Summary

In this 2006 study, authors Christian Funke, Timo Gebken, Lutz Johanning, and Gaston Michel investigate whether industries exhibit predictable return patterns following merger and acquisition (M&A) announcements. The paper finds that industries with positive reactions to M&A announcements tend to continue performing well, while those with negative reactions tend to underperform. This behavior suggests a systematic drift in industry returns, hinting at market underreaction to industry-wide information.

The authors construct a zero-cost trading strategy — going long on industries with positive announcement returns and short on those with negative ones. The strategy remains profitable even after adjusting for standard asset pricing factors (e.g., size, book-to-market).

Key Ideas

  • Persistence of Reaction (Page 2, Paragraph 2):
    Industry reactions to M&A announcements show temporal persistence. The average reaction in one month helps predict future intra-industry effects and returns, indicating underreaction in capital markets.

  • Return Predictability (Page 2, Paragraph 3):
    Monthly industry returns correlate with the previous month’s announcement reactions. A 72 basis point return spread (8.6% annualized) exists between industries with prior positive versus negative M&A reactions.

  • Investment Strategy (Page 3, Paragraph 2):
    A simple long-short strategy based on prior month’s announcement reactions yields significant profits — 105 basis points (raw) or 75 basis points (adjusted) in one month, with diminishing but significant returns over longer horizons.

  • Profit Attribution (Page 12, Paragraph 2):
    The short side of the strategy (betting against negatively reacting industries) contributes more to profitability, possibly reflecting short-selling constraints and stronger market underreaction in declining industries.

Data & Methodology

  • Source:
    M&A data from SDC U.S. Mergers and Acquisitions Database (1985–2002)
    Stock returns from CRSP
    Book values from COMPUSTAT

  • Key Data Fields:
    • M&A transaction value (minimum $25M)
    • Industry classification via 2-digit SIC codes
    • Daily and monthly stock returns
    • Book-to-market ratios
    • Size (market capitalization)
    • Cumulative Abnormal Returns (CARs) around announcement dates
    • Size and B/M adjusted industry returns
  • Method:
    The authors compute five-day CARs around each M&A announcement using a modified market model and track equal-weighted industry returns. They use Fama-French adjustments, rolling portfolio formation, and Fama-MacBeth regressions for robustness.

Potential Issues and Clarifications

1. Market Inefficiency Interpretation

  • Lack of full consideration of transaction costs:
    • Frequent monthly rebalancing can reduce net profitability.
    • Real-world execution costs and liquidity limits may erode returns.
  • Assumes widespread investor inattention:
    • Institutional investors may already factor in such reactions.
    • Strategy may not scale without impacting market prices.

2. Strategy Scalability and Implementation

  • Requires dynamic, monthly industry-level portfolio shifts:
    • Operationally intensive and cost-heavy for real-world application.
    • Susceptible to slippage in thinly traded industry components.
  • Equal-weighting may not reflect investable strategies:
    • Market-cap weighting could yield different (possibly lower) returns.
    • Larger stocks dominate in practice due to liquidity and accessibility.

3. Use of Historical Data (1985–2002)

  • Outdated market conditions and structure:
    • Current M&A landscapes and market microstructure have evolved.
    • Algorithmic trading and speed advantages could negate effect.
  • Changes in regulation and disclosure:
    • Post-SOX and Dodd-Frank environment alters corporate behavior.
    • Information dissemination is now faster and more transparent.

4. Risk Factor Adjustments

  • Relies on size and book-to-market controls only:
    • May omit relevant modern factors (e.g., momentum, quality, volatility).
    • Fama-French three-factor model may not fully capture risk exposures.
  • Statistical significance vs. economic significance:
    • Strategy alpha is significant, but actual risk-adjusted returns may vary.
    • Omitted variable bias or data-snooping could inflate t-stats.

Reflection

This paper provides compelling evidence that merger announcements carry predictive power beyond the firm level, extending to industry-wide effects. The strategy’s consistent outperformance — particularly from short positions — challenges the notion of market efficiency. While the strategy may face implementation barriers like transaction costs and liquidity, the strength and persistence of the return patterns suggest real trading and forecasting value for practitioners and researchers alike.

Reference