Two articles in the August issue of Journal of Asset Management discuss topics that relate to random portfolios.

Sharpe ratios

The first article is “The Sharpe ratio’s market climate bias: Theoretical and empirical evidence from US equity mutual funds” by Sebastian Krimm, Hendrik Scholz and Marco Wilkens (abstract).  SSRN claims a version of the paper is downloadable, but all I got was the abstract.

The observation in the paper (though not stated this way) is that random portfolios with  large idiosyncratic risk will tend to outperform the index when the index falls, and underperform it when the index rises.

Thus they correctly observe that the Sharpe ratio is not measuring performance well.  Their fix for this problem is a “normalised” Sharpe ratio.

That is, they want to create a complicated statistic that people aren’t going to understand and that still won’t work well.

Better is to use random portfolios to directly measure investment decisions.

Replacing investment managers

The second paper is “How much value should you expect to gain or lose by replacing your investment manager?” by Robin Penfold (abstract and SSRN version).

This is rather in the same spirit as random portfolios in that it performs a Monte Carlo to try to answer the question in the title.  I suspect that performance measurement (and other experiments) with random portfolios may be able to improve the Monte Carlo.

An interesting paper by the same author on SSRN is “Evaluating an Investment Manager in an Uncertain World” which uses Bayesian statistics.

Hat tip to MoneyScience.

Update: a more complete list of Robin Penfold papers.