Mean Reversion and Statistical Arbitrage › Mean Reversion and Statistical Arbitrage

Mean reversion is one of the most intuitive and enduring concepts in quantitative finance. The idea is simple: when prices or spreads deviate significantly from their historical norms, they tend to drift back toward those norms over time. This behavior creates opportunities for traders who can systematically identify and exploit these temporary dislocations.

Unlike trend following strategies that profit from persistent price movements, mean reversion strategies bet against extremes. When a spread between two related securities widens beyond historical bounds, mean reversion traders go long the underperformer and short the outperformer, expecting convergence. When an asset's price falls far below its fair value, they buy expecting recovery.

Statistical arbitrage, often called "stat arb," is the large-scale application of mean reversion principles. Rather than betting on individual pairs, stat arb strategies construct diversified portfolios of many small mean-reverting bets, using statistical models to identify mispricings and manage risk. This approach transformed from a niche strategy at a few quantitative hedge funds in the 1980s into a dominant force in equity markets, with stat arb firms accounting for a significant share of daily trading volume.

This chapter develops the theoretical foundations of mean reversion, builds practical implementations of pairs trading and statistical arbitrage, and examines the risks that have caused even the most sophisticated stat arb strategies to experience dramatic drawdowns.