Page 1 of 1

Anual estimation of weekly data

Posted: Thu Apr 10, 2014 10:38 am
by Kouvas
Hello there

I have a large dataset of firms with weekly data on stock prices. I want to estimate a yearly regression for each firm. Is there any code to do that (i.e. using the smpl command)? Thanks in advance!

Re: Anual estimation of weekly data

Posted: Thu Apr 10, 2014 11:38 am
by EViews Gareth
Could you be a bit more explicit as to what you're trying to do?

Re: Anual estimation of weekly data

Posted: Tue Apr 15, 2014 5:37 am
by Kouvas
I am trying to estimate yearly regressions using weekly data. Therefore, in each year I will have a large number of weeks (the number of weeks depends on the year). I am using the smpl command to update the sample for each year (for example year 2007 should ran using the command smpl !year !year where !year=2007). Is there an alternative way to do that?

Re: Anual estimation of weekly data

Posted: Tue Apr 15, 2014 9:25 am
by EViews Gareth
Not really, you'll have to use samples.

Re: Anual estimation of weekly data

Posted: Tue Apr 15, 2014 9:51 am
by Kouvas
Thanks for the prompt answer. Therefore I have to create samples for example using the command sample mysample1990 1990 1990 should create a sample that when used in weekly data includes all the weeks in 1990?

Re: Anual estimation of weekly data

Posted: Tue Apr 15, 2014 9:56 am
by EViews Gareth
Yes

Re: Anual estimation of weekly data

Posted: Tue Apr 15, 2014 10:03 am
by Kouvas
This provides a sample that includes the last week of the previous year. However if I use the comand sample mysample1999 1999+1 1999 then it creates the right sample. Is this something I should expect?

Re: Anual estimation of weekly data

Posted: Tue Apr 15, 2014 10:35 am
by EViews Gareth
EViews views a week as being in a year if any day of that week was in that year.