CMStatistics 2019: Start Registration
View Submission - CMStatistics
B1041
Title: Test for detecting risk equivalent or risk neutral portfolios Authors:  Daniel Gaigall - FH Aachen University of Applied Sciences (Germany) [presenting]
Marc Ditzhaus - Otto-von-Guericke University Magdeburg (Germany)
Abstract: We are interested in testing the marginal homogeneity or the exchangeability of the joint distribution of two portfolios. Thereby, we deal with the underlying time series of the portfolios as functional data. The test under consideration is of Cram\'er-von-Mises type and is applicable for random variables in a general Hilbert space. We suggest a bootstrap procedure to obtain critical values. From the theory of $U$-statistics, properties of the test under the null hypothesis and under alternatives are derived. A direct application is the detection of risk equivalent or risk neutral portfolios. We analyze real financial time series to demonstrate how the approach works in practice.