New theory uses LLMs to estimate firm risk when price data is scarce

Researchers have developed a new theoretical framework called DisclosureBeta to estimate a firm's beta, a measure of its sensitivity to market risk, when historical price data is limited. This approach models a large language model (LLM) as a noisy measurement channel for a firm's risk characteristics, incorporating this noise into the asset-pricing error budget. The theory provides a method for identifying and consistently estimating regime-conditional loadings, offering a lower bound on estimation precision that accounts for unavoidable disclosure-noise and misclassification terms. An adaptive combination of text-based and rolling-window estimators is proposed, which dynamically adjusts its weighting based on the quality and length of price history. AI

IMPACT Introduces a novel application of LLMs in quantitative finance for risk assessment, potentially improving financial modeling accuracy.

RANK_REASON Academic paper detailing a new theoretical framework for financial risk measurement. [lever_c_demoted from research: ic=1 ai=0.4]

Read on arXiv cs.CL →

• Adad
• arXiv
• Breitung
• DisclosureBeta
• Fama-French five factor model and the necessity of value factor: Evidence from Istanbul Stock Exchange

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →