Does portfolio resampling really improve out-of-sample performance? evidence from the Brazilian market.

Does portfolio resampling really improve out-of-sample performance? evidence from the Brazilian market.

• Author
• André Oliveira
• Co-authors
• Carlos Trucios , Pedro L. Valls Pereira
• Abstract

•  

  Markowitz optimization plays an important role in modern portfolio theory. However, it is well-known that Markowitz optimization is highly affected by the estimation error of the mean vector and covariance matrix, resulting in extreme and/or unrealistic portfolio weights, lacks of diversification and poor out-of-sample performance. To deal with this issue, Michaud and Michaud (1998) proposed a heuristic portfolio resampling approach which can deliver more diversified and better out-of-sample portfolio performance in practice. In this paper, we assess the performance of the Michaud and Michaud (1998) portfolio resampling approach in the Brazilian context and also introduce a new portfolio resampling scheme called factor-based portfolio resampling, which takes advantage of the factor structure of stock returns. The results suggest that portfolio resampling can be an easy to implement alternative to increase portfolio diversification, reduce transaction costs and improve out-of-sample performance in the Brazilian context

• Keywords
• Bootstrapping, Covariance Matrix, Diversification, Efficient Frontier, Markowitz, Portfolio Allocation.
• Subject Area
• Econometrics and Numerical Methods

Back Download