Residuals and ARIMA
Posted: Wed Sep 01, 2010 1:07 am
In what cases, if there is any, would one include a 'resid' variable as an explanatory variable in an ARIMA specification? I have an ARIMA (12,1,12), and if I include resid01 in the estimation, it fixes up a lot of things. Does this make sense, theoretically and practically?
Also, when conducting the Ramsey RESET with the resid01, the test looks fine. When I remove resid01 and conduct RESET again, p-values drop below .05. What is this showing me, that I have ommited significant variables?
Also, when conducting the Ramsey RESET with the resid01, the test looks fine. When I remove resid01 and conduct RESET again, p-values drop below .05. What is this showing me, that I have ommited significant variables?