Title: Portfolio choice: Balancing forecasting risk
Authors: Ingmar Nolte - Lancaster University (United Kingdom)
Sandra Nolte - Lancaster University (United Kingdom)
Ekaterina Kazak - University of Manchetser (United Kingdom) [presenting]
Yifan Li - Lancaster University (United Kingdom)
Abstract: Estimation noise is a well-known issue in empirical portfolio modelling. Estimated weights are known to have huge standard errors and bad predictive quality, which often results in an inferior out-of-sample portfolio performance compared to simple alternatives. Most of the recent literature concentrates on the improvement of covariance matrix forecasts, which would hopefully result in better portfolio performance. However, the proposed models often suffer from the dimensionality problem, such that the forecasting error still dominates the theoretical gain. We propose a portfolio choice model, which explicitly takes into account forecasting risk and avoids the dimensionality problem by forecasting a one-dimensional portfolio measure directly. We then define a forecasting error based on the realized measures and look for weight estimates which results in the more precise forecast in terms of the forecasting error variance and at the same time is not far from the optimal portfolio solution. The proposed approach is close to the James-Stein type of estimator, which balances bias-variance trade-off in a data-driven manner. The proposed method is shown to outperform the commonly used approaches in both simulation and empirical studies.