Lecture 6: Event study

Methodology
 9

Event Study

• Today we are looking at the application of a methodology : The Event Study.

• There is a survey paper on moodle (MacKinlay) and also an Excel

programme to run an event study. You are welcome to use this programme
for your dissertations.

• Event studies look at the impact of a ‘shock’ or Event on the risk-adjusted

share price. We are looking at the BP share price and the shock is the oil spill
in the Gulf of Mexico on 20th April 2010.

• We model the ‘expected’ share price of BP using the CAPM using data up to
the period immediately before the Event. We then use this model to measure
the ‘expected’ share price during the event. The difference between the actual
share price and the ‘expected’ share price is our estimate of the impact of this
event; This is the Abnormal Return(AR).
 9

Event Study: writing the methodology for a proposal

• The methodology is the key part of your proposal. You therefore would need
to outline on how you will use this model in your research.

• Identify the model used (the CAPM) and how it will be applied. It is
important to reference the source of your model and to give the relevant
equations.

• We need to identify the estimation and event windows and also the important
outputs that are being measure; these are the abnormal return (AR) and
cumulative abnormal return (CAR).

• See example of event study proposal on Moodle Lecture 6.

 9

Event Study: the proposal

• Example proposal (note that this does miss out some key elements that it should
include):

• 1. the CAPM (or alternative model such as the Market Model) being applied and
the Estimation and Event windows,
• 2. the ways in which returns are measured explained,
• 3. The measurement of the AR is explained within the context of the CAPM. The
CAR should also be explained here but it is not given in this example.
• 4. Test for statistical significance should be presented.

• Note that the key element missing from this proposal is that it does not examine the
question of how we measure the statistical significance of the impact of the event on
the share price. You will see more on this in the paper by MacKinlay.
 9

Event Study: the proposal

1. The CAPM and the Estimation and Event windows

ri ,t  rf ,t   i   i (rm ,t  rf ,t )   i ,t
 9

Event Study: the proposal

2. The returns used are continuously compounded returns that use the natural log
which is shown as ln

• Rt= ln( Pt/Pt-1)

 9

Event Study: the proposal

3. The Abnormal Returns (AR) are the residuals from the CAPM
(We can re-describe the equation in terms of the AR is we want to)

ri ,t  rf ,t   i   i (rm ,t  rf ,t )   i ,t

The residuals are

the Abnormal
Returns

• The Cumulative Abnormal Returns (CAR) are the sum of the AR at the point in
time being considered.
 9

Event Study: the proposal

4. Statistical Tests

• If the event had no impact on the share price we would expect the abnormal
return to be zero. So what we test is to see if the AR is statistically Significantly
different from zero.

• H0: AR = 0
H1: AR ≠ 0

• If the null is rejected and the alternative accepted we conclude that the event did
have a statistically significant impact on the share price.

• There are several different ways of doing this and some can be quite complex (See
paper by McKinley). However, we can do a simple t-test.
 9

Event Study: the proposal

• 4. Statistical Tests

• The t-statistic estimated is for the AR at a specific time (t) divided by the standard
deviation (S). The standard deviation is measured over the estimation window only
with M being the number of non-missing returns. Note that the degrees of freedom
parameter (2) shown in this case is for the market model.
 9

Event Study: An Example

• We will now have a look at the Excel file and how we undertake the event
study.

• There are only 3 data series that are needed: the share price, the ftse100 index
and the daily interest rates. We measure:

• 1. Daily returns of both the share and the market. These are the continually
compounded rate estimated as: Rt = ln(Pt/Pt-1),

• 2. Daily interest rate estimated as 90 day rate/365,

• 3. Excess Return (for both the share and the market). Estimated as daily
return – daily interest rate,
 9

Event Study

• 4. The CAPM
This is estimated over a period before the event. Often we want to see if there
is any impact before it is announced to identify if the news leaked into the
market before it was officially announced

In this example we use period t= -140 to -21 as our CAPM estimation

period,

• 5. The Abnormal Return and Cumulative Abnormal Returns

In this example we use t= -20 to + 20 as the event window. The AR is the

residual in the CAPM i.e. the difference between the actual return and the
that estimated by the model

The CAR is just the cumulative AR,

 9

Event Study

• 6. The Statistical Significance of the event

If the event did not affect the share price the AR and CAR should be zero.
We can therefore measure if the AR is statistically significantly different
from zero. There are a large number of different tests. We can use the t-test
for simplicity

• We will let you work this one out.