Early signs of stock market instability may appear in how traders place orders

The study conducted by researchers from The University of Tokyo and Rikkyo University suggests that early signs of stock market instability may manifest in the way traders place orders. Published in EPJ Data Science, the research utilized dynamical network markers (DNM) theory, originally developed for identifying predisease states, but applicable to complex networks like financial markets.

The team analyzed order and execution data from the Tokyo Stock Exchange, focusing on trading activities of 186 anonymized market participants over a period from late 2019 to 2020. They discovered that fluctuations in indicators representing the trading activity of certain participants tended to increase in the days leading up to several days marked by market instability.

However, this early warning sign was not evident when examining individual participants. Instead, the patterns emerged from the collective behavior of groups of participants. The researchers noted that while many previous studies have focused on aggregate market indicators, this study examined the trading behavior of individual market participants.

In certain cases, fluctuations in indicators representing certain participants' trading activity increased before a turmoil day, whereas a standard measure based on past price changes (lagged volatility) only rose afterward. This indicates that trading behavior might contain early signs of instability that are difficult to discern from prices alone.

Despite the promising findings, the approach currently lacks the capability to predict sharp price changes with practical accuracy. External factors, such as those experienced in February and March 2020, also affected the market without showing a clear rise in group fluctuations. Nevertheless, the researchers view this work as an initial step towards developing an early warning system that could alert financial authorities and stock exchanges to emerging market instability.