Black-Litterman Portfolio Optimization with Dynamic CAPM via ABC-MCMC: Empirical Evidence from the Vietnamese Stock Market

Abstract

This study proposes a robust portfolio construction method integrating the Black-Litterman model with the CAPM model. In this study, market status is determined using a risk-adjusted score-based index. This status is derived from the Sharpe rotation ratio, estimated through the Markov Chain Monte Carlo algorithm using approximate Bayesian calculations. The research model uses a weekly dataset of 33 representative stocks of the Vnindex from 2019 to 2025. The research results show strong evidence of the highly volatile Beta of the Vnindex, indicating very high market risk in Vietnam.

Empirical results show that the proposed active investment strategy significantly outperforms passive strategies using a static Markowitz model. Furthermore, sensitivity analysis showed that a shorter adjustment window (6 months) yielded higher forecasting accuracy compared to the standard 1-year timeframe, reflecting the rapid price volatility of the Vietnamese market. These findings led the team to recommend that "applying a proactive and conditional investment strategy will yield more positive results for investors in the Vietnamese stock market"