Research: Stablecoin De-Pegging May Have a "Critical Threshold," Severe Narrative Shocks Can Trigger Panic Selling

TechFlow reported that on September 7, the Renmin University FinTech Institute released a summary of stablecoin research published in the Journal of International Money and Finance. The research employs large language model (LLM) agents to analyze stablecoin decoupling risks, highlighting that such decoupling exhibits distinct non-linear and threshold characteristics. While arbitrage mechanisms can absorb moderate information pressure, once narrative severity hits a "critical threshold," the system may rapidly transition into a sustained decoupling state.

The study argues that crises are not merely triggered by a straightforward "bad news → price drop" mechanism, but may instead unfold through a self-reinforcing process: "severe narrative shock → rising fear → liquidity deterioration → mass retail selling → arbitrageur withdrawal → order imbalance → sustained decoupling." Furthermore, the research finds that during experiments with the highest severity narrative shocks, the average decoupling magnitude across different risk sources shows no statistically significant difference. This indicates that, within the model's framework, the severity of the narrative shock likely outweighs the specific content of the narrative.