Exploring Robust Methods for Testing Equality of Variances

Abstract

Assessment and test of equality of volatilities of stocks with similar returns is important for decision makers in their portfolio selection. Volatility estimation is equally important for pricing stocks and derivative securities. In this study, three robust tests for homogeneity of variance are considered in the selection of stocks under the conditions of high kurtosis and large skewness. Investigations involved the effectiveness of the Ballett’s test, robust methods of Levene's Test and Fligner-Killeen test in the test of homogeneity of variance of stocks. Simulated and insurance stock data were employed and empirical results findings show that the three test statistics are equally efficient in testing homogeneity of variance even as the datasets deviates from normality.