doing, had succeeded in taming a 2,700 percent annual inflation rate. Brazilian policy makers had a deep-seated and understandable fear of their economy's historical demon and felt that going to a float risked reviving inflation. But we thought floating the real—which was significantly overvalued—could prove essential to making an IMF program work. This left us with the unhappy choice between proceeding with a huge IMF program in a situation where the odds seemed unfavorable or trying to force a devaluation by refusing to lend unless Brazil agreed to a floating exchange rate. I thought there was only a fifty-fifty chance, at best, of an IMF support program working with a fixed, overvalued currency. But we were hesitant to push the issue too hard, because the inflation problem was genuine and Cardoso and his team could be right that letting the real decline would tip Brazil back into an inflationary spiral.

As with Russia, Indonesia, and South Korea, some in the U.S. administration thought a robust IMF program, backed by money from the ESF, should settle the problem. If Brazil had
X
billion in reserves and
Y
billion in loans coming due, a
Z
billion program should cover it. But there's one other number that is incalculable and can swamp all those other numbers in the calculation: the size of potential domestic capital flight. For some reason, most discussions of the subject ignore what can be the largest issue in crisis recovery. If Brazilians started losing confidence in the country's currency and converted their reals to dollars, then the potential would exist for vast additional outflows.

But proceeding with a loan to Brazil that fall was nonetheless compelling, even without the devaluation and despite the unfavorable odds. In Russia, concerns about politics and nuclear weapons had been the key. In the case of Brazil, the external environment tipped the balance. The fragility of the global economy was so great that I believed that a program was worth trying. Even buying only a few months of breathing room could be enough to get through a dangerous period of financial market strain. And demanding that Brazil float the real in the fall of 1998 in the face of the continuing instability in world markets could be very risky. A disorderly currency devaluation in Brazil could lead to additional currency disruption and contagion elsewhere, and Brazil could go into a deeper crisis. After Asia, Russia, and LTCM, and in an already fragile environment, another such shock to the system could lead to a true global meltdown.

To me, there was never a real question about whether or not to help Brazil, although the decision to go ahead without floating the exchange rate was a hard one. I remember sitting at a meeting in Larry's office and thinking to myself that this was another choice among bad options:
The odds aren't in our favor. But even if it fails, it could still be the right decision. While the chances of success are small, the risks of inaction are enormous. And even if we fail to save Brazil, we can probably defer the impact of the collapse for six or eight months, and that will more than justify the effort.
I did not believe these arguments would forestall criticism. If the Brazilian program failed, nobody was going to look at the quality of the decision or the benefit of deferral. They were just going say: you took a whole bunch of the American people's money and threw it down the drain.

I remember this as another case where Clinton's understanding and intuition were rather extraordinary. Larry and I had gone to the Oval Office and laid out our case for supporting an IMF program and putting U.S. money on the line alongside the IMF funds, not just as a backup option. We had persuaded other major governments to do the same. Clinton agreed that we should go ahead. But then he said, “You know, I felt very good about the Mexico program. I had a feeling it was going to work. I don't feel as good about this one working. It just seems chancier to me.”

The large IMF program did help Brazil for a while, during November and December. But then the instability returned with a vengeance in January 1999, when Brazil did not follow through on some of its planned fiscal actions, monetary policy was loosened prematurely, and Brazil's second-largest state refused to make debt payments that were due to the central government. Brazil's deteriorating fiscal position sent investors running for the exits once again, further diminishing the government's foreign reserves. With the coffers emptying, the central bank governor resigned. Brazil tried, as Mexico had, to modify its exchange rate policy in a staged way that didn't work. A controlled devaluation seldom, if ever, works once trouble has begun—and didn't in this situation, collapsing after only two days. The real plunged dramatically, losing 10 percent of its value in a day and more than 30 percent by the end of January. In less than three weeks, the new central bank governor was also out.