Re: st: RE: RE: RE: RE: Autocorrelation test for pooled cross section data

[Scott Merryman <scott.merryman@gmail.com>]

-xtregar, lbi-  report the Baltagi-Wu locally best invariant (LBI)
test statistic that rho = 0 and a modified
        version of the Bhargava, Franzini, Narendranathan (1982)
Durbin-Watson statistic be calculated and reported.


. webuse grunfeld,clear

. xtregar inve mval, fe lbi
(output omitted)

modified Bhargava et al. Durbin-Watson = .50577251
Baltagi-Wu LBI = .82334942

The BFN Durbin Watson is also easy to calculate:

. qui xtreg inve mval, fe

. predict double e,e

. gen double bfn = sum((e- l.e)^2)/sum(e^2)

. cl bfn in l, noobs

       bfn
  .5057725

The original BFN paper supplies upper and lower bounds for Durbin
Watson statistic.   For a large number of cross sections, if it less
than 2 then it would indicate positive serial correlation.


See:

Bhargava, A., Franzini, and W. Narendranathan. 1982. "Serial
correlation and fixed effects model." Review of
    Economic Studies 49, p. 533 - 549.


Scott

On Wed, Sep 29, 2010 at 10:37 AM, Elizabeth Dhuey
<elizabeth.dhuey@utoronto.ca> wrote:
> Due to the length of my time period, I have some concern regarding
> correlated errors over time. I have plotted the residuals and see some
> evidence of correlation over time.
>
> However, my attempt to find a formal test that produces a p-value is based
> on a request from a journal referee. My main issue is that I can't seem to
> find any formal test that is appropriate for pooled cross sectional data. My
> current thought is to calculate a durbin-watson statistic for each state and
> report the distribution of the statistic. However, I suspect that someone
> has already formalized a produce for this kind of data structure.
>
> Elizabeth
>
*
*   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/