Mean Reversion
Mean reversion is the financial theory and empirical observation that asset prices, returns, volatility, or other financial variables that have deviated significantly from their historical long-run average tend to return toward that average over time. Trading strategies based on mean reversion profit by buying assets that have fallen significantly below their historical norms and shorting assets that have risen significantly above them.
Key takeaways
- Mean reversion is most reliably documented in financial ratios (P/E, P/B), interest rates, volatility (the VIX tends to revert to its long-run mean), and currency valuations (purchasing power parity).
- Mean reversion in individual stock prices is less reliable and can be overwhelmed by momentum effects over shorter horizons.
- Statistical tests for mean reversion include the augmented Dickey-Fuller test, the Hurst exponent (H < 0.5 indicates mean reversion), and the variance ratio test.
- Pairs trading and statistical arbitrage are the most common implementations of mean-reversion strategies, exploiting temporary divergences between historically correlated securities.
- Mean-reversion strategies are vulnerable to regime changes and structural breaks — when fundamentals change permanently, what appears to be a mean-reverting deviation may instead be a permanent repricing.
Explanation
Mean reversion is one of the two fundamental market dynamics — the other being momentum — and understanding when each dominates is central to systematic trading and investment strategy design. The intuition behind mean reversion is deeply connected to economic theory: if prices deviate significantly from fundamental value, arbitrageurs and value-oriented investors should be attracted to the mispricing, buying what is cheap and selling what is expensive until prices return to fair value.
The empirical evidence for mean reversion varies dramatically across time horizons and asset classes. Over very short intraday horizons (seconds to minutes), market microstructure effects create mean reversion in bid-ask bounces and temporary order imbalances. Over medium-term horizons (days to weeks), mean reversion is observed in relative value relationships between correlated securities — the foundation of statistical arbitrage. Over long-term horizons (years to decades), mean reversion is evident in equity valuation ratios (the CAPE ratio mean-reverts), interest rates (influenced by central bank policy around neutral rates), and currency values (purchasing power parity as a long-run anchor).
The Ornstein-Uhlenbeck (OU) process is the canonical continuous-time model for mean-reverting dynamics. It specifies that the rate of change in a variable is proportional to its distance from the long-run mean, with a stochastic noise component. The speed of mean reversion (the mean-reversion coefficient κ) determines how quickly the variable reverts — high κ implies fast reversion, low κ implies slow. The half-life of a mean-reverting process — the expected time for half of a deviation to be eliminated — is calculated as ln(2)/κ and is a critical parameter for strategy design.
For hedge funds, mean-reversion strategies face two primary risks. The first is 'value trap' risk: an asset may appear cheap by historical norms because the underlying fundamentals have permanently deteriorated (think Kodak or Blockbuster). The second is timing and funding risk: even if the mean reversion is genuine, it may take longer to materialize than the fund's liquidity terms or leverage tolerance allows, forcing liquidation at precisely the wrong time.
Formula
dX_t = κ(μ − X_t)dt + σ dW_t; Half-life = ln(2)/κ
Example
A statistical arbitrage fund identifies that the spread between two historically correlated airline stocks (Airline A and Airline B) has widened by 3.2 standard deviations from its historical 60-day mean. The fund buys Airline A (the relatively cheap one) and shorts Airline B (the relatively expensive one) in equal dollar amounts. Based on the estimated Ornstein-Uhlenbeck half-life of 8 trading days, the fund expects the spread to return to its mean within 2–4 weeks. The spread does revert, generating a profit of approximately 0.8% of capital deployed on the round trip, consistent with the strategy's typical per-trade economics.
Related terms
Arbitrage Central Bank Equity Leverage Liquidity Lock Up Period Macro Fund Merger Arbitrage Portable Alpha Purchasing Power Parity Redemption Gate Relative Value