Most cakes are made from flour, milk, eggs, sugar, butter, bicarbonate of soda and something to add flavor.

To make a cake that tasted nice, you could try tasting the ingredients and then use them in proportion to how nice they tasted individually … but you’d likely finish up with a very strange cake.

Yet, judging by industry discussion/debate about which is the best performing asset class, at least some take this approach to portfolio construction.

This approach is indicative of a failure to fully understand diversification. There are two aspects to diversification. The first is splitting your bet. By having a series of small bets running simultaneously you minimise the likelihood of a really bad result (and of a really good result too.) The second is co-variance. By investing in a range of asset classes that perform out of sync with one another you can expect to obtain both risk and return benefits.

Here are two trick questions.

Suppose you knew which asset class was going to be the best performer over the next ten years. Would the best strategy have to be to put all your investments into that asset class? Suppose you added to an existing portfolio a new asset class that had lower return and higher risk than the original portfolio.

Must the new portfolio have lower return and higher risk than the original one? In both cases, the correct answer is "No".

If you are puzzled by this, consider the following hypothetical. There are three asset classes A, B and C. Over three years, A is the best performer - $1000 invested in A grows to $1,134 versus $1,129 and $1,128 for B and C, respectively

But what happens if you construct a 50/50 portfolio of B and C?

Not only does the diversified portfolio outperform each of its component asset classes, it also outperforms A!

Admittedly, the figures in the hypothetical were chosen specifically to illustrate the argument and it was assumed that the 50/50 portfolio was rebalanced* each year. But the basic point is that this result is mathematically possible … a fact that many would previously have had difficulty accepting.

The message for portfolio construction is that, to obtain the full benefits of diversification, it is not sufficient to consider the potential performance of the component parts in isolation. Rather, they must be considered in combination with one another.

Arguments about which is the best performing asset class are, in fact, evidence that the participants do not understand diversification - the central issue of portfolio construction. Even if you knew which was going to be the best performing asset class in the future, the best result is probably not going to be obtained by putting all your investments into that asset class. So why bother trying to crystal ball the future, let alone argue over the past?

If you want your client's investment 'cake' to taste great, focus on how best to combine the ingredients rather than the ingredients themselves.

*Re-balancing strategy is a non-trivial topic in its own right and beyond the scope of this article.

Source: Retirement Weekly, March 7, 2008 (Vol. 6, No. 10)