Usefulness of Bootstrapping in Portfolio Management

This paper contains a comparison of in-sample and out-of-sample performances between the resampled efficiency technique, patented by Richard Michaud and Robert Michaud (1999), and traditional Mean-Variance portfolio selection, presented by Harry Markowitz (1952). Based on the Monte Carlo simulation, data (samples) generation process determines the algorithms by using both, parametric and nonparametric bootstrap techniques. Resampled efficiency provides the solution to use uncertain information without the need for constrains in portfolio optimization. Parametric bootstrap process starts with a parametric model specification, where we apply Capital Asset Pricing Model. After the estimation of specified model, the series of residuals are used for resampling process. On the other hand, nonparametric bootstrap divides series of price returns into the new series of blocks containing previous determined number of consecutive price returns. This procedure enables smooth resampling process and preserves the original structure of data series.