As demonstrated by the severe market sell-off that followed the news, there is an even worse story than has yet been told in the collapse of Long Term Capital Management of Greenwich, Conn., the hedge fund deluxe that scorned the fellowship of anyone with less than $10 million to invest. People dislike the idea that nobody is in charge but have learned to live with it—that's capitalism, where the invisible hand moves in mysterious ways. This story, however, means that nobody knows what's going on—and that really scares the market.
LTCM was founded four years ago by the former chief trader for Salomon Brothers and two sidekicks, one of whom was advertised as the first bond trader in history to receive a $24-million bonus for one year's work. It was run by an executive committee including two Nobel Prize-winners in economics and a former vice chairman of the Federal Reserve Board who had been a Harvard professor before that.
The work that these people did to make money was programming computers to spot "anomalies"—unusual differences between the prices of financial instruments with well-established historical price relationships. One did not have to know the cause of such relationships—the fact that the computer found them was good enough. Big bets would then be placed in such a way that they paid off when prices "converged" to their usual patterns.
It is probably worth noting that the work done at LTCM, while not illegal or sinful, was totally without redeeming social value. This is not "investing"; it enables the production of no goods or useful services. It is betting.
The formulas that funds like LTCM use give a very high probability of small gains, but also a very low probability of a big loss. If the bets are big enough and the turnover is fast enough—and if you can play with money borrowed at low interest rates—the small gains can add up to a very large return on each set of bets. In its first three years, LTCM made profits half again as much as the rise in the stock market. In January, the fund gave $2.8 billion, about a third of its total capital, back to the shareholders, professedly because the opportunities to make money on anomalies had declined, more likely because the insiders had decided to keep a bigger piece of the profits for themselves. All of this, of course, must be totally hush-hush. If others know what you're doing, they'll copy you and blunt your edge. Thus the banks in supplying fortunes to funds like LTCM have been schooled not to ask the uses to which the money will be put.
But the mathematical models that power these funds fail when something significant happens that the computer doesn't anticipate, like a Kobe earthquake or a Russian default. LTCM had made an especially big bet that higher-interest corporate bonds would gain in relative price while U.S. Treasury paper would decline in relative price, thus bringing the comparative yield on the two into a more familiar alignment. Instead, the corporate bonds went down and the Treasuries went up, and LTCM lost on both sides of its bets. The smartest mathematician cannot write an equation for "event risk."
LTCM admittedly lost $2 billion in August and probably as much again in the first three weeks of September. Its partners, not quite humbled (for they kept saying that if they had more time their strategies would work), scavenged the world for cash. Most of the LTCM positions were financed with repurchase agreements—the fund would sell the bank the instrument in question, under contract to buy it back at a slightly higher price in a specified number of days. As the value of the securities held under repurchase agreements declined, the lenders asked LTCM for more collateral, and soon the cupboard was bare. A $500-million loan from a group of multinational banks came due on Sept. 23. LTCM didn't have the money.
On the last day, the fund was kept alive by a consortium of a dozen big banks and investment houses, brought together for this purpose by the Federal Reserve Bank of New York. The bankers, presidents and CEOs—from Merrill Lynch and Chase, Morgan and Travelers Group, Goldman Sachs and their ilk—were surprised to see one another. LTCM was into each of them for considerably more than a billion dollars, but none of them knew the fund's relations with the others.
William McDonough, president of the Federal Reserve Bank of New York, testified before the House Banking Committee on Thursday that LTCM would have had to be liquidated if he had not organized the rescue. Its assets were still greater than its liabilities, he thought, but any effort to sell them to raise cash would have knocked down the price so far that the creditors would have suffered losses. Federal Reserve Chairman Alan Greenspan testified that the carnage in the markets from this "fire sale" might have ended prosperity in our time.
The argument does not convince, for the banks and investment houses that held LTCM's portfolio under repurchase agreements would not have had to dump it on the market when they foreclosed and took possession. They didn't need the cash, and if they did, they could sell the bonds through a repurchase contract themselves, without affecting the prices. LTCM had a lot of contracts with its saviours for derivatives (a derivative is a financial instrument that "derives" its value from changes in the price of another instrument; in other words, a bet on price movements in the market). But these derivatives could not have busted the system, because each such contract was between LTCM and a single "counterparty" bank. No doubt the failure of the world's largest hedge fund to live up to its obligations under derivative contacts would have slowed significantly the growth of play in this casino, but most thoughtful observers think this would be a super idea, anyway.
The trouble was that the Fed did not know what the losses were and who would bear them. LTCM had relations with separate departments of different banks all over the world. The information system at the banks themselves is not good enough so that a CEO—or a bank examiner—can push a button and see all of his own bank's exposures.
When the Bank for International Settlements did a study of foreign exchange trading two years ago, it found that these trades typically did not settle for two or three days, and during that time most multinational banks were exposed beyond their total capital to the execution of these contracts. Among the factors that made the S&L resolution so difficult in this country was the lack of legal documentation of the assets at even the largest thrifts. Financing through repurchase agreements is an activity conducted among friends, where costs and corners get cut. And by encouraging banks to to settle derivative contracts with each other behind closed doors, rather than at a clearing house, the Fed for all its talk of "transparency" has made the fastest-growing area of banking totally opaque, even to the supervisors themselves.
Looking at the LTCM disaster—the amazingly imprudent total of the loans that the banks had made to this fund and the total absence of warning from the bank examiners now stationed permanently at all the big banks—the Fed must have suspected that the banks did not have good information about their own exposures and might not have the collateral for the repurchases in usable form. This sloppiness is decorously called "operation risk," and it is all over the landscape. Letting LTCM fail under those circumstances, with weeks or even months of confusion, would indeed have provoked worldwide insecurity, and market reactions even more severe than those we experienced.
Virtually everything in the LTCM portfolio that has been made public knowledge has been severely punished in the last few days—it would not be surprising if the entire $3.5 billion has been lost. It doesn't matter in the short term, because the banking supervisors will let the institutions involved carry their LCTM investment at full cost, whatever its value. And the bets on "convergence" may pay out in the three-year time horizon of the bailout, though the change in risk perception that made the computer models worthless is human, not mathematical.
But the paper that has not been sold will continue to overhang and depress a market that knows that some day it must be sold. Certainly, the Japanese experience is that nobody but the insiders benefits when the government organizes "convoys" of big banks to protect the appearance of value in ill-judged financial portfolios. We used to think we knew better.