Some Generalized Dependence Measures for Random Variables with Finite First Moments: a Simulation Study

Here we introduce some new linear dependence measures, namely the generalized covariation coefficient, generalized symmetric covariation coefficient and the generalized sign symmetric covariation coefficient. These measures can be applied for random variables which fulfill a certain linearity property and have finite first moments. Some basic mathematical properties of these measures are discussed in detail in pH and [5], where in particular they contain the classical Pearson correlation coefficient as the special case. We propose the estimator of the measures, and also doing simulation study for the S alpha S data. Application of these measures for finance modeling can be found in e.g. [6].