A Dynamic Asset Allocation Strategy with Macroeconomic Indicators
Abstract
This thesis proposes a novel approach to strategic asset allocation (SAA) that uses macro-econometric factors and a regime-switching regression model to capture systemic risk, which is crucial in setting target allocations for various asset classes and measuring associated risk premiums. The proposed model employs a Hidden Markov Model (HMM) to filter the four macro-econometric factors, which are the main source of market risk, and uses a mean-variance-based dynamic selection model to increase the portfolio's risk-adjusted return. The empirical analysis utilizes global bond and equity ETFs, as well as sector ETFs, to test the asset pricing model and evaluate the performance of the portfolio. Weekly financial data from January 01, 2002, to December 31, 2022, are used to construct the four macro-econometric factors, and the four weekly macro-econometric factors from September 06, 2016, to January 1, 2019, are employed to estimate the HMM and asset return parameters. The results show that the proposed model consistently outperforms the benchmark MSCI ACWI Index in the out-of-sample period from January 1, 2019, to December 31, 2022, and outperforms the selected ETFs and the mean-variance models in terms of the Sharpe Ratio. The proposed approach to SAA provides a promising avenue for investors to achieve superior risk-adjusted returns in the current global financial landscape.