Modeling Market Reactions with Bayesian Return Adjustment in Financial Event Studies

Abstrak

Purpose: This study introduces the Bayesian Mean Adjusted Return Model (BMARM) to expand financial event study analysis and address the limitations of frequentist approaches during geopolitical crises. The model is applied to assess Indonesian banking stock reactions to the Russia–Ukraine war, focusing on abnormal returns, volatility shifts, and market stabilization. Research Methodology: Using daily stock returns of 44 IDX-listed banking issuers from July 2021 to March 2022, this study combines event study methodology with Bayesian inference. Bayesian Paired Sample t-tests, prior distribution selection, abnormal return estimation (BAAR and BCAR), and robustness checks were conducted using JASP. Expected returns were generated using Bayesian mean-adjusted, market-adjusted, and market models. Results: The findings show short-term negative shifts in BAAR immediately after the event, indicating increased volatility and declining investor sentiment. However, BCAR reflects a gradual improvement, suggesting a partial market recovery. Bayesian tests show weak evidence of differences between pre- and post-event abnormal returns, and robustness checks reveal sensitivity to prior assumptions. Conclusions: BMARM offers a more adaptive and probabilistic assessment of market reactions than classical models, capturing uncertainty during geopolitical shocks. It supports market efficiency patterns in which short-term disruptions eventually transition to stabilization. Limitations: The model is computationally intensive, dependent on prior selection, and limited by sample size. Further adaptation is required for nonfinancial market applications. Contributions: This study advances Bayesian empirical finance by introducing BMARM as a novel framework for event studies in emerging markets.