Recognizing Intra-day Patterns of Stock Market Activity

Abstract

The aim of this comparative research is to recognize and assess intra-day seasonality of investors activity on a stock market using high-frequency data. Three indicators of intra-day investors activity based on different market characteristics are utilized: (1) hourly aggregated trading volume for a stock, (2) hourly percentage relative spread based on the highest and lowest prices of a stock, and (3) the modified version of the Roll's estimator for hourly effective spread based on the logarithmic ultra-short rates of return of a stock. The time-stamped data derived at five-minute intervals from the Warsaw Stock Exchange (WSE) is used. The data set covers the recent period from December 1, 2020 to April 30, 2021. The findings of computational experiments for real-data from the WSE show that visible U-shaped, J-shaped or reverse J-shaped hourly patterns dominate for the majority of equities and investigated indicators of intra-day market activity. What is important, the empirical results are homogenous. Moreover, the robustness tests and statistical analyses based on the rolling-window procedure confirm that results are robust to the choice of the analyzed sub-period. The findings are crucial from a practitioner's point of view as an empirical assessment and visualization of intra-day activity patterns can help investors to state how various stock market characteristics vary within a session. Therefore, it may be a useful, both formal and intuitive tool supporting decision-making processes.