Tokyo, Japan – The influence of financial markets cannot be overstated, and this influence highlights the need for thorough market analysis. However, being thorough can be difficult when markets are highly complex and ever-changing, prompting the search for a suitable approach to analysis. Fortunately, the increasing availability of detailed market data, together with advances in mathematical theories, offers new opportunities to better understand market behavior.
Researchers from The University of Tokyo and Rikkyo University have recently tested whether one mathematical theory could give early warning of market instability. Their findings were published in EPJ Data Science .
Dynamical Network Markers (DNM) theory was originally developed by the study’s senior author Kazuyuki Aihara of The University of Tokyo and colleagues to detect a pre-disease state. However, this theory can apply to more than just health, including other interacting, complex networks such as financial markets. DNM theory can help detect signs of worrying changes in advance by picking out changes in key elements. The researchers analyzed order and execution data from the Tokyo Stock Exchange.
“Our insight into the Tokyo Stock Exchange was taken from market participants identified by grouping anonymized virtual server IDs that displayed similar trading activities,” explains lead author Mariko Ito of The University of Tokyo. “To assess a snapshot of participant activities, we gathered one-minute measures of data.”
To assess the performance of the theory in relation to financial markets, the team quantified the trading activities of 186 anonymized market participants using data collected from late 2019 through 2020. The researchers noticed that fluctuations in indicators representing the trading activity of certain participants tended to increase in the days leading up to several of the five focal turmoil days, days on which the market showed clear instability.
This pattern is qualitatively consistent with DNM theory. When each participant was examined alone, there was no statistically clear sign ahead of the unstable days. The early signs may therefore appear in the collective behavior of groups of participants, rather than in the actions of single participants.
While many previous studies have investigated financial market instability primarily through aggregate market indicators, the present study examined the trading behavior of market participants. In some cases, fluctuations in indicators representing the trading activity of certain participants increased before a turmoil day, while a standard measure based on past price changes (lagged volatility) rose only after it. This suggests that trading behavior may contain early signs of instability that are hard to see in prices alone.
However, the approach is not yet capable of predicting sharp price changes with practical accuracy. Markets are also influenced by external factors. In February and March 2020, when external factors affected the market again and again, no clear rise was seen in the group fluctuations. Even so, the researchers see this work as a first step toward an early warning system that could help financial authorities and stock exchanges notice and respond to emerging market instability.
“My field is using mathematics to design systems that keep the flows in society moving smoothly. A stock market is also such a flow, with orders moving all the time,” says corresponding author, Yudai Honma of The University of Tokyo. “Instead of looking for the cause after prices change sharply, we want to look for signs of instability in the flow of orders before that happens.”
###
The article, “Identifying dynamical network markers of financial market instability,” was published in EPJ Data Science at DOI: 10.1140/epjds/s13688-026-00710-w.
The Institute of Industrial Science, The University of Tokyo (UTokyo-IIS) is one of the largest university-attached research institutes in Japan. UTokyo-IIS is comprised of over 120 research laboratories—each headed by a faculty member—and has over 1,200 members (approximately 400 staff and 800 students) actively engaged in education and research. Its activities cover almost all areas of engineering. Since its foundation in 1949, UTokyo-IIS has worked to bridge the huge gaps that exist between academic disciplines and real-world applications.
EPJ Data Science
10.1140/epjds/s13688-026-00710-w
Identifying dynamical network markers of financial market instability
13-Nov-2026