Transcript A Survey of Behavioral Finance

A Survey of Behavioral
Finance
Nicholas Barberis
Richard Thaler
Presented by Ryan Samson
Traditional vs. Behavioral
• Traditional
– Rational
• Correct Bayesian
Updating
• Choices Consistent with
Expected Utility
• Behavioral
– Some are Not Fully
Rational
– Relax One or Both
Tenets of Rationality
Behavioral Finance
Limits to Arbitrage
Psychology
Limits to Arbitrage
vs.
Market Efficiency
• EMH
– Prices Reflect Value
– Mispricings Corrected by Arbitrageurs
• Limits to Arbitrage
– Strategies May not be Arbitrage
– Problems Entering Position?
• Correct Prices => No Free Lunch
• No Free Lunch ≠> Correct Prices
• Why Care?
Theory Supporting
Limits to Arbitrage
• Fundamental Risk – Negative Shock and
no Perfect Substitute (e.g. Ford and GM)
• Noise Trader Risk – Continued
Widespread Irrationality
– Forced Liquidation (Separation of Brains and
Capital)
– Horizon Risk
– Trading in the Same Direction
Theory Supporting
Limits to Arbitrage 2
• Implementation Costs
– Commission
– Bid/Ask Spread
– Price Impact
– Short Sell Costs
• Fees
• Volume Constraints
• Legal Restraints
– Identification Cost
• Mispricing ≠> Predictability
Evidence Supporting
Limits to Arbitrage
• Mispricings Hard to Identify
– Test of Mispricing => Test of Discount Rate Model
• Twin Shares
– Royal Dutch (60%) and Shell (40%)
• Only Risk is Noise Traders
– => PriceRD = 1.5*PriceS
Evidence Supporting
Limits to Arbitrage 2
• Index Inclusions
– Stock Price Jumps Permanently
– 3.5% Average
• Fundamental Risk
– Poor Substitutes (best R2 < 0.25)
• Noise Trader Risk
– Index Fund Purchases etc.
Evidence Supporting
Limits to Arbitrage 3
• Internet Carve-Outs
– 3Com Sells 5% of Palm in IPO, Will Spin Off
Remainder in 9 Months
– 1 Share of 3Com will own 1.5 Shares of Palm
– PPalm = $95
– 3Com should be ≥ $142
– P3Com = $81
– Value of 3Com Excluding Palm = -$60
Evidence Supporting
Limits to Arbitrage 4
• Why?
• Very Few Shares of Palm available to
Short
• Arbitrage was Limited
• Mispricing Persisted
Psychology
• Beliefs
– Overconfidence
• 98% CI only captures 60%
• 100% is actually 80% and 0% is actually 20%
– Optimism / Wishful Thinking
• Unrealistic View of Personal Abilities / Prospects
• 90% of Drivers Claim Above Average Skill
• 99% of Freshman Claim Superior Intelligence
Psychology 2
• Beliefs Continued
– Representativeness
• Base Rates are Under-Emphasized Relative to
Evidence
• Sample Size Neglect in Learning Distribution
– (6 Tosses vs. 1000 Tosses)
• “Law of Small Numbers”
– Gambler’s Fallacy
– Conservatism
• Base Rates are Over-Emphasized Relative to
Evidence
Psychology 3
• Beliefs Continued
– Belief Perseverance
• Search for Contradictory Evidence
• Treatment of Contradictory Evidence
– Anchoring
• Initial Arbitrary Value and Make Adjustments
– Availability Biases
• Recent or Salient Events
Psychology 4
• Beliefs, Final Notes
– People Display Poor Learning in Application
– Experts Often do Worse
– Increasing Incentives Doesn’t Help
Psychology 5
• Preferences
– Expected Utility vs. Prospect Theory or
Ambiguity Aversion
• Prospect Theory
– Value of a Gamble is: π(p)*v(x)+π(q)*v(y)
– Utility Defined over Gains and Loses
– Concave over Gains, Convex over Losses
– Nonlinear Probability Transformation
• Especially Large Weight on Certain Outcomes
Psychology 6
• Ambiguity Aversion
– People Avoid Uncertain Probability
Distributions
– Aversion Changes Based on Perceived
Competence at Assessing Relevant
Distribution
• Preference for Familiar
Application 1:
Aggregate Stock Market
• 3 Puzzles:
– Equity Premium
– High Volatility in Returns and P/D Ratios
– Predictable Returns (D/P alone  R2 = 0.27)
Equity Premium
• Risk Premium Seems too High
• Possible Explanations Under Prospect
Theory
– Benartzi and Thaler
• Eπv[(1-w)Rf,t+1 + wRt+1 – 1], π and v as before
• Given Historical Returns, Investors are Indifferent
to w = 1 and w = 0
• Calculate Implied Length of t
• 1 Year (Taxes? Annual Reports?)
• Result is Myopic Loss Aversion
Equity Premium 2
• Possible Explanations Under Prospect
Theory Continued
– Need Intertemporal Model
– Barberis, Huang, Santos
• Utility From Consumption (Source 1) AND Utility
From Changes in Value of Risky Assets (Source 2)
• Utility From Source 2 Captures Loss Aversion (Not
Convexity, Concavity, or Nonlinearity of π)
• Explanatory power based on weight of Source 2
Equity Premium 3
• Possible Explanations Under Prospect
Theory, Final Notes
– Why?
– Regret
– Bounded Rational:
• P(C(Labor Income, Stock Returns) < Habit)
• P(C(Stock Returns) < Habit)
– t = 1 Year Based on Presentation
Equity Premium 4
• Explanations Under Ambiguity Aversion
– Max[Min[E[U]]] (i.e. Playing Malevolent
Opponent)
– Requires High Equity Premium
Volatility
• Rational Approaches Must Focus on Changing
Risk Aversion to Explain Volatility
• Explanations Under Beliefs
– Overreaction to Dividend Growth  Volatile Prices
• Law of Small Numbers
• Overconfidence in Opinion
– Overreaction to Returns
• Law of Small Numbers
– Confusion Between Real and Nominal Rates
Volatility
• Explanations Under Preferences
– Same Model as Used for Equity Premium
– Add zt, a State Variable, to Source 2 of Utility
– Several Price Increases  Less Scared
– Price Decreases  Scared
Application 2:
Cross-Section of Average Returns
• You Can Form Groups of Stocks w/
Different Average Returns, Not Explained
by CAPM
– Size Premium (Small Stocks +0.74%/month)
– Long Term (3 Yr) Reversal (8%/Yr)
– Price Ratios
• B/M (High B/M +1.53%/month)
• P/E (High P/E +0.68%/month)
– Momentum (6 Month Winners +10%/Yr)
Cross-Section of Average Returns
• Anomalies Continued
– Earnings Announcements (Over 60 Days +4%
for Good Over Bad)
– Dividend Initiation / Omission
– Stock Repurchases
• Problems w/ Anomalies
– Difficult Statistics (Cross-Sectional
Correlation)
– Data-Mining (Test Out of Sample)
• Multi-Factor Models
Cross-Section of Average Returns
• Explanations Under Beliefs
– Conservatism (Underweight New Info)
• React Slowly to Earnings Reports
– Representativeness
• Overreact Now, Reversal Later
– Overconfidence
• Ignore Unfavorable Public Info  Reversal
• Too Much Attention to Favorable Public Info 
Momentum
– All Imply P Around Earnings Report
Cross-Section of Average Returns
• Belief Based Continued
– Positive Feedback
•
•
•
•
Momentum
Post Earnings Drift
Long Term Reversal
A Result of Law of Small Numbers?
Cross-Section of Average Returns
• Belief Based w/ Institutional Friction (i.e.
Short Sell Constraints)
– Bearish Cannot Short  Reversal or
Momentum
– Effect of Higher Incentives on Short Prices
Cross-Section of Average Returns
• Preference Based Explanations
– Same BHS Model Applied to Individual
Stocks
• Price Reversal (Not Momentum)