andrews quandt coef stability procedure
Posted: Wed May 16, 2012 10:55 pm
I am using eviews 7.2 august 25 2011 build. While implementing a training for 20 people in Africa, we realised that Andrews Quandt reports appear to be wrong as two pairs of probabilities are reported identical (see attached screen shot). This was true on 20 EViews 7.2 softwares, but not on the 21st (they all were 7.2s). Incidentally shows that Congolese can be pretty sharp!