How To Use The Financial Cycle And Economic Cycle To Improve Asset Allocation?

This paper uses the empirical mode decomposition algorithm to explore new ways of constructing financial cycles and economic cycles,and then adopts support vector machine to forecast stocks,bonds,commodities,cash and real estate yields.Finally,the configuration strategy of the five major types of assets is given by Black-Litterman model.The results show that EMD-SVM model based on US data is superior to the SVM,BP and EMD-BP models in terms of MSE.In the forecasting step less than 10 years,the prediction contribution of economic cycle(intermediate frequency component 2-8 years)dominates,while in the forecasting step more than 10 years,the contribution of financial cycle(low frequency component 8-30 years)becomes more and more important.Meanwhile the back test of Black-Litterman model based on US data show that the portfolio performance based on EMD-SVM model and EMD-BP model are better than the market portfolio as well as the SVM and BP models,showing that the predicted model values obtained by EMD decomposition are more conducive to provide effective BL asset allocation recommendations.Based on China data,the EMD-SVM model is not always dominant.In the tactical asset allocation,the Sharpe rate based on EMD-SVM and EMD-BP model is better than the market,SVM and BP model,while in the strategic asset allocation,the four models performs worse than the market portfolio.Therefore,it is recommended that investors in China should adopt tactical asset allocation in order to achieve return higher than market.