Dynamic Spillover of Negative Public Opinion on P2P Liquidity Risk Based on Sentiment Analysis

Abstract

The peer-to-peer (P2P) lending industry has emerged as a significant component of the alternative finance landscape, yet it remains plagued by inherent fragility and information asymmetry. This paper investigates the dynamic spillover effects of negative public opinion on the liquidity risk of P2P platforms, employing advanced sentiment analysis techniques and econometric spillover indices. By constructing a comprehensive daily index of negative sentiment derived from web-scraped data across major investor forums and news outlets, we analyze the transmission mechanisms through which adverse information impacts platform funding flows and withdrawal behaviors. We utilize textual mining to quantify soft information and apply a generalized vector autoregression framework to capture time-varying connectedness. Our findings reveal that negative public opinion exerts a substantial, immediate, and asymmetric impact on liquidity risk, with the intensity of the spillover varying significantly depending on the regulatory environment and the specific nature of the negative news. The results suggest that investor panic, fueled by social contagion, creates a self-fulfilling prophecy that exacerbates liquidity crunches. Furthermore, we identify a distinct lag structure where the volatility transmission from sentiment to liquidity persists even after the initial shock subsides. These insights provide critical implications for platform managers regarding crisis communication and for regulators aiming to stabilize the fintech ecosystem against systemic risks associated with information contagion.