Lec 7 - Mesuring abnormal return in short term Flashcards
What do the following factors mean?
rit = total return in test period
E(rit) = expected return
ARit = abnormal return on a day
What is Rit made up of?
Return due to impact of the event of interest
Return due to common factors (expected return)
What is E(rit)?
Return if the event hadn’t happened (return due to common factors)
What is AR(it)?
Return due to the event in question
What aspect causes models to differ?
The estimation of expected return
What causes the measurement of AR to differ?
Horizon of test period (long vs short term)
The model used to estimate E(r)
In the market model, what is expected return?
Alpha hat + beta hat - which are estimated in the estimation period that we specify
What is the cumulative abnormal return?
When you have multiple days in the test period (I.e 2 days before and after the event day)
When you have the single day abnormal returns, you add them all up and get the cumulative abnormal returns
You get “CAR” which is all of it together
What if there is a large sample?
Use Average Abnormal Return (AARt) and Cumulative Average Abnormal Return (CAAR)
Find average across all firms for a single day
AArt formula:
CAARt formula:
What is CAPM (formula included)?
Capital Asset Pricing model
- Similar to market model, with its difference being CAPM works with observed returns without any adjustments
What is Fama French 3 Factor Model?
Under this - expected return is the risk free rate + the alpha estimates + the beta estimates * excessive market return + factor returns * the coefficients (gamma estimates * value factor returns) (delta coefficient * size factor returns) which gives = the expected returns
Then expected return - actual return = abnormal return under fama french 3 factor model
How can the T-statistic for CAR, AAR and CAAR be further calculated?
By building on the t-statistic for the single-day firm abnormal return
The t-statistic in principle is calculated as the abnormal return scaled by its standard error
What is standard error in the context of t-statistics?
The standard deviation of the abnormal return