The Low Volatility Anomaly: Why Low-Volatility Stocks Outperform

A central assumption in modern finance is that higher risk should be compensated with higher return. The Capital Asset Pricing Model (CAPM) formalizes this idea, predicting that stocks with higher volatility should, on average, deliver higher expected returns.

However, empirical research over the past several decades has documented a pattern that appears inconsistent with this prediction. Stocks with lower volatility often generate higher risk-adjusted returns than their more volatile counterparts. This phenomenon is commonly referred to as the low-volatility anomaly.

In this study, we examine whether this relationship appears in a contemporary dataset of U.S. equities.

The Belief

Standard asset-pricing theory predicts a positive relationship between risk and expected return. Within an equity universe, stocks exhibiting higher volatility should therefore compensate investors with higher returns.

Under this framework, portfolios constructed from high-volatility stocks would be expected to outperform portfolios composed of low-volatility stocks.

A growing body of empirical research suggests that the opposite pattern may occur in practice.

What We Tested

To evaluate this relationship, we measured 20-day realized volatility for each stock in our universe of U.S. small- and mid-capitalization equities.

Stocks were sorted into five volatility quintiles, ranging from the highest-volatility group (Q1) to the lowest-volatility group (Q5). Forward returns were then measured over the subsequent 20 trading days.

The dataset spans more than 25 years of daily data across several thousand U.S. equities.

The analysis therefore reflects a simple and transparent test:

• Signal: 20-day realized volatility
• Sorting method: cross-sectional quintiles
• Forward horizon: 20 trading days

What the Data Shows

The results reveal a clear and monotonic relationship between volatility and subsequent returns.

The lowest-volatility quintile (Q5) produced an annualized return of +21.5%, while the highest-volatility quintile (Q1) generated −2.5% over the same period.

This corresponds to a spread of approximately 24 percentage points per year between the two portfolios.

Importantly, the relationship is monotonic across quintiles. Each step from higher to lower volatility is associated with higher subsequent returns.

Rather than a linear risk-return trade-off, the data suggests an inverse relationship within this universe.

Robustness Checks

One common critique of the low-volatility anomaly is that it may be driven by illiquid securities or penny stocks, which can distort empirical results.

An influential study published by AQR in 2014 argued that much of the anomaly could be explained by these effects.

To address this critique, we applied several robustness filters to the dataset:

• Excluding stocks with prices below $5
• Restricting the universe to the most liquid tercile
• Evaluating the signal across different time periods

Across these tests, the signal remained statistically significant.

When excluding penny stocks, the effect became stronger, not weaker. When restricting the universe to the most liquid tercile, the signal remained statistically significant with a t-statistic above 12.

These results suggest that the observed relationship is not solely driven by illiquid securities.

Signal Stability Over Time

We also evaluated whether the signal weakens in more recent years.

Many well-known anomalies tend to decay once they become widely documented and adopted by market participants.

To assess this, the sample was divided into five time periods. The most recent window (2020–2025) produced the strongest signal observed in the dataset, with an information coefficient of 0.087.

This suggests that the low-volatility effect has not weakened in recent years within this universe.

Possible Explanations

Several behavioral explanations have been proposed for the persistence of the low-volatility anomaly.

One hypothesis is that investors exhibit a preference for high-volatility "lottery-like" stocks, which offer a small probability of extremely large returns.

If these stocks become systematically overvalued due to investor demand, their future returns may be lower than predicted by traditional asset-pricing models.

At the same time, lower-volatility stocks may receive less investor attention, potentially leading to persistent pricing inefficiencies.

Implication

The results suggest that, within small- and mid-capitalization equities, the relationship between volatility and future returns may differ from the prediction of standard asset-pricing theory.

Rather than receiving compensation for volatility risk, investors may experience lower returns when holding the most volatile stocks.

These findings are consistent with a growing body of literature documenting the low-volatility anomaly.

"This article is provided for educational and research purposes only and does not constitute investment advice."