Third-Party Research & Methodology Only

This section shares summaries of third-party academic research and descriptions of quantitative models. The content represents the findings of the original researchers, not the opinions or recommendations of Foxholm Financial. Foxholm Financial does not publish hypothetical or backtested performance metrics on its quantitative research pages. All content is restricted to methodology, signal construction, factor logic, and risk architecture. SEC rules require that investment advisers not present misleading performance data, and our methodology-only approach reflects that standard and the firm's fiduciary obligations.

Mean Reversion

Trading Concept Statistical Property Strategy Foundation

Mean reversion is the theory that prices, returns, and other financial variables tend to move back toward their long-term average over time. When an asset's price drifts far above or below its historical norm, mean reversion predicts that it will eventually return to that average.

The concept is one of the oldest ideas in financial markets. It stands in direct contrast to momentum (the tendency for recent winners to keep winning and recent losers to keep losing). Where momentum strategies buy strength and sell weakness, mean reversion strategies do the opposite: they buy assets that have fallen and sell assets that have risen, betting on a snapback to normal levels. The tension between these two forces shapes much of quantitative trading.

Definition

Mean reversion describes the statistical tendency of a variable to drift back toward its long-run average after moving away from it. In finance, this applies to asset prices, valuations, interest rates, volatility, and other measurable quantities.

Core Idea

If a price or return is unusually high relative to its historical average, mean reversion suggests a statistical tendency to fall. If it is unusually low, it suggests a statistical tendency to rise.

The "mean" is the long-term average that the variable gravitates toward. The "reversion" is the pull back to that average. The speed and strength of that pull vary by asset, time horizon, and market conditions.

Mean reversion does not guarantee that a price will return to its average. It describes a statistical tendency observed across large samples of data. Individual assets can diverge from their historical average permanently if the underlying fundamentals change. A company that loses its competitive advantage, for example, may never return to its prior valuation levels.

How Mean Reversion Works

The basic mechanism is straightforward. Prices fluctuate around a central value, sometimes overshooting to the upside and sometimes undershooting to the downside. After an overshoot, the price tends to fall back. After an undershoot, it tends to rise. This creates a pattern where extreme moves are followed by moves in the opposite direction.

Several forces drive this behavior. When a stock becomes overvalued relative to its fundamentals, profit-taking by existing holders and reluctance by new buyers create selling pressure. When a stock becomes undervalued, bargain hunters step in and push the price back up. These forces act as a gravitational pull toward the long-run average.

The strength of the mean-reverting pull depends on the asset and time frame. Short-term price movements in individual stocks often show mean reversion at the daily or weekly level, particularly after large moves. Broad market indices tend to show weaker mean reversion over shorter periods but stronger mean reversion over multi-year horizons. Interest rates and implied volatility also exhibit strong mean-reverting behavior.

In practice, mean reversion strategies identify assets that have deviated significantly from their average, then take positions that profit when the price reverts. This forms the basis of pairs trading, statistical arbitrage, and contrarian investing approaches.

Measuring Mean Reversion

Quantitative analysts use several tools to measure whether a time series exhibits mean reversion and how quickly it reverts. Each approach captures a different aspect of the behavior.

Z-scores. A z-score measures how many standard deviations (a measure of how spread out values are from their average) the current value sits from its historical mean. A z-score of +2 means the current price is two standard deviations above the average, suggesting it may be due for a decline. A z-score of −2 suggests the opposite. Mean reversion traders often enter positions when z-scores reach extreme levels, typically beyond +2 or −2.

Half-life of mean reversion. The half-life measures how long it takes, on average, for a deviation from the mean to shrink by half. A half-life of 10 days means that if a price is $10 above its average today, it will typically be about $5 above the average in 10 days. Shorter half-lives indicate faster reversion and more tradeable opportunities.

Ornstein-Uhlenbeck process. This is a mathematical model that describes how a variable moves randomly but is continuously pulled back toward a central value. It captures three key parameters: the long-term mean, the speed of reversion, and the amount of random noise. Practitioners use it to model interest rates, volatility, and spread relationships between correlated assets.

Augmented Dickey-Fuller (ADF) test. The ADF test is a statistical test that checks whether a time series is "stationary" (meaning it tends to return to a fixed average) or "non-stationary" (meaning it can drift without bound). A statistically significant ADF result provides evidence that a series is mean-reverting. This test is a standard first step in evaluating whether a pairs trade or spread is likely to revert.

Practical Example

Consider a spread between two highly correlated stocks in the same industry. Their price ratio has a long-term average of 1.50. A mean reversion trader monitors the ratio and acts when it deviates significantly from that average.

Observation	Price Ratio	Z-Score	Signal
Week 1	1.48	−0.3	No action (within normal range)
Week 4	1.52	+0.3	No action (within normal range)
Week 8	1.70	+2.5	Enter short position (ratio too high)
Week 12	1.55	+0.6	Ratio reverting toward mean
Week 15	1.50	0.0	Close position (ratio at mean)

In this example, the trader identifies a significant deviation in week 8 and takes a position that profits as the ratio returns to its historical average by week 15. The z-score provides a standardized measure of how far the ratio has moved, making it easier to compare opportunities across different asset pairs.

Mean Reversion vs. Momentum

Mean reversion and momentum are opposing forces in financial markets. They tend to operate on different time horizons and capture different types of investor behavior. Understanding where each dominates is important for building diversified strategies.

Dimension	Mean Reversion	Momentum
Core belief	Prices return to their long-term average	Recent trends tend to continue
Typical time horizon	Days to weeks (short-term), or multi-year (long-term valuations)	3 to 12 months (intermediate-term)
Trading action	Buy losers, sell winners (contrarian)	Buy winners, sell losers (trend-following)
Behavioral driver	Overreaction to news; prices overshoot, then correct	Underreaction to news; prices adjust slowly
Risk profile	Risk of "catching a falling knife" (buying into a permanent decline)	Risk of reversal at trend exhaustion
Common strategies	Pairs trading, statistical arbitrage	Trend following, cross-sectional momentum

Research suggests that momentum tends to dominate at intermediate horizons (3 to 12 months), while mean reversion is stronger at very short horizons (days to weeks) and very long horizons (3 to 5 years). Some quantitative portfolios combine both signals to capture different sources of return across time frames.

Known Limitations

Limitations to Keep in Mind

Structural breaks can invalidate the mean. If the underlying fundamentals of an asset change permanently (a company loses a major customer, an industry faces new regulation), the historical average may no longer be relevant. Betting on reversion to an outdated mean can produce large losses.
Distinguishing mean reversion from random noise is difficult. Many apparent mean-reverting patterns in short data sets are simply random fluctuations. Rigorous statistical testing (such as the Augmented Dickey-Fuller test) is required to distinguish genuine mean reversion from chance.
Timing is uncertain. Even when mean reversion is present, the time it takes for prices to revert can vary widely. A position may be correct in direction but require more time (and capital) than anticipated, creating significant carrying costs.
Transaction costs erode profits. Mean reversion strategies often trade frequently, particularly at the daily or weekly level. High turnover means that transaction costs, bid-ask spreads, and market impact can consume a large share of the theoretical profit.
Leverage amplifies risk. Because individual mean reversion signals tend to be small, many strategies use leverage to scale up returns. This amplifies losses when the expected reversion does not materialize or takes longer than expected.

Academic Origin

The empirical study of mean reversion in stock prices gained prominence in the late 1980s. Poterba and Summers published "Mean Reversion in Stock Prices: Evidence and Implications" in 1988 in the Journal of Financial Economics, presenting evidence that stock returns exhibit negative serial correlation (a tendency for above-average returns to be followed by below-average returns) over multi-year horizons. Their findings suggested that a substantial portion of stock price variation reflects temporary deviations from fundamental value.

In the same year, Fama and French published "Permanent and Temporary Components of Stock Prices" in the Journal of Political Economy, reaching a similar conclusion: stock prices contain a slowly decaying temporary component that produces predictable mean-reverting behavior in returns. These two papers, along with Lo and MacKinlay's "Stock Market Prices Do Not Follow Random Walks" (1988), challenged the prevailing view that stock prices move in a purely random fashion and laid the groundwork for quantitative strategies built on mean reversion.