Active portfolio management in the Andean countries’ stock markets with Markov-Switching GARCH models

Abstract

In the present paper we test the benefits of using two-regime Markov-Switching (MS) models in the stock markets of the MSCI Andean index (Chile, Colombia and Perú). We tested this with either, constant, ARCH or GARCH variances and Gaussian or t-Student log-likelihood functions. By performing 996 weekly simulations from January 2000 to January 2019 with each MS model, we tested the next investment strategy for a U.S. dollar based investor: 1) to invest in the risk-free asset if the probability of being in the high-volatility regime at t+1 is higher than 50 % or 2) to do it in the stock market index otherwise. Our results suggest that the Gaussian MS-GARCH models are the most suitable to generate alpha in the Chilean stock market and the Gaussian MS-ARCH in the Colombian one. For the Peruvian case, we found that is preferable to perform passive investing instead of active trading.