Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • How to compute the Chow Test with multiple variables to test for structural breaks

    Hi, I am running the following regression:
    Code:
    reg Y1 X1 X2 X3
    There are 100 observations, and I test for a structural break at observation 50.
    How would I use the Chow Test with interaction dummies to see whether I reject or not the null of a structural break?
    Last edited by Adrian Cernescu; 05 Mar 2021, 14:47.

  • #2
    Have a look at this thread, and in particular my post #4.
    https://www.statalist.org/forums/for...lity-over-time

    Basically you create a dummy

    Code:
    gen break = observation>=50
    and then you interact this dummy with all your regressors. Then you test that the dummy itself and all its interactions are significant. Like this (pseudo code follows):

    Code:
    reg Y1 X1 X2 X3 break break*X1 break*X2 break*X3
    
    test break break*X1 break*X2 break*X3

    Comment

    Working...
    X