Factor Investing Explained: Size, Value, Momentum, Quality

Factors are systematic sources of return that explain why some stocks outperform others. Over six decades of academic research — from the Capital Asset Pricing Model to the Fama-French five-factor model — has identified, tested, and debated these return drivers. This guide traces the full history, explains how each major factor is constructed, and examines the critical question of whether published factors survive post-publication.

What Is a Factor?

In the context of asset pricing, a factor is a characteristic of securities that explains cross-sectional differences in expected returns. If stocks with a particular attribute consistently earn higher (or lower) returns than stocks without that attribute, even after adjusting for risk, that attribute is a candidate factor.

The concept is foundational to modern finance. If markets were perfectly efficient and the only thing that mattered was overall market risk, then all stocks with the same market beta should earn the same expected return. The existence of factors beyond market beta implies either that these factors represent compensation for bearing additional risks (the risk-based explanation) or that they reflect persistent behavioral biases among investors (the behavioral explanation). In practice, most factors likely reflect some combination of both.

Factor investing — the practice of systematically tilting portfolios toward stocks that score well on documented factors — has grown into a massive industry. The rise of “smart beta” ETFs has made factor strategies accessible to individual investors, and today hundreds of billions of dollars are managed in explicit factor strategies worldwide.

The Evolution of Factor Models

CAPM: The Single-Factor World (1964–1965)

The story begins with the Capital Asset Pricing Model (CAPM), developed independently by William Sharpe (1964) and John Lintner (1965). The CAPM proposed a beautifully simple idea: the only factor that should matter for expected returns is a stock’s sensitivity to the overall market — its beta. High-beta stocks should earn higher returns because they are riskier (they fall more in market downturns), and low-beta stocks should earn lower returns because they are safer.

The CAPM equation is: E(Ri) = Rf + βi * (E(Rm) - Rf), where E(Ri) is the expected return of stock i, Rf is the risk-free rate, βi is the stock’s market beta, and E(Rm) - Rf is the market risk premium. Any return above what this equation predicts is “alpha” — return not explained by market risk.

The CAPM was elegant but empirically incomplete. By the late 1970s and 1980s, researchers had documented numerous “anomalies” — patterns in stock returns that the CAPM could not explain. Small stocks outperformed large stocks (Banz, 1981). Value stocks outperformed growth stocks (Stattman, 1980; Rosenberg, Reid & Lanstein, 1985). Stocks with high recent returns continued to outperform (Jegadeesh & Titman, 1993). These anomalies demanded a richer model.

Fama-French Three-Factor Model (1993)

In 1993, Eugene Fama and Kenneth French published “Common Risk Factors in the Returns on Stocks and Bonds” in the Journal of Financial Economics, Vol. 33. This paper introduced the three-factor model, which added two factors beyond the market:

• SMB (Small Minus Big) — The return of a portfolio of small-cap stocks minus the return of a portfolio of large-cap stocks. This captures the size premium: the historical tendency of small stocks to outperform large stocks.
• HML (High Minus Low) — The return of a portfolio of value stocks (high book-to-market ratio) minus the return of a portfolio of growth stocks (low book-to-market ratio). This captures the value premium: the historical tendency of cheap stocks to outperform expensive stocks.

The three-factor model explained most of the cross-sectional anomalies that had been documented up to that point. Portfolios that appeared to generate alpha under the CAPM turned out to have significant loadings on SMB and HML — their excess returns were compensation for size and value risk, not genuine alpha.

Carhart Four-Factor Model (1997)

In 1997, Mark Carhart published “On Persistence in Mutual Fund Performance” in the Journal of Finance, Vol. 52. Carhart extended the Fama-French model by adding a fourth factor:

• UMD / MOM (Up Minus Down / Momentum) — The return of stocks with high recent returns (past 12 months, excluding the most recent month) minus the return of stocks with low recent returns. This captures the momentum premium documented by Jegadeesh and Titman (1993).

The Carhart four-factor model became the standard workhorse model for evaluating mutual fund performance. If a fund’s returns can be fully explained by its exposure to market, size, value, and momentum factors, then the fund is not generating alpha — it is simply harvesting well-known factor premiums, which can be obtained more cheaply through index funds.

Fama-French Five-Factor Model (2015)

In 2015, Fama and French published “A Five-Factor Asset Pricing Model” in the Journal of Financial Economics, Vol. 116. This paper added two new factors to the original three:

• RMW (Robust Minus Weak) — The return of stocks with robust (high) operating profitability minus the return of stocks with weak (low) operating profitability. This captures the profitability premium: profitable firms tend to outperform unprofitable firms.
• CMA (Conservative Minus Aggressive) — The return of stocks with conservative (low) asset growth minus the return of stocks with aggressive (high) asset growth. This captures the investment premium: firms that invest conservatively tend to outperform firms that invest aggressively.

A notable finding of the five-factor model is that HML becomes redundant when RMW and CMA are included. The value factor’s explanatory power is subsumed by the profitability and investment factors. This suggests that what we call “value” may actually be a proxy for profitability and conservative investment — cheap stocks tend to be cheap because they are profitable and don’t invest aggressively, not because of some independent “value” characteristic.

How Factors Are Constructed

Understanding factor construction methodology is essential for anyone implementing or evaluating factor strategies. The Fama-French factors are constructed using a systematic sorting procedure:

The 2x3 Sort Methodology

At the end of each June, all NYSE, AMEX, and NASDAQ stocks are sorted independently on two dimensions. For HML, the sort is on size (market capitalization) and book-to-market ratio (B/M). Size breakpoints use the NYSE median — only NYSE stocks determine the cutoff between small and big, though stocks from all three exchanges are then placed into the appropriate groups. B/M breakpoints use the NYSE 30th and 70th percentiles.

This creates six portfolios: Small Value, Small Neutral, Small Growth, Big Value, Big Neutral, Big Growth. The HML factor is then computed as:

HML = (Small Value + Big Value) / 2 - (Small Growth + Big Growth) / 2

This construction ensures that the HML factor captures the value premium independent of size. The same methodology is applied for RMW (sorting on size and operating profitability) and CMA (sorting on size and investment, measured as asset growth).

SMB is computed as the average return of all small portfolios minus the average return of all big portfolios, averaged across the three sorting variables (B/M, profitability, investment). This ensures that the size factor is not contaminated by value, profitability, or investment effects.

The Major Factors in Detail

Market (MKT)

The market factor is simply the excess return of the broad stock market over the risk-free rate: MKT = Rm - Rf. It captures the equity risk premium — the compensation investors receive for bearing the systematic risk of owning stocks. Historically, the US equity risk premium has averaged approximately 6–8% per year, though estimates vary depending on the time period and methodology.

Size (SMB)

The size premium — the tendency of small-cap stocks to outperform large-cap stocks — was first documented by Rolf Banz in 1981. The economic rationale is that small firms are riskier: they have less diversified revenue streams, less access to capital markets, lower liquidity, and higher bankruptcy risk. Investors demand higher expected returns to compensate for these risks.

However, the size premium has been controversial. It was largely concentrated in micro-cap stocks and in the month of January. Since its publication, the size premium has been much weaker — some researchers argue it has disappeared entirely once you control for quality and profitability. Small stocks that are also profitable continue to outperform, but small, unprofitable stocks (which drag down the average) are terrible investments.

Value (HML)

The value premium — cheap stocks outperforming expensive stocks — is one of the most studied phenomena in finance. Value can be measured by book-to-market ratio (Fama & French’s preferred metric), earnings yield, cash flow yield, or dividend yield. The risk-based explanation is that value stocks are distressed firms facing genuine economic risks; the behavioral explanation is that investors systematically overpay for glamorous growth stocks and underpay for boring value stocks.

Value experienced a severe and prolonged drawdown from approximately 2017 to 2020, leading many to question whether the premium had permanently disappeared. This episode illustrates a fundamental challenge of factor investing: even factors with strong historical evidence can underperform for years, testing the conviction of even the most disciplined investors.

Momentum (UMD / MOM)

Momentum — the tendency of past winners to keep winning and past losers to keep losing over 3 to 12 month horizons — was formally documented by Jegadeesh and Titman in their 1993 Journal of Finance paper. The standard implementation buys stocks in the top decile of past 12-month returns (skipping the most recent month to avoid short-term reversal) and sells stocks in the bottom decile.

Momentum is notable for being the strongest factor in terms of raw returns but also the most dangerous. Momentum strategies are subject to momentum crashes — sudden, severe reversals that typically occur when the market recovers sharply after a downturn. The most famous example occurred in 2009, when the momentum factor lost approximately 40% in three months as beaten-down stocks snapped back violently.

Profitability (RMW)

The profitability factor captures the empirical regularity that firms with higher operating profitability earn higher stock returns. This was theoretically motivated by Robert Novy-Marx (2013), who showed that gross profitability (revenue minus cost of goods sold, scaled by assets) is a strong predictor of cross-sectional returns. Fama and French’s RMW factor uses operating profitability (revenue minus cost of goods sold, minus selling/general/administrative expenses, minus interest expense, all scaled by book equity).

Investment (CMA)

The investment factor captures the finding that firms which invest conservatively (low asset growth) outperform firms that invest aggressively (high asset growth). This is related to the “asset growth anomaly” documented by Cooper, Gulen, and Schill (2008). The economic intuition is that firms with abundant profitable opportunities do not need to invest aggressively — they are already generating strong returns on existing assets. Firms that invest aggressively may be empire-building, overpaying for acquisitions, or investing in low-return projects.

Factor Decay: Do Factors Survive Publication?

One of the most important questions in factor investing is whether factors continue to work after they are published and widely known. The evidence is mixed and sobering.

The Replication Crisis in Finance

In 2016, Campbell Harvey, Yan Liu, and Heqing Zhu published “...and the Cross-Section of Expected Returns” in the Review of Financial Studies. This paper catalogued over 300 factors that had been published in top academic journals and argued that most were likely false positives. The core argument was that the standard statistical threshold (a t-statistic of 2.0, corresponding to a p-value of 0.05) is far too low given the massive number of factors that have been tested. With hundreds of researchers testing thousands of potential factors on the same datasets, many will appear significant by pure chance. Harvey, Liu, and Zhu argued that a t-statistic of at least 3.0 should be required for a new factor to be considered credible.

The Replication Study

In 2020, Kewei Hou, Chen Xue, and Lu Zhang published a comprehensive replication study in the Review of Financial Studies that attempted to replicate 452 published anomalies. Their findings were striking: approximately 65% of the anomalies failed to replicate using the original authors’ methodology with updated data. Many factors that appeared highly significant in the original papers produced t-statistics well below 2.0 in the replication attempt. The failures were concentrated in factors with complex construction, small sample sizes, or reliance on micro-cap stocks.

Post-Publication Decay

McLean and Pontiff (2016) studied 97 factors published in top finance journals and found that, on average, factor returns declined by approximately 32% after publication. They attributed about half of this decline to statistical bias (in-sample overfitting) and about half to arbitrage (investors trading on the published factor, thereby reducing the premium). This means that even factors that are genuinely real will deliver lower returns going forward than their historical backtests suggest.

Practical Factor Investing

Smart Beta ETFs

The most accessible way to implement factor strategies is through smart beta ETFs, which systematically tilt toward specific factors while maintaining broad diversification and low costs. Major providers include iShares, Vanguard, and Dimensional Fund Advisors (DFA). A single-factor ETF targets one factor (e.g., a pure value fund), while multi-factor ETFs combine several factors in one vehicle.

Factor Timing: Possible but Difficult

Factor timing — increasing exposure to factors expected to do well and decreasing exposure to factors expected to do poorly — is one of the most debated topics in quantitative investing. Some evidence suggests that factors are more attractive when they are cheap relative to their own history (value spreads are wide), or when macroeconomic conditions favor them. However, implementing factor timing in practice is extremely difficult, and most academic evidence suggests that the benefits are modest after transaction costs.

Multi-Factor Portfolios

Because different factors have low or negative correlations with each other (value and momentum, for example, are negatively correlated), combining multiple factors in a single portfolio can improve risk-adjusted returns significantly. A portfolio that holds stocks ranking highly on value, momentum, and quality simultaneously will have a much more stable return stream than a portfolio targeting any single factor.

There are two main approaches to multi-factor construction: the portfolio mixing approach (hold separate single-factor portfolios and combine them) and the integrated approach (select stocks that score well on multiple factors simultaneously). The integrated approach is generally preferred because it avoids holding stocks that score well on one factor but poorly on another, and it reduces turnover and transaction costs.

The Factor Zoo: Separating Signal from Noise

With over 400 published factors and counting, the field faces what some researchers have called the “factor zoo” problem. How do we determine which factors are genuinely predictive and which are artifacts of data mining?

A reasonable consensus has emerged around a relatively small number of robust factors: market, size (conditional on quality), value, momentum, profitability/quality, and low volatility/low beta. These factors have been documented across multiple countries, multiple time periods, multiple asset classes, and by multiple independent research teams. They also have plausible economic explanations — either as compensation for risk or as a reflection of persistent behavioral biases.

Factors that are concentrated in micro-cap stocks, that work only in specific time periods, or that lack any economic rationale should be treated with extreme skepticism. The bar for a new factor should be high: it should work out-of-sample, survive transaction costs, have an economic explanation, and ideally appear in international data.

References

1. Banz, R. W. (1981). “The Relationship Between Return and Market Value of Common Stocks.” Journal of Financial Economics, 9(1), 3–18.
2. Carhart, M. M. (1997). “On Persistence in Mutual Fund Performance.” The Journal of Finance, 52(1), 57–82.
3. Cooper, M. J., Gulen, H. & Schill, M. J. (2008). “Asset Growth and the Cross-Section of Stock Returns.” The Journal of Finance, 63(4), 1609–1651.
4. Fama, E. F. & French, K. R. (1993). “Common Risk Factors in the Returns on Stocks and Bonds.” Journal of Financial Economics, 33(1), 3–56.
5. Fama, E. F. & French, K. R. (2015). “A Five-Factor Asset Pricing Model.” Journal of Financial Economics, 116(1), 1–22.
6. Harvey, C. R., Liu, Y. & Zhu, H. (2016). “...and the Cross-Section of Expected Returns.” The Review of Financial Studies, 29(1), 5–68.
7. Hou, K., Xue, C. & Zhang, L. (2020). “Replicating Anomalies.” The Review of Financial Studies, 33(5), 2019–2133.
8. Jegadeesh, N. & Titman, S. (1993). “Returns to Buying Winners and Selling Losers: Implications for Stock Market Efficiency.” The Journal of Finance, 48(1), 65–91.
9. Lintner, J. (1965). “The Valuation of Risk Assets and the Selection of Risky Investments in Stock Portfolios and Capital Budgets.” The Review of Economics and Statistics, 47(1), 13–37.
10. McLean, R. D. & Pontiff, J. (2016). “Does Academic Research Destroy Stock Return Predictability?” The Journal of Finance, 71(1), 5–32.
11. Novy-Marx, R. (2013). “The Other Side of Value: The Gross Profitability Premium.” Journal of Financial Economics, 108(1), 1–28.
12. Sharpe, W. F. (1964). “Capital Asset Prices: A Theory of Market Equilibrium under Conditions of Risk.” The Journal of Finance, 19(3), 425–442.