In this article from The IRS Report in April 2009, Chris Gilchrist talks about asset allocation.……

Is It Time To Abandon Asset Allocation Investment Planning?

The premise behind using asset allocation as a method of managing portfolios is that it reduces risk. This is true of all methods of applying diversification to portfolios, but as Warren Buffet has remarked: "Diversification may protect wealth, but concentration builds wealth." This means that you should not judge asset allocation in terms of how much money it has enable you to make in rising markets, when you could always have made more by putting more capital into the best performing, usually riskier, assets. The real test of diversification, and asset allocation, comes in bear merkets. Has it helped you to preserve wealth? The answer is that in 2008 it did not help. All the major asset classes tanked together: shares, commercial property, residential property and credit. Worse, virtually every sub-class also tanked: Asian shares, emerging markets, resources, private equity, infrastructure, and so forth; every type of corporate credit was marked down, the only fixed interest winners being developed government bonds; and residential property crashed from California through Western Europe to Australia. Nothing zigged. Everything zagged.

The result is that many practitioners of asset allocation methods have been left with sizeable quantities of egg on their faces. A 'cautious' portfolio containing a slug of sterling credit, commercial property and defensive UK equities is down 20% to 30% over the past twelve months. The models used by planners say this shouldn't have happened, and most of them will have given pretty clear indications to clients that they would be most unlikely to see losses on that scale.

So how are asset allocation practitioners responding to this failure of their methods? There are two responses, and unsurprisingly the majority of people are adopting the "Say nothing and hope it all turns out all right" approach. if you want to add an intellectual gloss to this, you can say that the fact of increasing downside correlation is well known but that it was the unprecedented and global scale of the deleveraging prompted by the credit crunch that caused such huge coordinated price drops in all types of assets.

This approach dodges the basic issue. If correlation does increase in a bear market, and appears to do so in a way that planning models don't predict, what is the point of using such models? They are, after all, meant to help us manage risk and on this basis they have failed. It is no good saying that extreme events cannot be predicted, and then applying a model that is based on investment probability. It is worth repeating that the probability of the events of the past nine months is so far outside the predictions of conventional models that they say the universe would have to exist for several times its expected duration before such events occurred again.

I have also drawn planners' attention to a compounding factor in the disaster. Analysts using conventional methods to estimate the 'equity risk premium' (ERP) - the return over and above the risk-free rate that you earn from owning shares, an average of 3.7% per annum over the past 109 years - have over the past 20 years consistently said that in future the ERP should be lower than in the past. Their conclusion is derived from Chicago school financial theory, and I have always been sceptical about it. But many planners used a lower ERP as an input to forward-looking models. The consequence was that in order to generate the return you wanted, you had to allocate a higher proportion of capital to risky asset classes such as equities and property. Yet the model also justified you claimed the risk wasn't as great as naive intuition suggested.

So, to sum up, the charge against asset allocation planning methods - especially when supported by stochastic modelling techniques that mathematise probabilities in a mechanical way - is that they underestimate potential returns and risk from risky assets and therefore allocate too much capital to them.

A few asset allocation practitioners in America are starting to formulate different responses. None have yet had the courage to throw the whole of finance theory overboard, but in my opinion this is only a matter of time. The Value At Risk models used by banks in accordance with the Basel II guidelines proved a disastrous failure. This should bury once and for all the belief that conventional finance theory has anything useful to offer regulators, and the same goes for investment planners. All their tools are based on the same discredited ways of measuring and predicting risk.

Indeed, some of the smarter non-conventional economists are now proposing much simpler common-sense measures for bank regulation, and US asset allocation practitioners are heading in the same direction. Rather than rely on theory, they are proposing giving more weight to history. For example, the Barclays Capital annual Equity-Gilt Study gives a range of the ERP over various time periods and using this and making a realistic allowance for catastrophe risk would result in smaller allocations of capital to risk assets for most investors.

The other lesson practitioners are starting to draw is that dynamic rebalancing is necessary. This means every six months selling some of those assets whose price rise has resulted in them accounting for a larger share of your capital, and buying assets whose prices have fallen. Even this wouldn't have stopped you feeling pretty glum at the moment, but at least you would have sold a chunk of shares in 2006-07. Add a tweak whereby the model increases allocation to cash after a period in which returns from most asset classes have been rising, and you have a methodology that looks a lot more robust than the maths-heavy, commen-sense-light version for which 2008 was, hopefully, a terminal event.

You can see all of Chris Gilchrist’s articles at www.TheIRSReport.com.

Chris is the editor of The IRS Report every month.

Call 0800 756 5437 or click here for more information.