The 2007 Quant Quake

In the first two weeks of August 2007, a wide range of quantitative equity market-neutral funds — run by different firms, using different research, with no obvious common ownership — all lost several percent in the space of a few days, then largely recovered much of it within about a week. The funds hadn't suddenly become bad at stock-picking; the trigger was that one or more large multi-strategy funds needed to rapidly reduce leverage (likely tied to unrelated losses in credit and mortgage markets, the same summer the subprime crisis was unfolding), and started liquidating their quantitative equity books fast, without much regard to price.

Because so many quant funds independently held similar factor tilts — long cheap, high-momentum, high-quality stocks and short their opposites, a crowded style that many managers had converged on through similar research — one fund's forced selling pushed exactly the stocks that other funds were also long, and bid up the stocks other funds were also short. Losses cascaded: each fund's own risk models then triggered further deleveraging, amplifying the very price moves that caused the initial losses, in a self-reinforcing spiral entirely disconnected from company fundamentals.

The episode is the textbook case study for crowding risk: strategies don't need shared ownership or shared research to become correlated in a crisis, they only need overlapping factor exposures, and a large enough forced seller can synchronize supposedly independent funds' returns purely through the mechanics of everyone holding similar positions at the same time.