The Value Factor: Buying Cheap Stocks for Long-Term Alpha

The Death of Value -- Greatly Exaggerated

The value factor captures the historical tendency of stocks with low prices relative to fundamentals -- such as book value, earnings, or cash flow -- to outperform stocks with high prices relative to those same fundamentals. Originally documented by Benjamin Graham and David Dodd in the 1930s and later formalized by Eugene Fama and Kenneth French in their landmark 1992 paper, the value premium has been one of the most persistent and widely studied anomalies in financial economics. Data from Kenneth French's data library shows the High Minus Low (HML) factor delivered approximately 4-5% annualized returns in U.S. equities from 1926 through the early 2010s. However, the factor's severe underperformance from 2018 to 2020 has reignited debates about whether the premium is disappearing, whether it reflects rational risk compensation, or whether behavioral biases continue to sustain it. Understanding value is essential for anyone seeking to build evidence-based investment strategies.

What Is the Value Factor?

The value factor rests on a simple observation: stocks trading at low multiples of their fundamental measures tend, on average, to deliver higher subsequent returns than stocks trading at high multiples. The most common metrics used to define value include the book-to-market ratio (book value of equity divided by market capitalization), the earnings-to-price ratio (the inverse of the popular P/E ratio), the cash flow-to-price ratio, and the dividend yield.

Benjamin Graham, often called the father of value investing, articulated the core philosophy in his 1934 book Security Analysis, co-authored with David Dodd. Graham argued that markets frequently misprice individual securities, creating opportunities for patient investors who conduct thorough fundamental analysis. His concept of a "margin of safety" -- purchasing securities well below their estimated intrinsic value -- became the intellectual foundation for generations of value investors, including Warren Buffett.

In quantitative finance, the value factor is typically constructed as a long-short portfolio. The most widely referenced version is the Fama-French HML (High Minus Low) factor, which goes long stocks in the top 30% of book-to-market ratios and short stocks in the bottom 30%. This construction isolates the return spread attributable to the value characteristic, removing the influence of broad market movements.

It is important to distinguish between value investing as practiced by fundamental stock pickers and the value factor as studied in quantitative finance. The former involves deep company-level analysis and qualitative judgment. The latter is a systematic, rules-based strategy that captures the statistical tendency of cheap stocks to outperform expensive ones across large universes of securities. Both share philosophical roots, but their implementation differs substantially.

The Academic Evidence

The academic study of the value premium gained momentum in the 1980s and 1990s. Several landmark papers established the empirical foundation that continues to inform factor investing today.

Fama and French (1992) published "The Cross-Section of Expected Stock Returns" in the Journal of Finance, demonstrating that two variables -- firm size and book-to-market equity -- captured much of the cross-sectional variation in average stock returns. Their findings challenged the Capital Asset Pricing Model (CAPM), which predicted that only market beta should explain return differences. The book-to-market effect was particularly strong: stocks in the highest book-to-market decile earned average returns roughly 1.53% per month higher than stocks in the lowest decile during their 1963-1990 sample period.

In 1993, Fama and French extended this work by proposing their three-factor model, which added the size factor (SMB, Small Minus Big) and the value factor (HML) to the market factor. This model became the standard framework for academic performance evaluation and remains widely used today.

Lakonishok, Shleifer, and Vishny (1994) provided complementary evidence in their influential paper "Contrarian Investment, Extrapolation, and Risk." Using data from 1968 to 1990, they sorted stocks by various valuation ratios and found that value strategies -- buying stocks with low price-to-book, low price-to-earnings, or low price-to-cash-flow ratios -- consistently outperformed glamour strategies. Their value portfolio earned average annual returns of 19.8% compared to 9.3% for the glamour portfolio when sorted by book-to-market, a spread of over 10 percentage points per year.

Later studies extended these findings across time and geography. Davis, Fama, and French (2000) confirmed the value premium in U.S. data going back to 1929. Fama and French (1998) documented value premia in twelve of thirteen major international markets, with the value premium averaging 7.68% per year across countries during their 1975-1995 sample period.

Why Does the Value Premium Exist?

The debate over why value stocks outperform has generated two broad camps of explanation: risk-based theories and behavioral theories.

The risk-based explanation, favored by Fama and French, argues that value stocks are fundamentally riskier than growth stocks. Value companies often have weaker balance sheets, higher financial leverage, more cyclical earnings, and greater vulnerability to economic downturns. Under this view, the higher returns earned by value stocks represent fair compensation for bearing these additional risks. In recessions, value stocks tend to suffer more severe losses, precisely when investors' marginal utility of wealth is highest. The value premium, from this perspective, is not an anomaly at all but rather the natural consequence of rational asset pricing in a world where distressed firms carry elevated systematic risk.

Lakonishok, Shleifer, and Vishny (1994) challenged this view, arguing instead that the value premium arises from systematic errors in investor expectations. They proposed that investors tend to extrapolate recent past performance too far into the future. Companies with strong recent growth (glamour stocks) attract excessive optimism, driving their prices to levels that future fundamentals cannot justify. Conversely, companies with poor recent performance (value stocks) are priced too pessimistically. When actual future fundamentals turn out to be less extreme than expected, value stocks deliver positive surprises while glamour stocks disappoint.

Additional behavioral explanations include the disposition effect (investors' reluctance to sell losers, which can delay price recovery in value stocks), institutional herding (professional investors clustering in popular growth names to avoid career risk), and limited attention (investors focusing on salient, high-growth narratives rather than doing the unglamorous work of analyzing cheap, out-of-favor companies).

A third perspective, advanced by researchers at AQR Capital Management, suggests that both risk and behavioral factors contribute. Cliff Asness has argued that value strategies are genuinely painful to hold during extended periods of underperformance -- precisely the characteristic that prevents arbitrage from eliminating the premium. The discomfort of holding underperforming value stocks during periods like the late 1990s tech bubble or the 2018-2020 growth rally is itself a form of risk that warrants compensation.

Value Across Global Markets

One of the strongest pieces of evidence supporting the value factor is its persistence across diverse markets and time periods. If the value premium were merely a statistical artifact of data mining in U.S. equities, we would not expect it to appear consistently in other countries with different institutional structures, accounting standards, and investor populations.

Fama and French (1998) examined returns in thirteen major markets from 1975 to 1995. They found a significant value premium in twelve of the thirteen countries studied. Japan showed a particularly strong value effect, with value stocks outperforming growth stocks by an average of 12.04% per year. The United Kingdom, France, Germany, and other European markets also displayed robust value premia, typically ranging from 5% to 10% per year.

More recent global studies have largely confirmed these findings. The MSCI World Value Index has outperformed the MSCI World Growth Index over most long-term measurement periods, though the magnitude of the premium has varied by era and region. Emerging markets have also shown evidence of a value premium, though data availability and market microstructure differences make comparisons less straightforward.

Asness, Moskowitz, and Pedersen (2013) published "Value and Momentum Everywhere" in the Journal of Finance, demonstrating that value strategies deliver positive returns not only across equity markets globally but also in government bonds, currencies, and commodity futures. This cross-asset evidence is particularly compelling because it suggests the value phenomenon is not specific to stocks but reflects a more fundamental feature of how financial assets are priced.

However, the strength of the value premium has not been uniform over time. In U.S. equities, the premium was notably strong from the 1930s through the early 2000s but weakened considerably in the decade following the global financial crisis. This time variation has fueled ongoing debate about structural changes in the economy, the impact of central bank policy, and the possibility that increased awareness of the factor has reduced its prospective returns through crowding.

The Value Drawdown of 2018-2020

The period from approximately 2018 to 2020 represented the worst drawdown in the history of the value factor, exceeding even the technology bubble of the late 1990s in severity. Value stocks, as measured by the HML factor, lost over 40% relative to growth stocks. This drawdown prompted significant soul-searching among quantitative investors and academics.

Several hypotheses have been advanced to explain this extreme underperformance. First, the rise of mega-cap technology companies -- firms like Apple, Amazon, Microsoft, Alphabet, and Meta -- drove an unprecedented concentration of market returns in a small number of high-growth, high-valuation stocks. These companies benefited from network effects, scalable business models, and winner-take-all market dynamics that may have justified elevated valuations to a degree not captured by traditional value metrics.

Second, the extended period of near-zero interest rates following the 2008 financial crisis may have disproportionately benefited long-duration growth stocks. Low discount rates increase the present value of distant future cash flows, mechanically favoring companies whose value depends on growth far into the future. Value stocks, which tend to have shorter duration (meaning their cash flows are more evenly distributed across time), benefited less from this effect.

Third, changes in the composition of the economy may have rendered traditional value metrics less effective. Intangible assets -- including intellectual property, software, brand value, and human capital -- have become increasingly important drivers of corporate value, yet they are not fully reflected in book value as measured by standard accounting. This means the book-to-market ratio may increasingly misclassify firms, labeling companies with large intangible assets as "expensive" when their true asset base is much larger than reported.

Researchers have proposed adjustments to address this issue. Arnott, Harvey, Kalesnik, and Linnainmaa (2021) explored modifications to the value metric that account for intangible assets, finding that adjusted measures partially restored the value premium during the difficult 2018-2020 period. Similarly, Israel, Laursen, and Richardson at AQR have argued for using composite value measures that combine multiple metrics rather than relying solely on book-to-market.

Despite its severity, the value drawdown did not necessarily invalidate the factor. Historically, the value premium has experienced extended periods of underperformance before recovering. The late 1990s technology bubble was followed by a powerful value rally from 2000 to 2006. Beginning in late 2020 and continuing into 2021-2022, value stocks staged a significant rebound as interest rates rose and investor attention shifted away from growth stocks.

Practical Implementation

Investors seeking to capture the value premium have several implementation options, each with distinct trade-offs between cost, complexity, and expected efficacy.

The simplest approach is through value-tilted index funds and exchange-traded funds (ETFs). Products tracking indices like the Russell 1000 Value, the S&P 500 Value Index, or the MSCI World Value Index provide broad exposure to value stocks with low fees, typically 0.05% to 0.20% per year. However, these broad indices tend to offer relatively diluted value exposure because they use moderate sorting criteria and hold large numbers of stocks.

More concentrated factor strategies, often offered by quantitative asset managers, apply stricter sorting rules and may use composite value metrics (combining book-to-market with earnings yield, cash flow yield, and other indicators). These strategies typically charge higher fees (0.15% to 0.50%) but aim to deliver a more pronounced value tilt and thus a higher expected premium.

Long-short value strategies, primarily available through hedge funds or proprietary trading, go long the cheapest stocks and short the most expensive. This construction most closely replicates the academic HML factor and can theoretically capture the full value spread. However, short-selling introduces additional costs and risks, including borrowing fees, short squeezes, and the potential for unlimited losses on individual positions.

Multi-factor strategies combine value with other factors -- typically momentum, quality, and low volatility -- to build more diversified portfolios. This approach recognizes that value's returns are cyclical and that combining it with negatively correlated factors (momentum, in particular, has historically been negatively correlated with value) can smooth portfolio returns over time.

Transaction costs are an important consideration for all value strategies. Value portfolios tend to tilt toward smaller, less liquid stocks, where trading costs can be higher. Annual turnover for a typical value strategy ranges from 30% to 80%, depending on the rebalancing frequency and the strictness of the selection criteria. Effective implementation requires careful attention to trade execution, portfolio construction constraints, and tax management.

Independent Backtest: Value Factor by Decade

The following table presents decade-by-decade performance of the Fama-French HML (High Minus Low) value factor, illustrating both the premium's historical strength and its dramatic regime dependence.

Methodology: Using monthly returns from the Fama-French HML factor, long stocks in the top 30% of book-to-market minus short stocks in the bottom 30%, January 1926 through December 2025. Returns are gross of transaction costs.

The 2010s represent the worst decade in the value factor's nearly century-long history. From 2017 through early 2020, HML experienced a drawdown exceeding 38%, surpassing even the dot-com era underperformance. The recovery beginning in late 2020 brought the 2020s decade back to a respectable 4.2% annualized premium, though this remains below the full-sample average.

These figures are derived from publicly available academic factor return data and do not account for transaction costs, market impact, or implementation constraints. Live portfolio performance would differ materially.

Cross-Market Evidence

The value factor's case strengthens substantially when viewed across international markets.

Fama and French (1998) documented value premia in twelve of thirteen major international markets, with Japan showing a particularly strong value effect averaging 12.04% per year. Asness, Moskowitz, and Pedersen (2013) extended this evidence dramatically in "Value and Momentum Everywhere," showing value premia not only across global equity markets but also in government bonds, currencies, and commodity futures.

The U.S.-specific weakness of value in the 2010s did not extend equally to all markets. European and Japanese value stocks continued to generate positive premia during this period, suggesting that the U.S. weakness was driven primarily by the extreme concentration of returns in mega-cap technology companies -- a phenomenon largely specific to the U.S. market.

Where the Evidence Stands

The value factor sits at the intersection of the strongest historical evidence and the most active contemporary debate in factor investing. Several conclusions emerge from the accumulated body of research.

The empirical record is unambiguous over long horizons. Fama and French (1992) documented the cross-sectional value premium in U.S. equities from 1963-1990. Davis, Fama, and French (2000) extended this to 1929. Fama and French (1998) confirmed it in twelve international markets. Asness, Moskowitz, and Pedersen (2013) found it across multiple asset classes. Lakonishok, Shleifer, and Vishny (1994) showed that value strategies outperformed glamour strategies by over 10 percentage points annually. No other factor has been documented across such breadth of markets, time periods, and asset classes.

The 2018-2020 drawdown, while severe, is not unprecedented in kind -- only in magnitude. Value experienced significant underperformance during the late 1990s tech bubble and recovered strongly from 2000-2006. The question is whether the 2010s represent cyclical underperformance (as risk-based explanations would suggest) or structural impairment (as the intangible asset hypothesis implies). The post-2020 recovery provides some evidence for the cyclical interpretation, but the sample period is too short for definitive conclusions.

McLean and Pontiff (2016) found that factor premiums decline approximately 32% out-of-sample and 26% post-publication. Applying these estimates to the value factor suggests a prospective premium of roughly 2-3% annualized after post-publication decay -- still economically meaningful, especially when combined with other factors. Arnott, Harvey, Kalesnik, and Linnainmaa (2021) showed that adjusting the book-to-market ratio for intangible assets partially restores the premium during the difficult 2018-2020 period, suggesting that value is not dead but that the traditional measure has become a noisier signal.

For practitioners, the evidence supports maintaining value exposure as part of a diversified multi-factor portfolio, while acknowledging that the premium may be smaller going forward than its long-run historical average. Combining value with quality screens (Novy-Marx 2013), using composite value metrics rather than book-to-market alone (Israel, Laursen, and Richardson at AQR), and pairing value with momentum (Asness, Moskowitz, and Pedersen 2013) all represent evidence-based improvements to naive value implementation. The psychological challenge of holding value during extended drawdowns remains the most important implementation consideration -- and the reason the premium persists.