Introduction Benchmarks Abnormal Return Cumulative Abnormal Return Average Abnormal Return CAAR Event Study Christopher Ting Christopher Ting http://www.mysmu.edu/faculty/christophert/ k: [email protected]T: 6828 0364 : LKCSB 5036 March 29, 2017 Christopher Ting QF 604 Week 2 March 29, 2017 1/23
23
Embed
Event Study - mysmu.edu · Purpose of Event Study XEvent studies allow you to estimate how asset prices, i.e., prices of stocks, FX, bonds, or CDS, react to announcements of economic
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
X Event studies allow you to estimate how asset prices, i.e., prices ofstocks, FX, bonds, or CDS, react to announcements of economicevents that include new information relevant for the value of theunderlying assets.
X Is market efficient in reflecting the information in an event?
X If strong-form market efficiency hypothesis is true, then anysignificant return deviations before the time of public announcementmay provide evidence of insider information leakage.
X Does the event have positive, negative, or neutral impact on returns?
Christopher Ting QF 604 Week 2 March 29, 2017 4/23
X In general, an event study is a systematic examination of theaverage impact of a certain event on the price of a certain type of(corporate) asset.
X It is very important to exclude all events announced jointly withanother piece of new, price relevant information.
X You compare the asset price that occurred as a result of theannouncement of the event (the realized price or return) with ahypothetical asset price that would have occurred if no event had beenannounced (the expected price or return).
X The difference between the realized and the expected asset price,called the abnormal return, can be attributed to the event and betested for statistical significance. This way, a reliable conclusion aboutthe price impact of specified events can be drawn.
Christopher Ting QF 604 Week 2 March 29, 2017 5/23
X Significant information in the event announcement is taken to beunanticipated news that causes the market to re-evaluate the stock’s(a) expected future earnings (thus also dividends)(b) risk-adjusted discount rate
XWith significant information impact, the expected value of ARiτ isnon-zero, conditional on market condition up to and including those atτ .
X Average abnormal returns over the event window would then besignificantly different from zero.
X Thus for the first 3 benchmarks, the null hypothesis is, for τ = −10 toτ = 10,
E(ARiτ
∣∣rmτ) = 0
Christopher Ting QF 604 Week 2 March 29, 2017 12/23
� The tests are based on the null that the return process’ mean leveland variance remain constant.
� Rejection of the null could suggest that either the mean changes orthe volatility changes (or both change) due to the eventannouncement.
� If we want to test only whether there is any mean level change whilenot fixing volatility at the level of the estimation period, then we can usethe sample variance of AAR during the event window, i.e.
σ2(AAR) =1
20
21∑τ=1
(AARτ − AARτ )
2
to construct the approximate t20 test statistic
AARτ
σ(AAR)
Christopher Ting QF 604 Week 2 March 29, 2017 21/23