Re: st: Regression with different firms

[Nick Cox <njcoxstata@gmail.com>]

We can't see anything that you don't tell us. "seems to work" and
"seems as if" are no help. Please give details in order not to waste
your own time by asking questions that no one can answer.

Nick

On Thu, Aug 9, 2012 at 7:44 PM, felix kreppel <felix.kreppel@gmx.de> wrote:
> I tried the following commmand:
>
> reshape long firm std, i(date) j(firm)
>
> This seems to work. My dataset, however, is quite large (1000 firms & 1000 volatilites). It seems as if Stata cannot perform the conversion. Am I doing something wrong, or is there another way to do this?
>
> The regression command I would end up with after reshaping would be
>
> regress firm std market_return, right?
>
> Thanks.
> -------- Original-Nachricht --------
>> Datum: Thu, 9 Aug 2012 14:34:09 -0400
>> Von: Fernando Rios Avila <f.rios.a@gmail.com>
>> An: statalist@hsphsun2.harvard.edu
>> Betreff: Re: st: Regression with different firms
>
>> Exactly. You might end up with a data set that looks like:
>> Year   firm_id  return  std  market_return
>> 1985     1
>> 1985     2
>> 1985     3
>> 1986     1
>> just check the reshape examples from wide to long.
>> and if you tried already something, send the commands you have tried,
>> so its possible to tell you where is the syntaxis mistake.
>> Fernando
>> On Thu, Aug 9, 2012 at 2:21 PM, felix kreppel <felix.kreppel@gmx.de>
>> wrote:
>> > Hello Fernando,
>> >
>> > Thank you for your quick response.
>> >
>> > I thought of this. Do you mean, that I should reshape my dataset in the
>> following way?
>> >
>> >
>> >       year   return std market_return
>> > firm1 1985
>> >       1986
>> >       1987
>> > firm2 1985
>> >       1986
>> >       1987
>> >
>> > and then run a simple linear regression?
>> >
>> > I tried different reshape commands-, but none of them seems to work..
>> >
>> > -------- Original-Nachricht --------
>> >> Datum: Thu, 9 Aug 2012 14:01:33 -0400
>> >> Von: Fernando Rios Avila <f.rios.a@gmail.com>
>> >> An: statalist@hsphsun2.harvard.edu
>> >> Betreff: Re: st: Regression with different firms
>> >
>> >> Felix,
>> >> Look at -help reshape-
>> >> HTH
>> >> Fernando
>> >> On Thu, Aug 9, 2012 at 1:58 PM, felix kreppel <felix.kreppel@gmx.de>
>> >> wrote:
>> >> >
>> >> > Hello,
>> >> >
>> >> > I have regression problem:
>> >> >
>> >> > I have a dataset consisting of yearly firm returns (for firm1-firm3)
>> and
>> >> the volatility of each firm in that month(std1-std3) and a data on a
>> >> monthly market return (market_return) over a specific timepriod.
>> >> >
>> >> > No I want to estimate the following regression over the whole time
>> >> period:
>> >> >
>> >> > return_firm_i = a*std_i + b* market_return
>> >> >
>> >> > Its not a problem to do this with only one firm:
>> >> >
>> >> > regress firm1 std1 market_return, robust
>> >> >
>> >> > However, I want to do this for the whole dataset and estimate single
>> >> coefficients.
>> >> >
>> >> > Could somebody help me?
>> >> >
>> >> >
>> >> > My dataset looks like this:
>> >> >
>> >> > date    firm1 firm2 firm3 std1 std2 std3 market_return
>> >> > 1985    0.03   0.04  0.05  0.3  0.4  0.5    0.6
>> >> > 1986   ..
>> >> > 1987    ...
>> >> >
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/