Paper Contents
Abstract
Dynamic hedging provides the means by which traders and institutions manage their exposure to price, volatility, and interest rate risks. Conventional models like Black, Scholes, Merton are based on assumptions of continuous markets and quantitative rebalancing only (Hull, 2022), thus they hardly take into account qualitative factors such as news, sentiment, and macroeconomic events which nowadays have a considerable impact on volatility (Tetlock, 2007; Loughran & McDonald, 2016).The recent improvements of large language models (LLMs) make it possible to derive structured insights from unstructured texts such as financial news and social media (Brown et al., 2020; Yang et al., 2025). The paper presents a hybrid dynamic hedging mechanism that uses LLM, generated sentiment analytics as an input to hedge ratio estimation. Sentiment, adjusted variables that represent the model's tone and the intensity of the text allow more flexible hedging decisions.They have been able to demonstrate through backtests of equity options portfolios (2018, 2024) that the employment of sentiment, informed hedging helps to lessen errors and promote stability in periods of volatility. In spite of difficulties like latency and transaction costs, the use of LLM for the implementation of hedging strategies is a customer, oriented, and adaptable method that is capable of mixing quantitative rigor with linguistic insight.
Copyright
Copyright © 2025 Mr. Vaivaw Kumar Singh, Dr. Kunal Sinha. This is an open access article distributed under the Creative Commons Attribution License.
