April 10, 2003

Notes: LTCM

A short summary of what went wrong at LTCM.

Posted by DeLong at April 10, 2003 09:35 PM | TrackBack

Comments

It's always seemed to me that one way of understanding the LTCM debacle is that they hit the 'house limit'. After all, if there's no house limit, there -is- a winning strategy for a gambler: just doubling your losing bet until you win. This is the version of the gambler's fallacy that isn't a fallacy-- probability -does- force things to happen if you can make an infinite number of bets.

Posted by: Matt on April 11, 2003 05:05 AM

"One hand, one million, no tears"

Posted by: Eric on April 11, 2003 05:50 AM

When Genius Failed: The Rise and Fall of Long- Term Capital Management, by Roger Lowenstein [ISBN: 0375758259] is the definitive source here.

LTCM unknowingly amassed a book of bets that were highly correlated, illiquid, and too big to unwind.

Three rookie trading errors commited by the pros.

Posted by: George Zachar on April 11, 2003 06:13 AM

LTCM's positions eventually recovered. But in the real world, you can't wait forever for that to happen when billions upon billions of dollars of other people's money are on the line. The i-banks that stepped in to rescue the fund knew what they were doing: They could provide liquidity and discipline and thus reap the rewards of LTCM's insights over the long run.

Posted by: sd on April 11, 2003 09:13 AM

I agree with Matt:

Consider a heavily left-skewed return distribution. Most often you win some, when you lose, losses are heavvy, but this seldomly happens.

Even if your avarage returns are zero, the average return from a finite sample is most likely to be positive.

You do well at first which makes people bet their money on you. When a return on the far left of the distribution finally hits you, you have quite a lot of money at the table.

If I was asked to shortly summarize LTCM that is. Plus that a left-skewness avert market, according to some recent literature, pays you quite good to hold left-skewness, there is a negative premium on non-diversifiable market coskewness.

Posted by: Mats on April 11, 2003 12:15 PM

On luck and success in the financial trading world (including more along the lines of Mats' and Matt's comments), I highly recommend Naseem Talib's "Fooled by Randomness". You can find a sampler in an article on Talib by Malcom "Tipping Point" Gladwell at http://www.gladwell.com/2002/2002_04_29_a_blowingup.htm.

Posted by: Tom Slee on April 11, 2003 05:08 PM

' When Genius Failed: The Rise and Fall of Long- Term Capital Management, by Roger Lowenstein [ISBN: 0375758259] is the definitive source here ....'

It's a good book, but I think that the book written by the editor of Risk magazine, Nicholas Dunbar (Inventing Money:the story of Long-Term Capital Management and the legends behind it) gives more insight because the author links the advances made in quantitative finance theory to the rise and fall of John Meriwether and his team. LTCM's use of a 'Martingale' strategy, which resembles gambler's fallacy, is explained very well in this book.

Posted by: Nescio on April 12, 2003 10:04 AM

Short and inaccurate. This is the Roger Lowenstein view of the world under which LTCM was following a sound strategy by was just tripped up by failing to consider liquidity issues. In fact, by the end, the majority of the profits were being made in two strategies which can't be characterised in this manner:

1) Straightforward "merger arbitrage"; placing bets that takeover transactions would go through.

2) "Selling volatility". LTCM was a massive writer of long-dated options, in the belief that equity volatility would fall and it rose.

Also this paragraph:

>>Had they been forced to mark their portfolios to market—as Long-Term Capital was—most U.S. banks would have been insolvent in the late 1980s and early 1990s. Most investment banks were illiquid in October 1987 when the market collapsed because they had made huge bridge loans to finance risky takeovers in companies whose values were evaporating.<<

is inaccurate in every important regard as far as I can see. For the first part, LTCM was not "forced to mark its portfolio to market". It wasn't regulated at all. It had collateral arrangements under which each individual position was marked to market, but there was no effective system for centralised aggregation of the risks on a daily basis; this was a lot of the whole problem.

For the second part, the statements made about the banking system in the 1980s and 1990s are also not true. The banks were, for the most part, forced to make massive adjustments to the real estate and problem country debt portfolios and they weren't insolvent. This particular urban myth comes from a congressional inquiry of 1991 (I think) when some loudmouth kept asserting that Citicorp was "economically insolvent", but saying it doesn't make it so.

Posted by: dsquared on April 14, 2003 11:01 PM
Post a comment