Monte-Carlo comparisons of the power of some tests of heteroscedasticity
White, Kenneth J.
Master of Arts
This thesis discusses the class of tests of heteroscedasticity developed by J. Szroeter. Specifically, for models with non-stochastic regressors, this thesis discusses some exact tests within the above class of tests utilizing existing tables of distributions of the Von Neumann ratio and of the Durbin-Watson bounding ratio. The power of these tests are then compared to the power of a Goldfeld-Quandt type test to determine the efficiency of Szroeter's tests. There are two ways available to us for calculating the power of a test. One way is to use the Imhof method and the other is the Monte-Carlo method. This thesis uses the Monte-Carlo method to calculate the power of the test as this method is computationally uncomplicated compared to the Imhof method.