We live in troubled times. Since the dawn of time, which in this case means the late 70's, the asset allocation of pension funds has been governed by a mean-variance-optimization (MVO) process that springs from the so called modern portfolio theory (MPT). Then came the TMT-crisis. Confidence was shaken, but the alternative that emerged - The Yale Model - never really threatened MPT as it was built on a parallel process but added illiquidity, leverage and (potentially) increased diversification to the recipe. Then came the Leman-crisis. As the authors state "We did not abandon modern portfolio theory; it abandoned us." The now emerging alternative - risk parity - thus abandons MVO.
On the book cover it reads "An efficient alternative to [MPT]", yet in the introduction Greg Davies (Global Head of Behavioral and Quantative Investment Philosophy at Barkleys Wealth) and Arnod de Servigny (Global Head of Discretionary Portfolio Management and Investment Strategy at Deutsche Bank Wealth Management) says that they are not trying to "set a new holistic standard, a so called successor to [MPT]". Both statements are correct. The book presents a full MVO-model in the sense that it is possible to use in practice. However, it's not a theoretical alternative to MPT as some parameters are subjectively chosen. The model is an attempt to improve the MVO-process based on insights from behavioural finance. You have to applaud this approach. Behavioural finance excels in finding flaws with MPT but the discipline is seldom used in this more constructive way.
So what are the issues with MPT that the authors aim to correct? First, the use of normally distributed variance as a representative of the statistical distribution is clearly invalid for most assets. On top of this, the normal distribution assumes a linear and stable trade-off between risk and return in investors utility function, where Daniel Kahneman and Amos Tversy with their Prospect Theory instead shows risk aversion to be non-linear and dependant on whether returns are positive or negative. Secondly, MPT assumes stable correlations between assets (giving stable diversification benefits) and hence ignores the evidence for time-variations and regime dependence in both correlations and returns.
The 2 main features of the book are a) a "behaviourised" utility/risk function and b) a regime switching model to handle the different correlations and returns during times of stress versus more tranquil markets. The risk inconsistencies are handled first by adding the higher moments of the distribution, like skew and kurtosis, to the risk measure and secondly by adding an individual, subjective, risk tolerance factor to the equation. The regime switching model uses analysis of historical data to distinguish between different states of the world. The authors find that once their model has switched to a regime it stays there between 2 and 5 years. In the different regimes very different assumptions of correlations and returns go into a MVO-process with very different asset allocations as a result.
This is a seriously geeky book with its combination of portfolio theory and financial psychology and it's a relatively heavy read for the less mathematically inclined. Yet it is an unusual and important book that addresses several of the most acute topics discussed in pension funds today. Regime switching models are quite the rage currently and as they often use volatility based signals to differentiate between regimes, the difference in resulting allocation versus risk parity-strategies might in effect not be that large. It is also obvious that these models are active allocation strategies and as such they cannot be used by all investors collectively, i.e. they are not macro consistent (in fact it would be highly un-stabilising for the market if too many used the strategy). That doesn't mean that the models couldn't be good for those who use them. However, the behaviourising of the utility function is clearly very innovative work. I've not seen it anywhere else. If I could have wished for a more comprehensive coverage of one related topic it would have been the time variance in expected returns.
The authors show a clear understanding of psychology when concluding that it is better to have a slightly suboptimal portfolio that is possible to stay with, compared to an optimal portfolio that is psychologicaly unbearable and as such is sold at exactly the wrong time. As J.M. Keynes put it: "There is nothing so dangerous as the pursuit of a rational investment policy in an irrational world."
This is a review by eqtbooks.com