Efficient Market Hypothesis Flashcards
Stock Return Predictability (21 cards)
Format for Stock Return predictability
-What is EMH
-Different forms of EMH
-Time series predictability
-Implications for market efficiency
-Short and Long run predictability
-Lagged information variables
-Implications for market efficiency
-Joint hypothesis problem of Fama 1970
-Implications for market efficiency
-Cross-sectional predictability
What is EMH?
Theory that states financial markets fully reflect all available information at a given time
-Impossible to consistently achieve above average returns
Forms of EMH
Eugene Fama 1970
-Weak form efficiency
-Prices reflect past market data
-Technical analysis is useless
-Semi-strong form efficiency
-Prices reflect all publicly available information
-Fundamental analysis cannot consistently generate excess returns
-Strong form efficiency
-Prices reflect all public and private information
-Insider trading cannot produce excess returns
Time-series predictability
-Asks whether past price patterns or other time-related information can be used to predict future returns
-Directly links to weak-form EMH
-All historic prices reflected in current prices
-Trading on past prices should not be profitable
-Historically seasonal patterns were impossible due to the general knowledge of their existence leading to arbitrage
Explanations for time-series predictability
-Behavioural
-Risk based explanations
-Higher risk after market declines
-Market frictions
-Transaction costs, liquidity restraints
Empirical findings of time-series predictability
Elton and Gruber 2017
-Seasonal patterns (Sydney Watchel)
-January effect - tax loss selling and repurchase
-Behavioural
-Short-term momentum
-Long-term reversal - over long-term past losers outperform past winners
-Volatility clustering - asset returns exhibit periods of high and low volatility
-Mean reversion - stock prices revert to long term averages
Implications of time-series predictability
-If markets aren’t fully efficient investors can exploit time-series patterns
-Passive management is preferred with a weak or inconsistent market
Short-term predictability
Momentum (Jegadeesh & Titman 1993)
-Stocks performing well in last 3-12 months will continue outperforming in the next 6-12 months
Cross-sectional momentum
-Forming portfolios with stocks with good past
Time-series momentum (Moskowitz 2012)
-Assets with positive returns continue while losers keep losing
Long-term predictability
Reversal (De Bonat & Thaler 1985)
-Poor performing stocks (3-5 years) tend to rebound
-Past winners underperform
Mean reversion (Shiller 1981)
-High P/E predicts lower future returns
Other lagged information variables
Dividend yield
-High D/P - High future returns
Term spread (Yield curve)
-Steep curve - High equity returns
Inflation
-High Inflation - Low future returns
Volatility
-High VIX - Low near-term returns
Implications of short, long term predictability and other lagged information variables
Weak form efficiency
-If past returns predict future, technical analysis may have some merit
Semi-strong efficiency
-Macro variables predictive power suggests public information is not instantly reflective
Highlighting the problem
-January effect
-Inefficiency - Tax-loss selling creates seasonal mispricing
-Model misspecification - CAPM underestimate small-cap risk
-Momentum Profits
-Inefficiency -Investors under react
-Risk - compensation for unmodeled risk
Resolving the problem
-Better asset pricing models
-Out-of-sample tests
-Behavioural experiments
The Joint Hypothesis problem
Eugene Fama 1970
“We cannot test efficiency without assuming a pricing model, and we cannot validate a pricing model without assuming efficiency”
-Asset pricing models such as CAPM / Fama-French models
-If test find inefficiencies
-Markets are inefficient
-Pricing model is misspecified
Implications of the Joint Hypothesis Problem
-Challenges in empirical research
-False rejections of EMH
-No ‘clean’ test without perfect model
-Behavioural finance vs Risk based
-Behavioural - anomalies suggest inefficiency
-Risk based view - anomalies stem from missing risk factors
-Practical consequences
-Active management works if anomaly reflects inefficiency
-Passive management work if anomaly reflects risk
-Model dependency - findings change as models evolve
-Fama - French model explained value / growth as risk premiums
Cross-sectional predictability
Ability to predict difference in returns
between assets over a single point in time (which stocks will outperform others)
-Value effect (cheap vs expensive)
-Momentum effect (past winners vs losers)
-Size effect (small vs large)
-Profitability & Investment
-Volatility effect
Value effect (Cross-sectional predictability)
Value effect (cheap vs expensive)
-P/B, P/E, high dividend yield
-Value stocks outperform growth stocks
Risk - Value stocks are distressed and riskier (Fama & French 1992)
Behavioural - Investors overextrapolate poor past performers
Momentum effect (Cross-sectional predictability)
Momentum effect (past winners vs losers)
-Past winners continue outperforming (Jegadeesh & Titman 1993)
Risk - Momentum captures macroeconomic risk exposure
Behavioural - Slow information diffusion (Hong & Stein 1999)
Size effect (Cross-sectional predictability)
Size effect (small vs large)
-Market capitalisation
-Small cap stocks outperform large (Banz 1981)
Risk - Small firms less liquid, more vulnerable to shocks
Data mining - Size effect weakened post-discovery (Schwert 2003)
Profitability & Investment (Cross-sectional predictability)
Profitability & Investment
-High profitability (ROE, ROA), low investment outperform (Fama & French 2015)
-Rational pricing - Disciplined investment earn higher returns
Volatility effect (Cross-sectional predictability)
Volatility effect
-Low volatility stocks outperform (Baker 2011)
Behavioural - Investors overpay for high-volatility stocks
Leverage constraints - Institutions prefer high-beta stocks (more volatile)
