Beware Layers of the Pyramid

Intelligent Investor Funds analyst Matt Ryan recently sat the third and final exam for his CFA candidacy. It’s strange (and unfortunately common amongst academic institutions) that they continue to teach things that they know are wrong. The CFA course teaches Value at Risk (VaR), a once widely used modelling tool used by financial institutions to measure risk that was completely discredited by the financial crisis.

The model looks back over historical prices and tells you that you can be 95% confident, for example that your losses won’t exceed a certain amount (it could be 99%, or 99.5%, it doesn’t really matter). The problem, as everyone found out in 2008, is that it doesn’t tell you how bad the worst 5% might be. It also doesn’t warn you about anything that hasn’t happened before because it isn’t in the historical data. House prices hadn’t fallen since residential mortgage backed securities were invented, so the historical data model tells you the risk of loss is zero.

It’s a very limited tool in the real world, as the collapse of Lehman, takeover of Bear Stearns and JP Morgan’s recent US$2bn trading loss makes patently clear. The CFA institute know that. The shortcomings get a few paragraphs in the text book.  But they keep on testing it nonetheless. Teach what is testable, not what it is right, seems to be the way.

Anyway, once we flush the codswallop out of his system, there’s plenty of useful material that came out of the course. One of my favourites was an explanation of the shortcomings of goal-based or pyramid investing.

They are referring to an approach where people segment their portfolio to meet specific goals. For example, you might separate money for an upcoming holiday from your regular savings account. You might think of your superannuation as a separate pool of savings from your non-super savings and invest accordingly. Or you might separate the children’s education from your own retirement funds.

Along these lines, financial planners often recommend clients develop their portfolio allocations using a ‘pyramid’ approach, where the bottom layer of the pyramid is filled with the safest assets to meet the client’s most important objectives. Higher layers are then allocated to progressively riskier investments for the client’s less critical objectives until the portfolio is complete.

This approach is easy to understand and intuitively seems ‘responsible’. In fact, we probably all do it to some extent. It can, however, lead to a flawed overall investment strategy. By segmenting your assets you may not consider the correlations between these assets, which means you may take too much risk or be excessively conservative. An example will help demonstrate the issue:

Joe has $300k of personal savings, and in 5 years time when Joe retires he will be eligible for a pension paying $50k per year. Joe’s goal is to have $400k of savings as a safety buffer by the time he retires. Joe is comfortable taking some risk, however he has been told by a friend that someone his age shouldn’t have more than 30% of assets in equities and so allocates his savings $90k to shares and $210k to bonds.

So what’s Joe’s problem? By focussing solely on his $400k target when allocating his savings, he hasn’t properly considered his pension which is already a large position in bonds. His friend’s recommendation of 30% equities is probably not a bad overall allocation, but Joe’s true portfolio is now more like 90% in bonds, more conservative than what he was trying to achieve.

At the heart of this is an issue known as mental accounting bias, segmenting your assets and then treating them like they were somehow different from one another.

It’s far from an optimal approach to investing. Whilst different goals may have different levels of importance, it’s better to consider your goals collectively and then decide how to use all your wealth to meet these goals.

Unlike VaR, there’s an insight here that is worth keeping.