Lo-MacKinlay Variance-Ratio Tests
Posted: Thu Apr 30, 2009 11:16 am
We had a posting elsewhere on the forum asking for something in this area, so I cooked up something that some may find to be of use.
Attached is a subroutine that I cooked up to perform the simplest version of Lo-MacKinlay variance ratio tests on the series in a group. The limitations are described in the code, but in particular it assumes homoskedasticity, doesn't use bias corrected variances, and doesn't perform any joint testing. To use it, you'll need to include the subroutine at the top of any program file that calls this.
It's undergone a bit of testing (but not much) and matches a variety of known results, but you should use at your own risk.
Attached is a subroutine that I cooked up to perform the simplest version of Lo-MacKinlay variance ratio tests on the series in a group. The limitations are described in the code, but in particular it assumes homoskedasticity, doesn't use bias corrected variances, and doesn't perform any joint testing. To use it, you'll need to include the subroutine at the top of any program file that calls this.
It's undergone a bit of testing (but not much) and matches a variety of known results, but you should use at your own risk.