Plausability of my Cumulative Abnormal Returns

#1
Dear all,

I have a question regarding the plausability of my CAR(2;366)=-40%. My tutor (in accounting, not statistics) claims that it should lie in the interval of +-10% to be reasonable.

He has been wrong before in matters of statistics so I registered an account here to ask the perhaps more knowledgeable people.

Thanks in advance for your replies.
best regards,
Michael
 

Miner

TS Contributor
#2
+/- 10% sounds like an accounting rule of thumb, or based on their past experience.

Statistical significance is a combination of both effect size and sample size, not one alone.
 
#3
Indeed it does sound that way.
It is significant at the 5% confidence level (based on a two-sided test) with a sample of roughly 500 observations.
His argument is basically that our results are not trustworthy and thus our calculations are worthless/wrong. I do not think that our calculations are wrong but can not state how a yearly CAR of this size still can occur in an event study.
 

Miner

TS Contributor
#4
In the absence of unusual external influence, he might be correct (purely from a personal experience / historical perspective). However, the market frequently does experience external influences. And a company can find themselves in dire straights. For example, I would not be surprised at your results if it were done for Blackberry. Are you analyzing a specific company, a portfolio or ...?
 
#5
I see. It is a portfolio of companies in different industries. The event is the announcement of the company's largest (based on deal value) merger or acquisition (at day zero). I do not know if this announcement characterizes as unusual external influence or not but this large merger/acquisition would have a large impact on the company and its share price.
 

Miner

TS Contributor
#6
I would call that an unusual external event. That announcement will set certain market expectations on the return. Often a company cannot live up to those expectations.