Chris Gilchrist: Perhaps the market just isn’t efficient

Chris Gilchrist FW Index 480

In a recent column I outlined three types of response to the 2008 crash. Perceptive readers have noticed that I missed one out.

In addition to long-only investors, tactical asset allocators and absolute returners, there is a fourth group, which you might think is the dominant one – yet I never speak to serious investment advisers who confess to being members.

I refer to the adherents of Modern Portfolio Theory. Their reponse to 2008 is pure denial. MPT methods work most of the time but not when there’s a crash – this just about sums up the depth of the intellectual response from theorists and practitioners.

Many advisers who have taken exams in recent years have learnt about MPT and claim to understand it, at least in principle.

As I have pointed out before, those without university-level maths will struggle with the detail of MPT because its probabilistic theories (think algebra and calculus) are beyond us.

But that does not mean we can not form judgments about its relevance and usefulness, since these depend on common sense and a grasp of history rather than of statistics.

I regard it as unfortunate that the exam creators at the CII and elsewhere continue to regard MPT as holy writ. 

More than 10 years of increasingly sceptical academic commentary has yet to make it into the exam syllabus.

But now that the FCA itself has published a paper on behavioural finance, there is no excuse for continuing to ignore the fundamental flaws in MPT.

First comes the lack of rationality of investors. Markets are prone to bursts of wild irrationality of which the dotcom bubble of 2000 is a near-perfect example. MPT is predicated on rationality, and not just any rationality, but super-rationality in which investors have perfect information and can forecast the future.

Then there is the asymmetry of risk preference.

MPT demands symmetry in relation to profit and loss but we know that it takes about £2 of profit to balance £1 of loss in investors’ own experience and that this shapes their behaviour, inevitably making it diverge from MPT models.

These two aspects of MPT have been adequately demolished by behavioural economics and their use as axioms in a model cannot be justified by anything other than brilliant success.

If MPT is so good, why aren’t there scores of portfolios rigorously managed on the “efficient frontier” that have beaten all comers?

Perhaps one reason is that many academics have concluded that the market is not efficient.

The most elegant proof of this is the use of simple “value” filters to create portfolios which have outperformed the market indices for the best part of 100 years in the US and UK. According to MPT, this just can not be – but it is.

The use of extra ugly maths to adapt MPT part-way to the real world is like the Ptolemaic astronomers’ invention of epicycles to try to match observed irregular planetary orbits to their idealised Platonic model. Kepler’s simple and elegant laws of planetary motion sent the Ptolemaic calculations to the dustbin.

Until we get a Kepler for finance – “chaos maths” may deliver a much improved model in the next decade – those navigating the treacherous terrain of the investment world will do better to use observation and history rather than the deeply flawed MPT map.

Chris Gilchrist is director of FiveWays Financial Planning, edits the IRS Report newsletter and is the author of the Taxbriefs Guide, The Process of Financial Planning