The Economic Value of Nonlinear Predictions in Asset Allocation

摘要: Predictions of asset returns and volatilities are heavily discussed analyzed in the finance research literature. In this paper, we compare linear nonlinear predictions for stock- bond index their covariance matrix. We show in-sample out-of-sample prediction accuracy as well impact on allocation results short-horizon investors. Our data comprises from German DAX stock market REXP joint matrix over period 01/1988 - 12/2007. The comparison a approach is focus study. that while accuracies weak terms statistical significance, performances based result significant Jensen's alpha measures Sharpe-ratio further improved by predictions.