Time series Break tests

[jhh]

hello I am using eviews 8 to carry out Bai and Perron Structural Break tests. I have a very simple question.
When I perform the L+1 vs L breaks i get three dates (1977 1987 1998) in sequential break dates and three (1977 1987 2003) in the repartition break dates.
One date is different in these. My question is that should i consider all four dates (1977 1987 1998 2003) in my model as breaks Or just the ones with sequential procedure??

[trubador]

Merging the results in this way is not a good idea from an econometric perspective. I think you should now consider the output from the practical point of view. You can search for more information regarding the years 1998 and 2003. If you can label or identify a structural change around these periods, then you are fine. And remember that these are point estimates and are subject to uncertainty as well.

[guiri]

please how can i conduct the bai and perron's test on the mean and variance equation for garch models .It works for the mean Eq but for the variance eq. i have an error msg . Thanks