Dear all,
I run an ECM model and these are my post-estimation results:
Durbin-Watson (autocorrelation): 2.1638
Breusch-Godfrey (autocorrelation): 0.4418
Breusch-Pagan (heteroscedasticity): 0.0001
ARCHLM (conditional heteroscedasticity): 0.0208
RAMSEY (omitted variables): 0.3565
Jarque-Bera (normality): 0.1119
How can heteroscedasticity be corrected for ? Again, if any of these tests are significant, does that invalidate the regression results?
Thanks!