This paper emphasizes the importance of identifying changes in financial cycles when predicting monthly US excess stock returns for the period 1977 - 2017. Incorporating regime switching into the predictive models improves the quality of the excess return forecasts in terms of market timing ability, economic value and stability. The Markov Switching models consisting of predictor variables selected based on their performance during bull and bear markets performs especially well. A mean-variance investor would be willing to pay several hundreds basis points to switch from the static benchmark portfolios to one of these portfolio strategies.

Additional Metadata
Keywords Keywords: Return predictability, Markov Switching models, factor models, variable selection, Lagrange Multiplier test, market timing, economic value
Thesis Advisor Kole, H.J.W.G.
Persistent URL
Series Econometrie
Can, H. (2019, April 3). To Switch or Not to Switch: Return Prediction and. Econometrie. Retrieved from