Spectral Volume Models: Universal High-Frequency Periodicities in Intraday Trading Activities
Researchers introduce spectral volume models using Fourier analysis to reveal persistent, high-frequency patterns in intraday trading volumes. These periodic bursts, found in both U.S. and Chinese stock markets, account for much of the variance in trading activity. The study links these patterns to algorithmic trading with scheduled instructions. Recognizing them can improve volume predictions and trading strategies like VWAP (volume-weighted average price), offering potential excess returns. The findings highlight algorithmic trading’s strong influence on market behavior.
What it examines
This paper introduces spectral volume models using Fourier analysis to detect and explain high-frequency periodic patterns in intraday trading volumes. The study aims to uncover universal periodicities in US and Chinese stock markets, linking them to algorithmic trading behaviors and improving predictions and trading execution.
What it concludes
The results show that identifying high-frequency periodicities enhances volume prediction and trading performance, offering insights into algorithmic trading and market efficiency. Applications include better trade execution, price discovery, and generating excess returns. Future research may explore more markets and refine models for broader financial use.
Evidence objects
Researchers unveil 'spectral volume models' using Fourier analysis to expose persistent, high-frequency trading patterns in both U.S. and Chinese stock markets, revealing these periodicities are universal and often hidden by noise.
key_findings bullet 1 · key_findings · validation V0
Surprisingly, these regular bursts of trading activity explain a significant share of trading volume variance, suggesting algorithmic trading with scheduled instructions is a dominant force shaping global market behavior.
key_findings bullet 2 · key_findings · validation V0
The studys novel methodology improves volume prediction and VWAP execution, offering fresh insights into market microstructure, though future research should test more diverse markets and consider regulatory impacts for broader generalizability.
key_findings bullet 3 · key_findings · validation V0
This paper uniquely applies Fourier analysis to intraday trading volumes, revealing persistent, universal high-frequency periodicities across major markets. By linking these patterns to algorithmic trading and demonstrating improved volume prediction and VWAP execution, it offers a novel, impactful approach, making it compelling and original for market microstructure and high-frequency trading research.
key_findings bullet 4 · key_findings · validation V0
Raw abstract and provenance
We develop spectral volume models to systematically estimate, explain, and exploit the high-frequency periodicity in intraday trading activities using Fourier analysis. The framework consistently recovers periodicities at specific frequencies in three steps, despite their low signal-to-noise ratios. This reveals persistent and universal high-frequency periodicities in the United States and Chinese stock markets in recent years, and the dominant frequencies explain a significant fraction of the total variance of intraday volumes. We provide evidence that this phenomenon likely reflects the behaviors of trading algorithms with repeated and regular trading instructions. Finally, we demonstrate that uncovering such high-frequency periodicities improves intraday volume predictions and volume weighted average price execution qualities, yields insights for price informativeness of algorithmic trading, and generates excess returns. This paper was accepted by William Lin Cong, finance. Funding: This work was supported by the National Key Research and Development Program of China [Grant 2022YFA1007900], the National Natural Science Foundation of China [Grants 12271013 and 72342004], and the Peking University’s Fundamental Research Funds for the Central Universities. Supplemental Material: The online appendix and data files are available at https://doi.org/10.1287/mnsc.2024.06215 .
Source row: 1807 · abstract type: unknown