Scholarlink

Narrative economics, popularized by Shiller in 2017, studies the role of powerful, often viral stories in shaping economic decisions and consequent market outcomes. This approach helps to understand how narratives and formed opinions can drive investor behavior and lead to significant financial phenomena, such as bubbles and crashes. In this paper, I propose an agent-based trading model integrated with opinion dynamics to investigate the impact of such narratives on financial market volatility. Building upon my foundational research [5, 6], this study examines the dynamics of market behavior under two competing narratives, focusing on three hypotheses: self-reinforcement, herding, and an additive response to inputs. Using empirical validation, the findings indicate that all three factors—when hyper-parameters are optimally calibrated—can influence market volatility, validating all proposed hypotheses. The source code is publicly available on GitHub, facilitating further research and replication.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Competing Narratives in Financial Markets: An Opinionated Agent-Based Model for Simulating Volatility

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Arwa Bokhari

摘要

Narrative economics, popularized by Shiller in 2017, studies the role of powerful, often viral stories in shaping economic decisions and consequent market outcomes. This approach helps to understand how narratives and formed opinions can drive investor behavior and lead to significant financial phenomena, such as bubbles and crashes. In this paper, I propose an agent-based trading model integrated with opinion dynamics to investigate the impact of such narratives on financial market volatility. Building upon my foundational research [5, 6], this study examines the dynamics of market behavior under two competing narratives, focusing on three hypotheses: self-reinforcement, herding, and an additive response to inputs. Using empirical validation, the findings indicate that all three factors—when hyper-parameters are optimally calibrated—can influence market volatility, validating all proposed hypotheses. The source code is publicly available on GitHub, facilitating further research and replication.