Addressing Market Regime Changes and Heavy-Tailed Returns in Portfolio Optimization via Bayesian VAR and Elliptical Black-Litterman

2026-06-08

AI Summary

This study addresses the limitations of conventional deep reinforcement learning approaches in portfolio optimization, which often overlook the heavy-tailed nature of asset returns and regime shifts in financial markets, leading to poor performance under extreme risk scenarios. To overcome these issues, the authors propose a novel framework that integrates a Bayesian Average Vector Autoregression (BAVAR) model with an elliptical-distribution-based Black-Litterman extension (BLED), embedded within a TD3 reinforcement learning architecture. The framework employs a Transformer to generate dynamic market views, a CNN to estimate time-varying risk aversion, and a Student’s t-distribution to capture return heavy tails, enabling adaptive asset allocation through multi-scale temporal modeling. Evaluated on ten years of data from 29 Dow Jones constituents, the method achieves a total return of 57.26%, an annualized Sharpe ratio of 1.72, and a Sortino ratio of 2.70, significantly outperforming existing benchmarks.

Abstract

Deep reinforcement learning (DRL) frameworks for portfolio optimization have shown promise for their ability to learn allocation rules dynamically from market data. However, these models fail to account for fat-tailed returns, which characterize actual market behavior with more frequent extreme events. Furthermore, historical data is treated homogeneously, without accounting for temporal importance, leading models to fail during regime changes. We propose a new BAVAR-BLED algorithm that combines methods derived from Bayesian-Averaging Vector Autoregressive (BAVAR) and the Black-Litterman model using Elliptical Distributions (BLED) within a TD3 architecture. BAVAR captures a set of vector autoregressive representations that consider multi-scale temporal features, enabling adaptive allocation decisions based on regime-aware estimates of return expectations and dispersion matrices. These estimates serve as prior inputs to BLED, a model that uses Student's t-distributions, allowing for more realistic fat tail return estimates. The BAVAR-BLED algorithm uses transformer networks for view construction and CNNs for risk-aversion estimates, which modify dynamic allocation decisions based on market conditions. An evaluation of 29 Dow Jones Industrial Average constituents over a decade-long market period shows that BAVAR-BLED significantly outperforms state-of-the-art methods, achieving Sharpe and Sortino ratios of 1.72 and 2.70, respectively, and total returns of 57.26%.

Problem

• portfolio optimization
• market regime changes
• heavy-tailed returns
• fat tails
• regime shifts

Innovation

• Bayesian VAR
• Elliptical Black-Litterman
• Heavy-tailed returns
• Market regime changes
• Deep reinforcement learning