Event Study analysis

[Fran Zayn]

Hello everyone,

thank you again for helping on my other problem: https://www.statalist.org/forums/for...reting-results

Now I have to do additonally an Event-Study about the same issue. I used the princeton method, however I have a few questions if I have done it right and I hope you can help me.

So just a quick update: I have daily quotes of 12 countries and I want to test if the quotes differ after the event than before.

So my first question is how do I calculate the returns on Stata since I want to use this formula as the return:

return = spread_i,t - spread_i,t-1

my dataset looks like this:

Another question I have which is more important. How do I calculate the expected return?
I was thinking about to apply a constant mean model, which looks like this

So the expected return results from the average returns of 250 trading days before my event window of 30 days
However, I want this on a rolling window. (For the next day the, 250 days stay constant but then I estimate the mean between -30 and -279 and so on)

Do you know how I can program this on stata?
I already did it on Excel but I am confused how to do it on stata.

Thank you very much you guys helped me already alot!

[Fran Zayn]

Hello everyone,

So my first question is how do I calculate the returns on Stata since I want to use this formula as the return:

return = spread_i,t - spread_i,t-1

So I tried this code but it only works for country 1

egen id=group(company_id) forvalues i=1(1)12 {
gen return = spread[_n] - spread[_n-1] if id==`i'
}

[Fran Zayn]

allright i programmed it on excel.

I have now some comprehension questions.

What is this estimation window done by princeton? What does that mean?

gen estimation_window=1 if dif<-30 & dif>=-60

do i have to put in here the -31? and -280? of my expected return respectively?



and also what does this term express?

reg ret market_return if id==`i' & estimation_window==1

for market return I entered my expected return as you can see in the formula above.
Is that the right apporach?

Thank you for your help.

[Phil Bromiley]

You didn't get a quick answer. You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex.

You need to explain what you want to do - only people in your little research area know what you mean by the princeton method for example. You also need to work your problem a lot more before coming to us with a laundry list of questions.