Research Methods
Lecture 1
Erasmus Data Service Centre (EDSC) à use to access data
Can always contact them to ask for advice about which data platforms to use.
For sustainability data, go to the manual on the website of EDSC. All data is already
downloaded and posted there.
Three types of analysis:
Cross-sectional
Time series
Panel
How to deal with outliers?
Data transformation: “pull” extreme” values to the mean
Winsorizing: replace extreme values with upper/lower cut-off value
Truncating (=trimming): delete extreme observations
After treating outliers, always recalculate descriptive statistics. Also note that none of these
are applied to returns.
Portfolios can be:
Equally-weighted
Value-weighted
Price-weighted
Event Studies
A study that examines empirically what the stock price responses around corporate events
are. You look at similar announcements (events). It assesses the impact of corporate events
on firm value. Corporate events are situations when information about the firm is released
to the market. The idea is that if markets are efficient, a change in stock price around
announcement should reflect impact of event on firm value.
Under a semi-strong form, stock price should immediately jump to new equilibrium level at
announcement. However, there is evidence of under/overreaction, implying violation of
EMH. If you suspect market inefficiency, long-run event study can be useful to detect long-
term reversal to equilibrium price.
An event study focusses on abnormal returns to control for overall market movements. We
often need a cross-sectional analysis in an event study because it helps to understand:
- The sources of abnormal returns
- How abnormal returns vary across firm
- How abnormal returns depend upon characteristics of the event
, - Because aggregate abnormal returns may be zero, despite clear announcement
effects
Steps:
The event date is normally the announcement date. Note that announcements are
sometimes made at strategic moments. This may show up in abnormal returns before the
announcement. Stock price can also react before the announcement due to information
leakage.
Estimation window: to establish what a normal expected return is
Event window: most important, where you analyse what happens to the stock price of the
firm that I am looking at
In practise therefore, multiple event window lengths are used à robustness check
Event studies aim to estimate ‘abnormal’ return occurring around a specified corporate
event.
, Three categories of earnings announcements:
By relating CARs to firm characteristics (must be known when announcement is made) and
running a regression, you can:
Limitations/drawbacks of an event study:
Things to worry about:
Lecture 1
Erasmus Data Service Centre (EDSC) à use to access data
Can always contact them to ask for advice about which data platforms to use.
For sustainability data, go to the manual on the website of EDSC. All data is already
downloaded and posted there.
Three types of analysis:
Cross-sectional
Time series
Panel
How to deal with outliers?
Data transformation: “pull” extreme” values to the mean
Winsorizing: replace extreme values with upper/lower cut-off value
Truncating (=trimming): delete extreme observations
After treating outliers, always recalculate descriptive statistics. Also note that none of these
are applied to returns.
Portfolios can be:
Equally-weighted
Value-weighted
Price-weighted
Event Studies
A study that examines empirically what the stock price responses around corporate events
are. You look at similar announcements (events). It assesses the impact of corporate events
on firm value. Corporate events are situations when information about the firm is released
to the market. The idea is that if markets are efficient, a change in stock price around
announcement should reflect impact of event on firm value.
Under a semi-strong form, stock price should immediately jump to new equilibrium level at
announcement. However, there is evidence of under/overreaction, implying violation of
EMH. If you suspect market inefficiency, long-run event study can be useful to detect long-
term reversal to equilibrium price.
An event study focusses on abnormal returns to control for overall market movements. We
often need a cross-sectional analysis in an event study because it helps to understand:
- The sources of abnormal returns
- How abnormal returns vary across firm
- How abnormal returns depend upon characteristics of the event
, - Because aggregate abnormal returns may be zero, despite clear announcement
effects
Steps:
The event date is normally the announcement date. Note that announcements are
sometimes made at strategic moments. This may show up in abnormal returns before the
announcement. Stock price can also react before the announcement due to information
leakage.
Estimation window: to establish what a normal expected return is
Event window: most important, where you analyse what happens to the stock price of the
firm that I am looking at
In practise therefore, multiple event window lengths are used à robustness check
Event studies aim to estimate ‘abnormal’ return occurring around a specified corporate
event.
, Three categories of earnings announcements:
By relating CARs to firm characteristics (must be known when announcement is made) and
running a regression, you can:
Limitations/drawbacks of an event study:
Things to worry about:
