EViews.com

Hello,

I apologise in advance if this has a really simple solution that I've overlooked (although I hope it does!).

I have estimated a linear regression and am trying to run a rolling chow test in order to identify any break points in my sample period.

Ideally I'd like a graph showing the observation number on the x-axis and the Chow F statistic (or corresponding p-value) on the y axis.

I couldn't find any specific function to do this, and have spent ages trying to write a program that will do it for me, but to no avail.

Any help would be VERY gratefully received. :D

Should be relatively straight forward. Just a for loop that loops over the dates/observations you want to perform the chow test on, then freeze each chow test into a table, and extract the test statistics.

Of course, an easier way would seem to be to use the Quandt Andrews test instead, since that is just a Chow test at every point.