MULTIPLE TESTING OF SIGN SYMMETRY FOR STOCK RETURN DISTRIBUTIONS
Multiple statistical procedure for testing elliptical model for stock return's distribution is proposed. Sign symmetry conditions are chosen as individual hypotheses for multiple testing. Distribution free uniformly most powerful tests of Neyman's structure are constructed for individual hypotheses testing. Associated stepwise multiple testing procedure is applied for the real market data. Numerical experiments shows that hypothesis of elliptical model is rejected. At the same time it is observed that the graph of rejected individual hypotheses has unexpected structure. Namely this graph is sparse and has a few hubs of high degree. Removing this hubs leads to non-rejection of hypothesis of elliptical model.