Knowledge that Transforms

Financial sentiment analysis models, which extract meaning from vast amounts of unstructured data, play a crucial role in sentiment-driven financial decisions. However, the complex and domain-specific language used in finance poses unique challenges for adversarial attacks. To address these challenges, we propose a novel, white-box attack methodology leveraging a pre-trained general-purpose language model (GPT-4o). We employ carefully designed instructions and incorporate a new loss function based on embedding similarity to ensure semantic coherence while producing syntactically diverse samples. Our experimental results demonstrate that both FinBERT and Fin-GPT, leading models in financial sentiment analysis, exhibit significant susceptibility to our proposed adversarial attacks. Specifically, the sentiment predictions of these models were successfully altered for a substantial proportion of the samples across three public datasets, including Financial Phrase Bank (FPB), Twitter Financial News Sentiment (TFNS), and Sentimence and Entity Annotated Financial News (SEntFiN). Our findings emphasize the need for enhanced robustness in financial classification models against adversarially targeted attacks. By understanding and addressing these vulnerabilities, it is possible to improve the reliability and security of automated financial systems.

Regulatory punishment in an oligopolistic credit rating market can be costly. Utilizing the Chinese bond market’s unique features, particularly a third-party rating agency, we investigate the regulatory suspension of Dagong Rating by Chinese regulators and its market impact. The punishment initially deters Dagong but diminishes the quality of its ratings post-punishment, altering market competition. Upon returning, Dagong inflates ratings to regain market share, reflecting a “temporary suppression” strategy. Non-Dagong agencies respond by adjusting their ratings; higher power agencies lower ratings, while lower power agencies raise them to stay competitive. Investors remain skeptical of these inflated ratings. Despite Dagong’s suspension, we find no significant differences in bond or stock price reactions between Dagong-rated and non-Dagong-rated firms, suggesting investors did not penalize Dagong-rated entities. This study highlights the complex dynamics and unintended consequences of regulatory interventions in the credit rating market.