Exploring the Behaviour of Actively Managed, Maximally Diversified Portfolios

Abstract

Maximising returns is often the goal of asset management, but in the current (2020) low-interest investment environment, plagued by political, trade and economic uncertainty, managing portfolio risk also plays a significant role. Maximally diversified (MD) portfolios are assembled with an emphasis on risk management, not return outperformance. This approach can yield considerable benefits for risk- averse investors. While some work has been done applying this technique to passive portfolios, little to none has been undertaken on active portfolios, restricted by tracking errors (TEs) and evaluated relative to a benchmark. For the first time, actively managed maximum diversification portfolios are scrutinised over time. In booming markets, actively managed MD portfolios generate significant outperformance, but during recessionary periods, no significant benefits emerge. Returns and Sharpe ratios are weak and volatilities high (albeit lower than other strategies). As TE increases, actively managed MD portfolio weights become less confined and adjust ever closer to the overall (unconstrained) MD portfolio weights. Fewer benefits are realised as TEs increase for actively managed MD portfolios compared with the unconstrained alternative.

Notes

1 This is counterintuitive but can be demonstrated empirically and mathematically. Even the benchmark portfolio’s risk and return can be reproduced exactly, using portfolio weights that differ from those which constitute the benchmark. The reason is that there is not a one-to-one mapping between an amalgamation of portfolio weights and a given combination of return and risk. Mappings are instead many-to-many and all positions within the constant TE frontier have the same TE.