Harvesting Term and Credit Risk Premia: A Regime-Based Fixed-Income Allocation Framework

Dynamic asset allocation in fixed-income portfolios requires systematic rules for rotating duration and credit exposure as rate and spread conditions evolve. We develop a regime-based framework that harvests the term premium and the credit risk premium through two independent hidden Markov models with Student’s t emission distributions. The rate regime is identified from the Nelson–Siegel level factor net of the shadow short rate, which captures the effective stance of monetary policy including unconventional measures; the credit regime is identified from the BBB credit spread. Unlike predictive approaches prone to overfitting, detected regimes map directly to fixed duration and credit weights grounded in economic reasoning, with no parameter optimization. The framework is validated out-of-sample in euro and US dollar markets through walk-forward recalibration. Net of transaction costs, it outperforms a static equally weighted government credit benchmark in both markets, with higher Sharpe and information ratios and smaller maximum drawdowns, and the out-of-sample Sharpe ratio remains close to its in-sample value, suggesting that the regime classification captures economic structure rather than in-sample noise.