Ludwig B. Chincarini has written a compelling book. The thesis of The Crisis of Crowding: Quant Copycats, Ugly Models, and the New Crash Normal (Bloomberg/Wiley, 2012) may not be a paradigm shift, but the detail with which it is documented makes for fascinating reading.
In the beginning was Long-Term Capital Management and the 1998 crisis. As Chincarini contends, “the 2008 financial crisis really began 10 years earlier.” LTCM had a portfolio of trades, mainly relative value and convergence trades in the fixed-income swap and bond markets of OECD countries, that seemed to be statistically and economically uncorrelated. LTCM had, however, also ventured into “less lovely territory,” including directional bets that made up about 20% of the portfolio. They were short the U.S. swap spread, a position that carried with it an unlimited downside risk in the event of a world crisis when traders would flee to quality and the spread would widen. They were also short long-term volatility.
“LTCM knew these trades weren’t perfect, but may have kept them on in the interest of diversification. But crowds were developing, and that was affecting trades. With every call to a dealer, with every cocktail party between dealers, and with every quant trying desperately to reverse engineer LTCM’s success, word traveled and more investors copied LTCM’s moves. The beautiful trades were getting ugly, with smaller expected profits and a new danger: the danger that copycats might rush for the exit at inopportune moments and cause dramatic changes among trade correlations. That’s just what happened in August and September 1998, when Russia defaulted on its debt, the copycats ran for the exits, and the Titanic of hedge funds sank into the chilling water.” (pp. 68-69)
Chincarini claims that ignoring the broad lessons from LTCM’s failure was, “in part, the cause of a much greater systemic financial collapse in 2008 and 2009.” (p. 101) He thinks that there were twelve lessons to be learned, among them: “1. Understanding saturation and interconnectedness is important for measuring risk and liquidity. 2. There are limitations to any risk management system. Risk measures should include some aspect of valuation. 3. Appropriate leverage depends on underlying volatility, but any amount of leverage with any amount of underlying risk can lead to bankruptcy.” (p. 119)
Before the Bear Stearns collapse, however, came the quant crisis in 2007. It lasted only about ten days but did irreparable damage to many quantitative equity funds. Theories abound as to what (or who) caused the crisis. The author argues that “a shocked, crowded space caused the quant crisis. As quantitative managers began closing positions, they put pressure on other managers to close positions or face margin calls. That pushed prices even lower.” (p. 137)
We know only too well what followed in 2008, at least in its broad strokes. Chincarini, who conducted interviews with many of the players in the various financial crises, again fills in some important details and offers lessons for the future.
Oh, yes, and then there was the flash crash. The author maintains that the Waddell-Reed trade was simply too small to have caused the flash crash. It’s possible that “HFT activity caused data overload in some of the NYSE Arca’s out-of-date systems. Jittery markets combined with a flood of orders and old computers to create a computer glitch at the NYSE Arca. This directly caused the Flash Crash. Chaos ruled those three minutes because faulty data scared off liquidity providers. Amazingly enough, the liquidity providers all ran for the exits at the same time. Crowd behavior erased liquidity just when traders needed it the most, and just as the market saw in the LTCM crisis, the Quant Crisis, and the subprime crisis.” (p. 321)
Chincarini’s analysis should be mandatory reading for anyone who manages money, trades in size or designs those trades, or thinks about financial regulation. And, by the way, even for the rest of us it is a darned good read.