digest
Short Summary
Main recommendation: Keep a long-term value tilt in a diversified portfolio despite recent underperformance, because the value premium has a strong risk-based and behavioral basis and periods of weak performance are expected. Ben Felix argues that the U.S. value premium has weakened since 1991 but cannot be confidently said to have disappeared, and that value has historically offered both excess returns and diversification benefits.
Key Points
- A value stock is one with a low price relative to a fundamental metric such as book value or earnings, and the value premium is its expected excess return over the market.
- Fama and French's 1992 Three-Factor Model treats value and small-cap stocks as independent systematic risks, which supports the view that the premium is compensation for risk.
- Behavioral explanations, such as investor overoptimism about glamorous firms (from Piotroski and So's 2012 paper), support the premium, and limits to arbitrage (for example, a pension officer avoiding multi-year underperformance) help it persist.
- From July 1991 to June 2019, the U.S. value premium averaged 1.17% per year and was not statistically different from zero; Fama and French's 2020 paper found the monthly data too volatile to conclude the expected premium had changed from the earlier period.
- Value investing can be painful for long stretches, but a 2018 bootstrap study by Fama and French found a 23% chance of a negative premium over 23 years and 9% over ten years.
Details Worth Noting
- Concrete instructions: The speaker gives no step-by-step procedure, but the actionable guidance is to keep an existing long-term value tilt rather than abandon it, since recent periods are within expected variation. For index investors, a value tilt requires adding separate value index funds to the portfolio, which adds complexity. Ben Felix does not condone market or factor timing, but following AQR co-founder Cliff Asness's view, an investor who already holds value could consider modestly increasing value exposure ("sinning a little") when the value spread is historically wide, which was the case as of August 2019 (97th percentile of historical cheapness).
- U.S. value premium history: 4.9% per year (July 1963 to June 1991) and 2.39% per year (July 1926 to June 1963), with the premium weakening afterward.
- Recent underperformance: value trailed the market by 6.88% per year over the three years ending December 2019 and by 2.24% per year over the decade.
- Diversification: during some ten-year periods when the U.S. market underperformed Treasury bills, U.S. value delivered a positive premium.
- International results (July 1991 to June 2019, annualized excess over local market): Japan value +6.14%, U.K. +0.29%, emerging markets +3.02%, Asia Pacific +5.4%, Australia +2.95%, Canada -0.53%.
Bottom Line
Value's long-run premium remains plausible on theoretical grounds, but it requires tolerance for long underperformance and extra portfolio complexity, and it should be judged over decades rather than recent years.
