Analysis of the intensity of trading on the Ukrainian stock market

Abstract

In this research, we present a new approach for analysis of the intensity of trading on the Ukrainian stock market. In order to model the operations throughout the day, the non-homogeneous Poisson process with some unknown intensity function is used. The proposed approach is tested on the high frequency data concerning the trading operations associated with UX-index assets, with the unknown rate function estimated by kernel-based method with Epanechnikov kernel. From the analysis, it can be clearly seen that "the lunch time effect" is noticeable in the intensity profile, as well as the "day-of-the-week" effect, i.e. trading activity on the market depends both on the time of the day and day of the week with high statistical significance. This yields that stock price models with day-based switching correspond to reality better in comparison to standard models.