CMStatistics 2022: Start Registration
View Submission - CMStatistics
Title: Robust and efficient Breusch-Pagan test: A beta-score LM test for heteroscedasticity in linear regression models Authors:  Nirian Martin - Complutense University of Madrid (Spain) [presenting]
Abstract: In Econometrics, the Breusch-Pagan test-statistic has become an iconic application of the Lagrange multipliers (LM) test, also recognized as the Raos scores test for composite null hypotheses. We shall introduce beta-score LM tests for heteroscedasticity in linear regression models, for which the degree of robustness and efficiency is regulated through a non-negative tuning parameter, being $\beta=0$ the classical Breusch-Pagan test-statistic, the most efficient one under absence of outliers. A very elegant expression is obtained, with a similar interpretation to the one for the classical case. The test statistic is constructed by extending the methodology from identically distributed to nonidentically distributed individuals, for composite null hypotheses. Detailed theoretical justifications for robustness and efficiency properties are given. A simulation study illustrates the finite-sample behaviour of several Breusch-Pagan beta-score LM test statistics.