Attention and Underreaction: Why Some Anomalies Work Better in the Dark

This paper shows that many return anomalies—especially those tied to earnings, profitability, and momentum—perform much better when media coverage is low. Limited investor attention and short-sale constraints create predictable mispricing in these overlooked stocks.

💡 Takeaway:
Underreaction-related anomalies earn higher alpha among low media coverage stocks, especially from the overpriced short side. Media attention suppresses anomaly returns by reducing mispricing.


Key Idea: What Is This Paper About?

This study explores why many prominent return anomalies (e.g., momentum, ROE, earnings surprises) are stronger among firms with low media coverage. Using media attention as a proxy for investor awareness, the authors show that limited attention leads to underreaction, and that short-sale constraints prevent arbitrage correction—especially for overpriced stocks. This drives stronger anomaly returns among “ignored” firms.


Economic Rationale: Why Should This Work?

📌 Relevant Economic Theories and Justifications:

  • Limited Attention: Investors don’t fully process firm-specific news for less-covered firms.
  • Underreaction: Price adjustments to good/bad news are slow, creating predictable returns.
  • Limits to Arbitrage: Short-sale constraints prevent correcting overpricing on the short side.
  • Behavioral Biases: Inattention leads to expectation errors by both investors and analysts.

📌 Why It Matters:
Many anomalies are more profitable because they exploit systematic expectation errors. When attention is low, mispricing lasts longer, especially among hard-to-arbitrage stocks.


How to Do It: Data, Model, and Strategy Implementation

Data Used

  • Time Period: 2000–2018
  • Assets: US equities (CRSP + Compustat + RavenPack)
  • Attention Proxy: Monthly Dow Jones news count (log-transformed)
  • Anomalies Studied:
    • PEAD, MOM, ROE, PERF, PMU, RMW
    • Additional: SUE, RE, IndMom, ROA, NEI, FP, OS

Model / Methodology

  • Double Sorts: Stocks grouped by media coverage and anomaly signal
  • Fama-MacBeth Regressions: Control for firm size, beta, liquidity, ownership, and interaction terms
  • Event Study: Earnings announcement surprise differences (CAR and SUE) across coverage levels
  • Arbitrage Cost Proxies: Idiosyncratic volatility, illiquidity, institutional ownership, bid-ask spreads

Trading Strategy (Based on Results)

  • Signal: Focus on underreaction anomalies in low media coverage stocks
  • Execution:
    • Long: Undervalued firms with strong fundamentals or momentum
    • Short: Overvalued firms with bad signals and low coverage
  • Enhancement:
    • Filter stocks with high arbitrage cost (illiquidity, high idio-vol, low institutional ownership)
    • Rebalance monthly; earnings announcement periods are key windows

Key Table or Figure from the Paper

📌 Explanation:

  • FF5 Alpha spread: 0.97%/month for low media firms vs 0.24% for high media
  • Most alpha comes from short leg of anomalies (e.g., −1.21%)
  • Confirms investor inattention drives underreaction, and media attention moderates it

Final Thought

💡 Anomalies aren’t dead—they’re just hiding where no one’s paying attention. 🚀


Paper Details (For Further Reading)

  • Title: Attention and Underreaction-Related Anomalies
  • Authors: Xin Chen, Wei He, Libin Tao, Jianfeng Yu
  • Publication Year: 2023
  • Journal/Source: Management Science
  • Link: https://doi.org/10.1287/mnsc.2022.4332

Read next