Can we invest on the basis of equity risk premia and risk factors from multi-factor models ?

We examine two investment algorithms built on the weekly data of world equity indices for emerging and developed countries in the period 2000-2015. We create seven risk factors using additional data about market capitalization, book value, country GDP and betas of equity indices. The first strategy utilizes the theoretical value of equity risk premium from the seven-factor Markov-switching model with exogenous variables. We compare theoretical with the realized equity risk premium for a given index to undertake the buy/sell decisions. The second algorithm works only on eight risk factors and applies them as input variables to Markowitz models with alternative optimization criteria. Finally we note that the impact of risk factors on the final results of investment strategy is much more important than the selection of a particular econometric model in order to correctly evaluate the equity risk premium.