Re: st: Remedy for serial correlation in Panel Data

[David Greenberg <dg4@nyu.edu>]

The xtregar command takes first-order serial correlation into account
in the estimation process by doing a Cochrane-Orcutt procedure on the
data. You should keep in mind that you have a very small data set for
analysis. David Greenberg, Sociology Department, NYU

On Mon, Apr 9, 2012 at 7:38 PM, San K <devank@gmail.com> wrote:
> This is my understanding with autocorrelation. It occurs because
> 1. Specification error. You are missing some important independent variable.
> 2. Seasonality (similar to point 1)
> 3. Wrong functional form
> 4. Data manipulation
>
> I also have trouble with autocorrelation but couldn't nail it.
>
>
>
> On Sat, Apr 7, 2012 at 3:54 AM, D. Demetriou <dd358@cam.ac.uk> wrote:
>> Hi everyone,
>>
>> Just a quick question: what is the best way to account for Serial
>> correlation in Panel Data (T=10, N=9)?
>>
>> Some (e.g. Cameron&Triverdi) say that autocorrelation is more a problem in
>> long macro panels=> I shouldn't worry according to them
>>
>> Others say that it is certainly a problem, irrespective of circumstances,
>> since it violates one of the basic assumtpions of panel data and gives
>> 'wrong' standard errors => I should worry very much according to them
>>
>> So?? Any ideas on this?
>>
>> PS: I am using 'xtserial' test (Wooldridge test for autocorrelation in Panel
>> Data) to detect whether or not serial correlation is a problem.
>>
>> Yours,
>>
>> DD
>>
>> *
>> *   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/
>
> *
> *   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/

*
*   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/