Regime-based portfolio optimisation : a hidden Markov model approach for fixed income portfolios

This paper presents a methodology for incorporating a regime-based approach to portfolio optimisation, specifically for fixed income portfolios. By segmenting the market into distinct periods or "regimes" characterised by different market conditions, such as high or low volatility, investors can adjust their portfolio strategies accordingly. We propose a two-stage approach: first, applying Principal Component Analysis (PCA) to fixed income indices to represent the yield curve, and second, using the variance of the first principal component to fit a Hidden Markov Model (HMM) that identifies high and low volatility regimes. This regime-based methodology allows portfolio managers to make more informed decisions, potentially improving risk-adjusted returns. We apply this approach to portfolio optimisation, targeting a specific value-at-risk (VaR), and compare the results to a traditional approach that does not account for regime changes.