Serial correlation LM Test - Help!
Posted: Thu Sep 25, 2014 6:38 am
by hayleyvictoria
When test my ols regression for autocorrelation with the Breusch-Godfrey LM test, I obtain the correct coefficients for c, x and lagged ressiduals, correct F statistic, Prob, and chi square result however my standard errors (t and prob), rsquared, adjusted rsquared, se of regression and mean dep var are all very slightly out on what they should be. anyone have any ideas what could be causing this please? Its eviews 7. Thanks.
Re: Serial correlation LM Test - Help!
Posted: Thu Sep 25, 2014 7:11 am
by startz
Possibly the number of observations has changed to allow for lags?
Re: Serial correlation LM Test - Help!
Posted: Thu Sep 25, 2014 10:24 am
by EViews Glenn
To what are you comparing your results? How do you know what the results "should be"?
Note that Startz's conjecture, while reasonable, isn't correct. As described in the manual (and at the top of the test output):
Following the suggestion by Davidson and MacKinnon (1993), EViews sets any presample values of the residuals to 0. This approach does not affect the asymptotic distribution of the statistic, and Davidson and MacKinnon argue that doing so provides a test statistic which has better finite sample properties than an approach which drops the initial observations.