This paper presents a practical investment framework for dynamic asset allocation strategy based on changes in macro environments. To identify economic regimes, we first construct macro indicators that monthly track the U.S. economic activity and inflation, respectively. We then show that the regimes divided by changes in growth and inflation successfully partition the historical performance in asset classes and suggest the dynamic strategy of tilting exposures to the attractive assets based on the regimes. Out-of-sample analysis in performance indicates that the dynamic approach outperforms the static approach after taking transaction costs into account, improving risk-adjusted returns significantly. The result has an important implication for portfolio managers who seek to effectively develop a dynamic asset allocation strategy throughout economic cycles to enhance the long-term portfolio performance.​

JEL Classification Numbers: G11, G15, E32
Keywords: Dynamic asset allocation; Economic regimes; l1 trend filtering