|
Abstract:
|
In this paper, we will discuss methods for testing the equality of regression coefficients across two samples when the data is multicollinear. Such tests are well-developed using OLS methods. However, there is no method for conducting such tests when OLS estimation is not an appropriate method. For example, when multicollinearity exists in the data, OLS estimation techniques have been found to produce estimates with large variances; yielding wide confidence intervals, unreliable tests, and incorrect signs. Thus, OLS techniques will not suffice in conducting such tests when the data is multicollinear. Therefore, there is a need for a method to test the equality of regression coefficients between two or more regression equations when the data is multicollinear. In this paper, we will present two methods for conducting such a test using ridge regression estimators and a stacked estimator.
|
