Testing the stationarity of sign coincidence in market network
The problem of stationarity of sign coincidence of returns is considered. Statinarity of sign coincidence of a pair of stocks is tested by two sample Kolmogorov-Smirnov and Chi-Square tests. Multiple comparison pocedures, such as Bonferroni and Holm procedures, are employed to test stationarity of sign coincidence in market network and to control the family-wise error rate (FWER). The method is validated for testing stationarity of stock's prices and returns. It is shown that the hypotheesis of stationarity is rejected for prices and it is not rejected for returns and their sign coincidence on some significance level.