While many researchers have studied the performance of style investing strategies such as value, growth or small caps, studies dealing with the performance of smart beta portfolios are limited. This study tests the performance of a dynamic asset allocation strategy based on various smart beta portfolios that rely on a Markov regime-switching model based on macroeconomic regimes. Results and backtests show that using Markov regimes increases the performance of a dynamic smart beta portfolio based on Markov regimes compared to a static benchmark in-sample, and that such performance begins to erode when utilized out-of-sample considering one friction (trade costs). Also, this study finds that the choice of the economic variable used to estimate the Markov regime switching model is important for the performance of smart beta portfolios using Markov regimes based on macroeconomic indicators.