AI-Driven Retail Trading and the Institutional Complementarity Trap: Why China's Market Reform Requires Simultaneous, Not Delayed, Change
This paper employs a five-population asymmetric evolutionary game to demonstrate that China's stock market modernization requires simultaneous, rather than delayed, reform of T+1 settlement, price limits, and short-selling constraints, as AI-driven retail trading can inadvertently lock the market into a transient but persistent instability that only a comprehensive, three-gate policy intervention can overcome.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
In the world of finance, markets are often viewed as machines that naturally improve over time. The logic suggests that as technology advances, bringing smarter tools to everyday investors, the system should become more efficient, fair, and modern. This belief rests on the idea that better tools will force old, rigid rules to adapt or disappear. However, a new study challenges this comforting narrative by looking at how different parts of a system interact. The researchers focus on a concept called institutional complementarity, which simply means that certain rules work well together but fail miserably when separated. Imagine a three-legged stool: if you remove one leg, the stool doesn't just stand on two; it collapses. The study asks what happens when a powerful new force, like artificial intelligence, is introduced into a system where the legs are already locked in a specific, rigid arrangement. The question is not just about technology, but about whether adding a new tool to an old, broken structure actually fixes the problem or just makes the collapse more complicated.
The researchers behind this work, based at HSE University and Ural Federal University, built a complex computer simulation to test this idea using China's stock market as a case study. They were particularly interested in why the massive adoption of AI trading tools by millions of retail investors has not led to the modernization of the market's core rules. Instead of a smooth transition to a mature, global-style market, the system seemed stuck in a hybrid state where high-tech tools operate within a framework of strict, outdated regulations. To understand this, the team broke the market down into five distinct groups of players: everyday investors, large financial institutions, the government rules themselves, foreign investors, and the mechanisms for betting against the market. They then watched how these groups interacted over time, simulating thousands of days of trading to see where the system would naturally settle.
The simulation revealed a surprising and somewhat unsettling dynamic. When the researchers started the model with a mix of conditions, the system quickly fell into a trap. Within a short period, equivalent to about six months of trading, the market appeared to stabilize in a specific configuration: high levels of AI trading, strict settlement rules that prevent same-day selling, and limits on how much prices can move in a day. This state looked stable to anyone watching the short-term data. However, when the simulation was allowed to run for a much longer period, roughly six years, this "stable" state began to crumble. It turned out to be a temporary illusion. The system slowly drifted away from this high-tech trap and settled into a different, more traditional state where the AI tools were still used, but the market remained dominated by older, state-controlled institutions and conservative rules. The technology had not modernized the market; it had merely prolonged the life of a temporary, unstable situation before the system reverted to its traditional roots.
The study explicitly rules out the idea that fixing just one part of the system will work. The researchers tested various scenarios where they removed a single constraint, such as allowing same-day trading or lifting limits on short-selling. In almost every case, the system failed to reach a modern, efficient state. Removing one rule while leaving the others in place simply caused the system to reconfigure itself around the remaining restrictions, often landing in a new, unstable state or sliding back to the traditional model. The only way to reach a truly modern market configuration was to change all three major rules simultaneously. Even then, the timing was critical. The simulations showed that if the reforms were not launched together within a specific window, the system would lock itself into the old, rigid patterns, making it nearly impossible to escape later. Delaying the changes, even by a significant margin, caused the opportunity for modernization to vanish entirely.
A particularly striking finding concerned the role of artificial intelligence itself. The researchers discovered that the amount of AI adoption matters, but not in a simple way. If AI use is moderate, it can actually help the system settle into a traditional, stable state. But if AI adoption becomes too intense without accompanying rule changes, it can trap the market in a rigid, high-tech loop that is difficult to break. Worse still, the simulation suggested that if a market had already managed to reach a modern, efficient state, an excessive surge in AI trading could actually cause it to collapse back into the old, rigid trap. This creates a paradox where the very technology expected to drive progress can, under the wrong conditions, reinforce the very barriers it was supposed to overcome.
The researchers also looked at how long these different states last when the market is disturbed by random noise or shocks. They found that the traditional, conservative market state is fragile; even small disturbances can push it out of place relatively quickly. In contrast, the modern, efficient market state is much more resilient and persistent, once it is actually reached. However, getting there is the hard part. The study concludes that the path to a modern market is not a gradual evolution but a sudden, coordinated leap. It requires a simultaneous overhaul of settlement rules, price limits, and short-selling mechanisms. Without this synchronized effort, the market is destined to remain stuck in a cycle where advanced technology is absorbed into an outdated framework, creating a system that looks modern on the surface but operates with the rigidity of the past. The findings suggest that for policymakers, the key is not to wait for the right moment to fix one piece of the puzzle, but to commit to changing the entire picture at once, before the system locks itself into a pattern that cannot be undone.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.