AI-Induced Market Thinning: An Empirical Analysis of Trading Participation Using Volume, Trade Frequency, and Activity Metrics

Abstract

Artificial Intelligence (AI) and algorithmic trading platforms have revolutionized today's financial markets through their growing levels of automation, predictive capacity, and speed. AI-driven trading helps the market run smoothly in normal times, but when these trades are coming from the same predictive signals, it can have unintended consequences of market participation and, in the short term, a liquidity contraction. This paper explores the issue of AI market thinning by conducting an empirical study on trading participation based on different metrics such as volume, frequency of trades, and active trading interval. It's a research that analyzes how market participation changes during periods of high activity, with AI signals, by examining the price movements, trading volume, transaction frequency, and activity indicators of historical intraday market data. Model-based threshold conditions, based on predictive trading indicators and probability-based classification techniques, are used to identify AI-generated buy and sell signals. The study compares periods of normal market activity with periods of strong market activity as signaled by artificial intelligence, and finds that there is no evidence of a depressing effect on market activity from synchronized algorithmic trading. Market thinning can be quantified by trading volume, number of transactions and active trading intervals. Initial estimates suggest there could be substantial contraction of participation, with fewer trades executed and shallower liquidity levels, due to concentrated AI-driven trading signals. The study also analyzes these results in the context of the market microstructure theory, which points to a higher degree of order flow diversity and a lower trading participation as evidence of thinner markets and greater liquidity fragility. This study is meaningful because it combines the advantages of AI signal analysis with empirical market participation data, providing insights into the impact of automated trading systems on financial market liquidity and stability. The results could help regulators, institutional investors and market analysts understand the risks of participation in a synchronized environment with AI-driven trading and enhance the tools they have to monitor liquidity for resilience in increasingly automated financial markets.