Regime shifts present significant challenges for investors, because these shifts cause within-period performance to depart significantly from the ranges implied by long-term averages of means and covariances. But these regime shifts also present opportunities for gain. We show how to apply Markov-switching models to identify and forecast regimes characterized by turbulence, inflation, and GNP. We tested a simple investment strategy using these regime forecasts to scale exposure to specific risk premia, and we found that it outperformed constant exposures. We also applied the same methodology to shift exposures across stocks, bonds, and cash. Again, we found that a dynamic process outperformed static asset allocation, especially for investors who seek to avoid large losses.