ON THE CHOW TEST WHEN THE DEGREES OF FREEDOM ARE INADEQUATE

Date

1978

Author

Erlat, Halûk

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

22
views

47
downloads

Cite This

We consider the problem of testing for structural change by means of dummy variables (in the context of two subsamples), where one of the subsamples does not contain sufficient observations for the regression function to be fitted. We start out by proving an assertion by Valentine (1971) which states, within the context of the full set of coefficients , that if any k+l-T2 (where k+1 is the number of explanatory variables plus the intercept term and T i s the size of the second subsample) dummy variables are deleted from the initial unrestricted form and the joint significance of the remaining T^ dummy variables are tested, then the test statistic utilised to test for structural change will be the same as the one derived by Chow (1960) . Subsequently, we extend this result to (a) testing for structural change in a subset of the coefficients with the remaining subset apriorily assumed to be (i) exhibiting, (ii) not exhibiting structural change; (b) to testing for structural change when ther are variables which are specific to each subsample.

URI

https://hdl.handle.net/11511/111018

Collections

Department of Economics, Article