Re: st:AR(1) error notice: sample may not include multiple panels

[Clive Nicholas <clivelists@googlemail.com>]

Amanda Fu wrote:

> Sorry for bothering you with  a question about a simple AR(1)
> estimation. I searched related discussion in the archives , but still
> cannot figure out why the error message comes out. Any suggestions
> will be helpful.
>
> I want to estimate a univarate AR(1) model (without any controls). The
> data set is a panel data,  including 1000 observations for 5 years
> (200 variables. The one I am interested is enroll taking values 0,1) .
> . tsset id years
> . arima enrollment,ar(1)
> "sample may not include multiple panels"
>
> I must have missed something here. Can someone give me some hints?
> In addition, instead of using ARIMA, may I know if there is any easy
> alternative to do the AR(1) ?

ARIMA models are designed for _single_ time series, not panel data.
Since you're looking to specify AR(1) effects, that also rules out
fixed- and random-effect models under -xtreg-, since parameter
estimates are almost always biased and inconsistent in the presence of
AR(1). Your only real alternative is investigate instrumental variable
models, particularly IV-2SLS (-h xtivreg-), assuming you can find
suitable variables to 'instrument' with the lagged variable. That's
where the fun starts, although I use the word 'fun' advisedly.

-- 
Clive Nicholas

[Please DO NOT mail me personally here, but at
<clivenicholas@hotmail.com>. Please respond to contributions I make in
a list thread here. Thanks!]

"My colleagues in the social sciences talk a great deal about
methodology. I prefer to call it style." -- Freeman J. Dyson

*
*   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/