A comparative analysis of resampling efficiency and Black-Litterman portfolio optimization

It is common knowledge that the traditional Mean-Variance (MV) approach presents several non-negligible criticalities such as unintuitive and highly concentrated portfolios, input sensitivity, and estimation error maximization. Black-Litterman (BL) model and Resampled Efficiency (RE) techniques are advanced methods that help to generate better allocations than the traditional “a la Markowitz” method. On the one hand, the Black-Litterman model represents a well-known approach that overcomes this issue by assuming partial information on the expected returns. By blending a reference market distribution with subjective views on the market, the approach yields optimal portfolios that smoothly reflect those views. On the other, portfolio resampling, following a heuristic approach allows the portfolio manager to visualize the estimation error in traditional portfolio optimization methods. Starting from a thorough review of the literature about traditional portfolio theory and a discussion of its limitations, the two approaches: BL and RE are introduced. The core of this work will focus on the implementation of the Black-Litterman model together with resampling techniques for portfolio allocations, carrying out an empirical examination of the usefulness of those technologies in reducing estimation errors.