Who would have thought that the front line in the war against modern portfolio theory would be held by the legal profession, but that is just is what appears to be happening.
Modern portfolio theory depends on efficient markets and general equilibrium, and its implied random independent price movements, to support its mean variance optimization derived efficient frontier.
If the price is right, and you know the expected return and risk of all investments and assets, and you have tools that can combine all those assets into an efficient portfolio, and it is impossible and expensive to consistently beat the market by active management, then a fiduciary would be in breach of his or her responsibilities if they did not use these tools and inputs to construct and manage a portfolio.
That is effectively what the legal position was, at least under US law, with the American Law Institute’s 3rd restatement of trust law.
The problem is that market prices are not always right: they can be wildly wrong and dangerously so when it comes to planning and managing the ability of a portfolio to meet future liabilities. If markets are not efficient, and you ignore the very real valuation risks of investment, then by combining assets while ignoring their non equilibrium, non efficient market risks, fiduciaries are being negligent, because they are ignoring risks which the model does not manage.
Modern portfolio theory’s central assumption that all prices are right, has no doubt had a helping hand in the build up of financial market and economic excess, because this so called efficiency did not just impact investment, it also impacted central bank management of (asset focused) money supply growth: the whole system had moved so far out of whack.
It is as we turn to Fiduciary responsibility for the world’s developed financial services’ market places, that I feel we must acknowledge the critical weakness and blind ambivalence of modern portfolio theory towards significant absolute and relative valuation risks.
That does not mean that we do not need structure, we do, but we need to pay more attention to valuation and structural economic risks that affect market dynamics, prices and their returns.
Diversification based on relative price movements (this is still a correlation bound structure) and absolute valuation (one key element of risk in a non general equilibrium world), that is sensitive to the impact of significant risks to the ability of assets to meet future needs (prices are path dependent out of equilibrium), would result in a dynamically diversified structure better able to manage market risks.
The mean variance structure however depends on linear inputs and is incapable of working with assets that are significantly over/under valued (out of equilibrium)– assets would be excluded, or wildly over weighted, such is the sensitivity of these models.
General equilibrium is critical to the operation of Mean Variance Optimisers. Stable inputs are needed for efficient frontier production and much of the work of the last 20 years has been spent trying to make these inputs more stable and the outputs less sensitive to risk and return assumptions that many have ignored the real issues.
I feel that the dominance of modern portfolio theory has severely hampered the development of disciplines and structures that manage significant market and economic risks and that while many of the lessons and benefits of indexation versus most of the needless so called active management still remains, there is no such thing as an MVO derived efficient frontier.
We need to work back from an equilibrium structure to develop a framework that is sensitive to both liabilities and their timing and the excesses of market risk. Once you are out of equilibrium the efficient frontier cannot be calculated, and asset allocations cannot be quantified. You lose the benefits of diversification because you cannot capture a movement towards or away from equilibrium.
An interesting post on this issue can be found at the Psy-Fi blog, and I will be extracting excerpts from a number of reports on this issue in the next few posts.