How to test for marketefficiency.
FIN 352 – Professor Dow
How to test?
Fama:Test the efficientmarket hypothesisusing different information sets.Three categories:WeakSemi-StrongStrongSome tests directly use this categorization, others do not.
Weak form efficiency
All past-price information is fully reflected in stock prices.Can’t use past prices to forecast future prices.If true, technical analysis is not useful.
Semi-Strong Efficiency
All public information is fully reflected in stock prices.If true, fundamental analysis is not useful.
Strong Form Efficiency
All information is reflected in stock prices.Implies that trading on insider information shouldn’t be profitable.Not trueBut not legal
Examples of tests
A) Patterns in stock prices.B) Back-testing trading rules.C) Do categoriesofstocks earn abnormal returns?D) Event studies.E) Do stock prices move “too much?”F) Bubbles.G) Do some investors outperform the market?
A) Patterns in Stock Prices
Serial Correlation > 0, MomentumSerial Correlation < 0, Mean ReversionSerial Correlation = 0, Random WalkWeak Form EMH predicts random walk
B)BacktestingTrading Rules
See if trading rules are profitable when applied to historical stock price data.Data MiningIn-Sample vs. Out-of-Sample
C) Do some types of stocks earn abnormal returns
Value stocksSmall stocksOr is it microcap/neglected stocks?Is it is risk premium?
D) Event Studies
Abnormal returns: Stocks earn greater returns than they “should”:Ri– E(R)Theory implies that stocks should earn abnormal returns when news first comes out, but not afterwards (stock prices are quick to adjust to news)Book gives example where they use excess returns (Ri-Rm) to measure response to event. Response is slower than it should be.
E) Bubbles
Increases in asset prices not justified by “fundamentals”At some point, bubbles pop!Shouldn’t have bubbles if markets are efficient.Recent experience with real estate and stock price bubbles.
F) Do stock prices move “too much”
Theory: Stock price is the present value of expected future dividend payments.Stock prices shouldn’t vary more than dividends or earnings do.But there is more variationSimilar idea to bubbles: stock prices move based on psychological reasons rather than fundamental reasons.
G) Do some investors outperform others?
Why do some investors do well?LuckHigher riskSkillMutual funds tests
What do we know
Markets are broadly efficient, but some important exceptions.BubblesSome people understand the economy better – but do you?
Implications for investing
Build around index funds: Well-diversified and low costDo bubbles imply market timing?Do you want to engage in fundamental analysis?
0
Embed
Upload