Why optimal diversification cannot outperform naive diversification: Evidence from tail risk exposure

“In the literature on optimal portfolio choice, the outperformance of the 1/N portfolio strategy relative to the optimal portfolio strategy in out-of-sample asset allocation tests is largely attributed to estimation error in the optimal portfolio strategy…….academic research proposes various extensions of the Markowitz model to reduce estimation errors with the goal of improving the performance of the Markowitz model.

However, despite the considerable effort required to handle estimation error in the optimal portfolio strategy, this approach does not consistently dominate naive diversification. Recently, DeMiguel, Garlappi, and Uppal (2009b) report that none of the sample-based mean–variance models and almost none of the sophisticated extensions of the Markowitz rule consistently outperform the 1/N strategy.”

Leave a Reply