Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: Event study
From
Nick Cox <[email protected]>
To
"[email protected]" <[email protected]>
Subject
Re: st: Event study
Date
Fri, 12 Apr 2013 14:05:31 +0100
Should be
egen benchmark = total(abnormalreturn / (day == -1)), by(country)
Nick
[email protected]
On 12 April 2013 13:55, Nick Cox <[email protected]> wrote:
> I doubt that "event study" is universal jargon. Most things could qualify.
>
> That aside, your example is not clear.
>
> There are at least two things you can calculate, (1) change since time
> datum and (2) change from day to day
>
> The first is (for example)
>
> egen benchmark = total(abnormalreturn / (day == -1)), by(country
> gen deltachange = abnormalreturn - benchmark
>
> and the second is
>
> bysort country (day) : gen change = abnormalreturn - abnormalreturn[_n-1]
>
> The variable -deltachange- you give seems to be one thing for some
> observations and another thing for other observations, so at least one
> of us is confused.
>
> Nick
> [email protected]
>
>
> On 12 April 2013 13:39, Künzel Sven <[email protected]> wrote:
>
>> I am doing an Event-Study, which caused some troubles. In order to detect
>> abnormal reactions between event day 0 and post-event day + 1 with the help
>> of the event study, I want to compute the change between the days t and t-1.
>> Just below you see a data sample.
>>
>> country abnormalreturn day DeltaChange
>> 1 10 -2
>> 1 11 -1 1
>> 1 12 0 1
>> 1 15 1 3
>> 1 16 2 4
>> 1 14 3 -2
>> 1 13 4 -1
>> 1 12 5 -1
>> 2 35 -2
>> 2 37 -1 2
>> 2 39 0 2
>> 2 32 1 -5
>> 2 35 2 3
>> 2 46 3 11
>> 2 34 4 -12
>> 2 37 5 +3
>>
>> Is this possible? Which code would be the right one?
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/