Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: Event study 300 companies using a SUR to estimate standard market models

From   Doryon Zalko <>
Subject   st: Event study 300 companies using a SUR to estimate standard market models
Date   Sat, 14 Aug 2010 00:24:24 +1000


I am conducting an event study on each stock in the ASX 300 using the standard market model. I am using  an estimation period of 100 days, and a period before the event window of 40 days.

I need to use a seemingly unrelated regression for estimating the 300 standard market models. But I am having trouble programming Stata to run SUR during the estimation period and then predict the abnormal return over the event window all at once.

Is this even possible? If so how would I go about it?

Prior to using SUR I was estimating the standard market model using a doloop:

gen predicted_return=.
egen id=group(gvkey) 
 forvalues i=1(1)300 { 
        regress stockreturn ri if id==`i' & EP1==1 
        predict p if id==`i'
        replace predicted_return = p if id==`i' & event1==1 
        drop p
sort id event1 datadate
gen abnormal_return=stockreturn-predicted_return if event1==1
sort id event1 datadate
by id: egen cumulative_abnormal_return= sum(abnormal_return)
sort id datadate
by id: egen ar_sd= sd(abnormal_return)
gen testew =(1/sqrt(5)) * ( cumulative_abnormal_return /ar_sd)
drop predicted_return abnormal_return 

Where, EP1 and event1 variables taking on the value of 0 or 1. For EP1, 1 for estimation period, 0 otherwise.
For event1, 1 for event window, 0 otherwise.

Is it possible to adapt this into the seemingly unrelated regression code? Hints would be appreciated.

Thanks for your help

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index