📖 市场异象（Anomalies）

Equity Investments

I. Lesson Focus

This lesson examines market anomalies that appear to generate persistent risk-adjusted excess returns, challenging the Efficient Market Hypothesis (EMH). Candidates must be able to classify anomalies (calendar, fundamental, technical), explain their theoretical foundations using behavioral finance, compute risk-adjusted performance, and critically evaluate whether anomalies remain exploitable after transaction costs, liquidity risk, and data-mining bias. The material integrates directly with the Fama-French factors, momentum strategies, and behavioral explanations covered in equity valuation and portfolio management.

II. The Problem

Imagine you manage an equity fund and observe that small-cap stocks have delivered an average excess return of 4.2% in January over the past decade, while low price-to-book stocks have outperformed high price-to-book stocks by 7.8% annually. These patterns appear to offer “free lunches.” Do they contradict semi-strong form efficiency? Can systematic strategies reliably capture them after costs and risks? This lesson dissects the evidence, economic explanations, decay patterns, and practical pitfalls so candidates can accurately assess anomaly validity in both exam scenarios and real-world applications.

III. Efficient Market Hypothesis (EMH) Review and Definition of Anomalies

The Efficient Market Hypothesis asserts that asset prices fully reflect available information. It exists in three forms: weak (technical analysis useless), semi-strong (fundamental analysis useless), and strong (even insider information useless).

Market anomalies are pricing patterns that produce statistically significant, persistent risk-adjusted excess returns inconsistent with EMH. They are typically grouped into three categories: - Calendar anomalies (time-based patterns) - Fundamental anomalies (patterns linked to financial characteristics) - Technical anomalies (patterns based on price and volume)

A genuine, replicable anomaly constitutes a direct challenge to EMH, particularly the semi-strong form.

IV. Calendar Anomalies

1. January Effect

Small-capitalization stocks exhibit significantly positive abnormal returns in January. First documented by Rozeff and Kinney (1976). Primary explanations include: - Tax-loss harvesting: investors sell loss-making stocks in December and repurchase in January, pushing up small-cap prices. - Window dressing: portfolio managers sell risky small stocks at year-end and repurchase in January. - Higher risk: small stocks may carry elevated systematic risk in January.

Empirical evidence shows the effect has weakened substantially since the 1980s, especially after transaction costs and risk adjustment.

2. Weekend Effect

Stocks tend to earn negative or the lowest average returns from Friday close to Monday open. Explanations involve the release of bad news over weekends and selling pressure on Monday when traders cannot react immediately.

3. Turn-of-the-Month Effect

The last few trading days of the month and first few days of the next month generate significantly higher returns than mid-month periods. Improved liquidity and salary inflows are key drivers.

V. Fundamental Anomalies

1. Value Anomaly

Stocks with low price-to-earnings (P/E), low price-to-book (P/B), or high book-to-market (B/M) ratios tend to outperform growth stocks over long horizons. The HML (High Minus Low) factor in the Fama-French three-factor model captures this premium.

Formula:
Value premium ≈ E(R_HML) = Return on high B/M portfolio − Return on low B/M portfolio

2. Size Anomaly

Small-cap stocks have historically outperformed large-cap stocks. The SMB (Small Minus Big) factor isolates this effect.

Formula:
SMB = (1/3)(Small Value + Small Neutral + Small Growth) − (1/3)(Big Value + Big Neutral + Big Growth)

3. Profitability and Low-Volatility Anomalies

Stocks with high gross profitability, stable earnings growth, low beta, or low idiosyncratic volatility tend to deliver superior risk-adjusted returns. This contradicts the traditional CAPM prediction that higher risk commands higher expected return.

4. Momentum Anomaly

Stocks that performed well over the past 3–12 months continue to outperform over the subsequent 6–12 months (Jegadeesh and Titman, 1993). This anomaly blends technical and fundamental characteristics.

Momentum strategy formula:
Momentum Return = Return on past winners (t+1) − Return on past losers (t+1)

VI. Behavioral Finance Explanations of Anomalies

Most anomalies are best explained by behavioral biases: - Overconfidence → momentum and reversals - Loss aversion and the disposition effect → momentum continuation - Representativeness heuristic → undervaluation of value stocks - Herding → amplification of calendar effects

These biases cause prices to deviate from fundamental value, creating exploitable opportunities. However, arbitrage is limited by noise-trader risk, implementation costs, and career risk (Shleifer & Vishny, 1997).

VII. Sustainability and Decay of Anomalies

Many anomalies weaken significantly after academic publication and widespread investor awareness (McLean & Pontiff, 2016). Reasons include: - Inflows of institutional capital that arbitrage away the mispricing. - Out-of-sample failure (data-snooping bias). - Underestimation of transaction costs, liquidity risk, and tail risk.

A common exam trap is assuming that a statistically significant historical anomaly remains economically exploitable net of all frictions.

Worked Cases

Case 1: January Effect Risk-Adjusted Analysis

An analyst observes 1990–2020 data: small-cap stocks returned 8.4% in January, large-caps returned 3.1%. Risk-free rate = 1.2%, small-cap β = 1.35, large-cap β = 0.95, market excess return in January = 4.8%.

CAPM expected returns:
Small-cap: 1.2% + 1.35 × 4.8% = 7.68%
Actual = 8.4% → Alpha = 8.4% − 7.68% = +0.72%

Large-cap: 1.2% + 0.95 × 4.8% = 5.76%
Actual = 3.1% → Alpha = 3.1% − 5.76% = −2.66%

Conclusion: Even after beta adjustment, small-caps show positive January alpha, but after 0.8% transaction costs the net alpha becomes insignificant.

Case 2: Value versus Growth Long-Term Simulation

Assume $100,000 initial investment, 1980–2022:
- Low P/B portfolio: 13.8% annualized, volatility 18.2%
- High P/B portfolio: 9.4% annualized, volatility 15.6%
- Market benchmark: 10.9% annualized

Cumulative wealth:
Low P/B ≈ $100,000 × (1.138)^43 ≈ $28.5 million
High P/B ≈ $100,000 × (1.094)^43 ≈ $4.8 million

Sharpe ratios: Low P/B = (13.8 − 3.5)/18.2 ≈ 0.57; High P/B = (9.4 − 3.5)/15.6 ≈ 0.38. The value strategy is superior on a risk-adjusted basis, yet the spread narrowed dramatically after 2000.

Case 3: Momentum Strategy — 6-Month Holding Period

Stocks ranked by prior 6-month returns:
- Winner decile: +42%
- Loser decile: −18%

Subsequent 6-month returns: winners +11.2%, losers −2.4%, market +4.1%.

Momentum profit = 11.2% − (−2.4%) = 13.6% over 6 months (≈ 27.2% annualized before costs). With monthly rebalancing, transaction costs of 8–12% per year can reduce net alpha to near zero.

Traps

Key Formulas

• SMB = (1/3)(Small Value + Small Neutral + Small Growth) − (1/3)(Big Value + Big Neutral + Big Growth)
• HML = (Small Value + Big Value)/2 − (Small Growth + Big Growth)/2
• Momentum Return = R_Winners,t+1 − R_Losers,t+1
• Alpha = R_i − [R_f + β_i(R_m − R_f)]
• Risk-adjusted excess return = Actual return − CAPM expected return
• Net alpha after decay ≈ Gross anomaly return − transaction costs − liquidity premium − implementation lag losses

Practice Questions

Q1. According to the Fama-French three-factor model, which factor most directly corresponds to the value anomaly?
A. SMB
B. HML
C. MOM
D. CMA

Q2. The primary traditional explanation for the January effect is:
A. Institutional window dressing
B. Tax-loss harvesting creating buying pressure in January
C. Companies releasing good news in January
D. Small-cap stocks having betas significantly below 1 in January

Q3. Which anomaly presents the most direct challenge to the semi-strong form of the EMH?
A. Excess returns from technical analysis
B. Excess returns from a value strategy based on publicly available book-to-market ratios
C. Excess returns from insider trading
D. Failure of the random-walk test

Q4. A stock returned +35% over the past six months while the market returned +12%. In a 6-month momentum strategy, this stock would most likely be classified in:
A. The loser portfolio
B. The winner portfolio
C. The neutral portfolio
D. The reversal portfolio

Q5. McLean and Pontiff (2016) found that anomaly excess returns decline by approximately what percentage after academic publication?
A. 0%
B. 26%
C. 58%
D. More than 100%

Q6. Calculation: A small-cap value portfolio returns 12% monthly, β = 1.4, market excess return = 6%, risk-free rate = 0.3%. Its CAPM alpha is closest to:
A. +3.3%
B. +2.1%
C. +4.5%
D. −1.2%

Q7. Which of the following is least likely to explain the momentum anomaly?
A. Investor overconfidence and self-attribution bias
B. The disposition effect causing winners to be sold too early
C. Slow diffusion of fundamental information
D. Instantaneous information reflection in an efficient market

Q8. When evaluating whether an anomaly is investable, a CFA candidate should least focus on:
A. Out-of-sample test results
B. Transaction costs and turnover
C. The impact factor of the academic journal that published the anomaly
D. Tail-risk and liquidity-risk exposures

Answers

Takeaways

• Market anomalies are persistent risk-adjusted excess return patterns inconsistent with EMH and fall into calendar, fundamental, and technical categories.
• The most prominent anomalies—value (HML), size (SMB), and momentum—are largely explained by behavioral biases rather than risk mis-measurement.
• Publication of academic research typically causes substantial decay; net-of-cost, net-of-risk alphas are often close to zero.
• Calendar anomalies such as the January and weekend effects have weakened markedly in institutionalized markets.
• Proper evaluation of anomalies requires out-of-sample testing, full risk adjustment, and subtraction of all implementation frictions.
• Behavioral biases (overconfidence, disposition effect, herding) provide the dominant theoretical foundation for why anomalies persist.