Time Diversification and Financial Forecasts for Bond and Equity Markets

Abstract

Depending tactical asset allocation upon the forecasting model choice, it is often discussed the existence of a time diversification opportunity. We focus the analysis on MSCI and JPMorgan benchmarks for equity and fixed income market in Italy, Switzerland, UK, Europe, USA, Japan, and Far East. The topic of this paper consists on the comparison of three methods to estimate the risk premium: the historical return, the building block and the CAPM, which help to optimise asset allocation decisions through different holding periods. Time series have been chosen in order to avoid structural breaks, recorded after 2000. We give an answer to some relevant questions (i) how long risk premiums take to become stable among reliable intervals? (ii) is it possible to define ex-ante criteria to choose forecasting and time diversification methods? We demonstrate how forecasts efficacy change through different holding periods; moreover, historical returns start loosing their (slight) ability to estimate the expected ones with lower holding periods than already experimented in literature. The second result is that reliability ranking depends on the preference for a time horizon.