Events Studies and Market Efficiency

1
Events Studies and Market Efficiency

• An event study intends to study the economic
  impact of some event on a firm
• CEO dies
• Merger announced
• Litigation

2
Event Study Methodology

• Methodology
• Assume the stock price P-1 the day before the
  announcement is efficient
• Captures true value of the firm, given all
  available info
• Assume that the closing stock price on the day of
  the announcement P0 (day after the announcement?)
  is efficient
• Captures true value of firm, given all available
  info
• Wed like to interpret the return (P0-P-1)/P-1
  as the impact of new information

3
Methodology

• But we need to correct for market movements
  during the day.
• Use the CAPM Estimate the firms beta using
  historical data, then compute
• If we are interested in short-horizons, just use
• Or, use Fama-French

4
Abnormal Return

• Abnormal return on each day is
• AR Actual return E(r)
• You can also look at abnormal returns over longer
  windows.
• Then, define cumulative abnormal returns from day
  0 to day T as

5
Event Study Takeover Announcements
Announcement Date
6
The Definition

• A capital market is said to be efficient if asset
  prices fully reflect available information.
• What kind of information?
• Strong form (all obtainable private information)
• Semi-strong form (all public information)
• Weak form (prices, or history of prices)

7
Tests of semi-strong EMH

• If semi-strong EMH holds, we should find that
  stock prices respond to new information quickly
  and accurately.
• Flat tail in Event Studies (after the blue
  line)
• In other words, publicly-known events should
  not result in sustained profit opportunities.

8
Evidence

• Most event studies do not conclude that there are
  profit opportunities
• Large ones, anyway
• The market seems to be relatively semi-strong
  efficient (to the most obvious events). Any
  exceptions?
• Some continuing positive return after good
  earnings announcements.
• Most violations of SSEMH are more subtle and
  temporary.

9
Weak form EMH

• If the weak form of the EMH is true, then
  investors will be unable to predict future
  unexpected returns using price data.
• There are no price patterns that are good or
  bad news.
• Which would imply there is no autocorrelation in
  stock returns

10
Evidence of Random Walk
11
Better Test of Weak Form

• Look for autocorrelation in returns.
• Pull off data for a stock (or an index)
• In column H, put returns
• In column I, put lagged returns (i.e. returns
  shifted down by one day)
• Look for patterns. If you find none, this is
  consistent with weak-form.

12
Conclusions

• If the weak-form EMH is true, it should be
  impossible to earn abnormal profits from
  employing a trading rule based on price data.
• Which would imply, if true, that technical
  analysis will not work.
• This is with good reason If a simple trading
  rule existed, enough people would use it that
  its efficacy would disappear!

13
Autocorrelation at long horizons

• Do stocks have positive / negative / zero
  autocorrelation at long horizons?
• See 3/19 spreadsheet on my web page
• What does this say about Market Efficiency?