Book: When Genius Failed
Author: Roger Lowenstein
This is a story about Long-Term Capital Management (LTCM), a Connecticut based hedge fund that made waves with unreal returns for a few years before blowing up spectacularly. As it fell, it almost took down the financial markets with it, forcing the hand of New York Fed to organize a Wall St-funded rescue.
- LTCM relied on convergence trades: When it began, LTCM relied primarily on convergence trades, essentially betting that similar assets that are currently priced differently will sell at closer prices at a later date. Some think of convergence trade of as arbitrage trade, but it is not so, since arbitrage trades are done using same assets/cash flows, not merely similar.
- Equity volatility was one of the asset classes that LTCM bet on heavily: By itself, this is not a game-changer. What made it so are following two factors-
First, equity volatility has no ‘logical’ extremes, as it is a derived asset class. Equity volatility cannot be manufactured, and there are no traditional demand and supply forces to smoothen its fluctuation. Second, LTCM traded so much equity volatility that it constituted a signification portion of the entire market. Thus, if they tried to unload their positions, they would end up moving the market against themselves by virtue of the sale, thus nullifying the point of the sale.
- Rear view mirror: Models of LTCM were based on recurrence of the past. Even bigger problem was the assumption of sufficient liquidity in the markets. When liquidity dries up, prices of all asset classes plummet, especially so in case of OTC derivatives. This factor played an even bigger hand in the 2007 financial crisis.
- In crisis mode, correlation becomes 1: When liquidity gets badly eroded, correlation of unrelated entities get closer to 1. Before LTCM blew up, the Asian financial crisis and the Russian default had made the markets jittery causing a huge run to safety to Treasuries. So much so, that even the on-the-money and off-the-money notes ended up far closer prices than is usual. Even US Corporate bonds, which had little exposure to Asian or Russian economies got punished amidst the flight to safety.
My closing thought: LTCM got successful using their powerful models, and the success gave the portfolio managers the internal permission to lever up, at times 100 to 1, leaving very little room for error. The past success also made it easy for managers to ignore signs coming from places other than pure market data, bringing them closer to disaster even faster.