John Meriweather started LTCM, a
hedge fund, in 1993, upon being 'asked to leave'
his position as the head of
fixed income at Salomon Brothers in 1991 due to Salomon's
treasury auction rigging scandal.
Fixed Income in the 80's and Solly in particular was one of the most
prestigious banks involved in fixed income (and trading in general), so despite the scandal Meriweather preserved a great deal of credibility on the street.
Meriweather then teamed up with Nobel Prize Winners Robert Merton (from Black-Scholes-Merton theory, not to be confused with Morton from Heath-Jarrow-Morton model) and Myron Scholes (from Black-Scholes-Merton).
Banking on their enormous reputation LTCM was able to become higly leveraged, and put on interest rate swaps for no initial margin over the market rate. In this way they could take advantage of betting on small spreads in the fixed income market (like between mortgage-backed and highly rated corporate bonds), using their elaborate, technically sound models to detect the arbitrage. Using its ability to write swaps, they could also take advantage of liquidity price differences. In its first two years it returned 43% and 41% respectively, and finished with 7 billion in capital.
Now that they were making all of this profit, the huge investment banks wanted to buy in as well. UBS, for example, bought in for nearly $1 Billion USD.
Mistake number 1: In order to capitalize on market inefficiencies you must never become so big your actions become the market. LTCM did not reduce its positions relative to the capital reduction, so the leverage did not decrease.
Mistake number 2: The probability that a given equity will reach a certain price is 100%, assuming that it is a random walk. However, in real life, not in models, if it goes under a certain amount, and you are leveraged enough, your lenders are going to demand more on margin. If you don't have that much cash available, you're fucked. You have to sell out more assets to cover the margin.
Mistake number 3: Russian bonds. Here, the issue is risk inherent in your model. LTCM had taken a significant position on Russian bonds, because their models told them that they were mispriced. However, it did not take into account liquidity risk at all (i.e, Who is going to trade with you? To make a transaction in the financial markets, you need need a buyer and a seller).
So, in 1998 Russia melts down. LTCM had taken positions betting that the spread between US governments and Russian bonds would converge; obviously when Russia declares a moratorium on debt, investors jump out of Russia and invest in the stable US, increasing the US-Russia spread. LTCM starts losing money; then LTCM's counterparties (Mistake #2) start calling for more on margin.
Meriweather still believed that this was temporary. He focused on that random walk-based model and thought that all they needed was some capital to ride them through. After all, probability theory says that they will make it all back (It just doesn't say when).
Meriweather then sent out an urgent call for more capital, from such investors like Soros, Warren Buffet, Jon Corzine (Goldman Sachs), but the money was not coming. One reason might have been from Mistake #1; these huge institutional investors could have been making a lot of money trading against LTCM's portfolio. LTCM lost 2.1 Billion USD that August. Now nobody wanted to lend to LTCM to cover these margins, as the risk was too great; if LTCM collapsed the outcome could be disastrous.
If LTCM collapsed, it would then cause a mass close-out in all of the OTC derivatives markets, investment banks pulling out of LTCM would have to rebalance Basel-required hedges, and the market would move against them, shooting market prices for all derivatives, fixed income, and equities down. This would cause a panic that might cripple the world's markets.
The Federal Reserve then caught wind of the risk and rushed to do something: Alan Greenspan as Chairman and Peter Fischer from the NY Fed gathered bankers in an emergency meeting from Bankers Trust, Barclays, Bear Stearns, Chase, Credit Suisse First Boston, Deutsche Bank, Goldman Sachs, J.P. Morgan, Lehman Brothers, Morgan Stanley, Credit Agricole, Banque Paribas, UBS Warburg, Salomon Smith Barney, and Societe Generale to contribute nearly 300m each in order to bail out LTCM in a window of about 2 days; two days to save the entire global financial system.
If the financial system failed worldwide, then governments would fall, businesses would collapse, people would have lost their jobs, their savings.
After this debacle people began investigating how one institution could cause so much havoc; which is somewhat addressed in the Basel II Financial Accords.
Source: Lessons From The Collapse Of Hedge Fund, Long-Term Capital Management (by David Shirreff, a case we studied)