You are on page 1of 22

Behavioral Finance

Alok Kumar Yale School of Management 8 December 1999

• • • • • Efficient Market Hypothesis (EMH) Expected Utility; Rational Expectations Few Examples Prospect Theory (Kahneman and Tversky) Behavioral Heuristics and Biases in Decision Making • Implications for Financial Markets

. • Cannot “make money” using “stale information”. • Michael Jensen: “there is no other proposition in economics which has more empirical support than the EMH”. – Strong form: all public AND private information. – Semi-strong form: all public information.Market Efficiency • Fama: “The market price at any time instant reflects all available information in the market”. • Three forms – Weak form: past prices and returns.

They exhibit “biases” and use simple “heuristics” (rules of thumb) in making decisions. – Investors may sell winning stocks and hold onto losing stocks (Odean). – investors trade actively (Odean). . • Empirical Evidence on investor behavior: – investors fail to diversify. – extrapolative and contrarian forecasts.Challenges to EMH • Investors are not “fully rational”.

.Expected Utility Theory • A theory of choice under uncertainty for a single decision-maker. • Expected Utility = p1*u1 + p2*u2 + … + pn*un. Example: transitivity. cancellation. p: probability of an event u: utility derived from the event • Based on several strong assumptions about preferences.

Rational Expectations Paradigm • All investors are identical. Expected Utility + Rational Expectations => Market Efficiency . • All investors use “Bayes rule” to form new beliefs as new information becomes available. • All investors are utility maximizers. • All investor predictions are accurate.

– Momentum in stock prices: short-term trends (6-12 months) continue. • Challenges: – Excess market volatility – Stock price over-reaction: long time trends (1-3 years) reverse themselves. . – Size and B/M ratio (stale information) may help predict returns.Are Financial Markets Efficient? • Weak form of market efficiency supported to a certain extent.

. • Inclusion of a stock in the S&P500 index results in significant share price reactions. • 50 largest one-day stock price movements: occurred on days of no major announcements.6% decline without any apparent news. Example: AOL rose 18% on the news of its inclusion in the index.Stock Price Reaction to Non-Information • Crash of 1987: 22.

Information processing limitations. . • Investor Sentiment: beliefs based on heuristics rather than Bayesian rationality. • Investors may react to “irrelevant information” and hence may trade on “noise” rather than information.Role of Investor Behavior • Bounded Rationality: “satisficing” behavior. Example: memory limitations.

• Window-dressing: add to the portfolio stocks that have done well in the recent past and sell stocks that have recently done poorly. . • Herding: may select stocks that other managers select to avoid “falling behind” and “looking bad”.“Irrational” Behavior of Professional Money Managers • May choose a portfolio very close to the benchmark against which they are evaluated (for example: S&P500 index).

a 50% chance to gain $0. . • Initial endowment: $500. a 50% chance to lose $0.An Example • Initial endowment: $300. Consider a choice between: – a sure gain of $100 – a 50% chance to gain $200. Consider a choice between: – a sure loss of $100 – a 50% chance to lose $200.

. • Problem framed as a loss: decision maker is risk seeking. 28% chose option 2. => A reversal in Choice • Problem framed as a gain: decision maker is risk averse. 64% chose option 2. • Case 2: 36% chose option 1.Reversal in Choice • Case 1: 72% chose option 1.

1. – $5 million with prob 0.89 and $0 with prob • Case 2: consider a choice between: – $1m with prob 0.10 and $0 with prob 0. $1m with prob 0.Allais Paradox • Case 1: consider a choice between: – $1 million with certainty. $0 with prob 0.89. – $5m with prob 0. .

Allais Paradox: Explanation u(1m) > 0. 0.01*u(0m) Add 0.10*u(5m) + 0.11*u(1m) + 0.0.90*u(0m) Violates Expected Utility Theorem! .10*u(5m) + 0.89*u(1m) + 0.89*u(1m) to both sides.89*u(0m) .89*u(0m) > 0.

• Decision maker analyzes “gains” and “losses” differently. • Incremental value of a loss is larger than that of a loss. • Gambles are evaluated relative to a reference point.Prospect Theory • Proposed by two psychologists: Daniel Kahneman and Amos Tversky. . “the hurt of a $1000 loss is more painful than the benefit of a $1000 gain”.

especially ones that involve uncertainty? • Commonly Used Heuristics – Availability: “familiarity breeds investment”. – Reliance on the judgement of other people (Keynes beauty contest analogy). . “Patterns in random sequences”. – Representativeness: judgement based on similarity.Behavioral Heuristics and Decision-Making Biases • What strategies do decision makers use when faced with difficult decisions.

THTHTHHHHHH -> P(T) = ?.Gambler’s Fallacy • Investors may apply law of large numbers to small sequences. • Which of the 2 sequences is more likely to occur in a fair coin tossing experiment? – HHHHHHTTTTTTHHHHHH – HHTHTHHTHTTHTHHTTH . Example: fair coin tossing. P(H) = ?.

• Anchoring and adjustment: can create under-reaction.Some more Heuristics • Overconfidence: people overestimate the reliability of their knowledge. • Disposition Effect: sell winners. . – Excessive trading • Framing Effect • Regret Aversion: anticipation of a future regret can influence current decision. hold on to the losers.

.Fashions and Fads • People are influenced by each other. • Informational Cascades • Positive Feedback • Example: excessive demand for internet IPOs. Extremely high opening day returns. There is a social pressure to conform. • Herding behavior: “safety-in-numbers”.

Can arbitrage opportunities exist? • Yes! – Real-world arbitrage is always risky. – Close substitutes (needed for arbitrage positions) may not be available. – Fundamentally identical assets may NOT sell at identical prices. No riskless hedge for the arbitrageur. . – Arbitrageur faces“noise trader” risk: mispricing can become worse before it disappears.

hence it does not counter irrational disturbances. – Prices may not react to information by the “right” amount. – Prices may react to non-information. . – Markets may remain efficient. • Limited arbitrage: arbitrage is never riskfree.Behavioral Finance: Two Major Foundations • Investor Sentiment: creates disturbances to efficient prices.

. How much? Not clear! • Both “social” and “psychological” must be taken into account in explaining the behavior of financial markets. • Market “anomalies” may be widespread. • Behavioral Finance: does not replace but complements traditional models in Finance.Summary • Investor behavior does have an impact on the behavior of financial markets.