Obtain market model residuals from eventstudy2

[Gabriele More]

Dear Stata Community!

I am using the command eventsudy2. I am trying to obtain the residuals from the market model regression over the (-1,0) window, but I can't see how to obtain them from the command. Do you have any suggestion?

Thank you

[Thomas Kaspereit]

Dear Gabriele,

In event studies in financial research, the residuals represent the abnormal returns, either during the event window (out-of-sample) or during the estimation window (in-sample).

Since eventstudy2 outputs abnormal returns, you can easily access them, e.g. via the arfile.dta

You can even redefine the event window using the evwlb and evwub options to make it overlap with the estimation window, and the ARs will provide you the residuals.

Best Thomas

[Gabriele More]

Thank you prof Thomas. In effect, my question could have been formulated better. I am trying to obtain the variances of the CARs. The market model generates a measire of variance for each abnormal return, and I need to extract the standard deviations of the market model residuals. Thank you, Gabriele

[Thomas Kaspereit]

Thank for clarifying. No, those variances do not constitute output of eventstudy2, nor can they be easily accessed. Of course, they are internally used to calculate test statistics. However, they are are not provided as outputs. This would be outside the intended scope of my command.

Best Thomas

[Gabriele More]

ok, thank you. Is there a way to obtain them? I actually need them as weights to a WLS regression and i don't know if it is possible and how to obtain them from Stata...

[Thomas Kaspereit]

It is possible, Gabriele, by diving deep into eventstudy2.ado's source code. You need advanced Stata programming skills to do so.

A more efficient way to achieve your goal is to use the Princeton code snippets for event studies. A google search will lead you there.

These snippets are also nicely described in

https://www.sciencedirect.com/scienc...19850121000328

Best Thomas

[Gabriele More]

ok, thank you! I will look into it. That's very useful and helps me a lot. Thank you

[Gabriele More]

alternatively, if I want to extract the standard deviation of the residuals from the market model during the estimation period, can I extract them from the eventstudy2 command? Or is it the same answer as before? Thank you