Standard asset allocation practice (“strategic asset allocation”) produces unstable results.

Remember Harry Markowitz, father of Modern Portfolio Theory which strategic asset allocation is based on, describes two stages in his Nobel award winning paper:

The first stage starts with observation and experience and ends with beliefs about the future performances of available securities.

The second stage starts with the relevant beliefs about future performances and ends with the choice of the portfolio. This paper is concerned with the second stage.

Portfolio Selection, 1952

If strategic asset allocators need to have “relevant beliefs about future performances” then why not start with “observation and experience” and assume stocks will not always outperform bonds for example.

Why not do away with the practice of building return assumptions on index return averages over a century? No individual investor has a long-term 100 year holding period.

In Nobel laureate Robert J. Shiller’s ground breaking research on PEs and returns, he found that over the long-term (which he defined as 10 years), US stock returns were higher for stocks with lower starting PEs and vice versa.

Independent investment researcher, Crestmont Research illustrates this fact:

Therefore wouldn’t it be better to take the average returns at various levels of PE to develop the expected return on stocks?

Just like strategic asset allocators commonly use the current yield on a bond, such as a 10 year treasury, as the expected return on bonds. Why not use the current PE on the S&P 500, such as CAPE, to determine the expected future return on stocks?

Using CAPE, for example, seems to me to be a better representation of “relevant beliefs about future performances” than to assume an average or re-sampled assumption based on a century or more of stock data.

A strategic asset allocator should also have relevant beliefs about future correlations.

In their research supporting their paper “When Diversification Fails,” T Rowe Price fund managers found:

One of the most vexing problems in investment management is that diversification seems to disappear when investors need it the most. We surmise that many investors still do not fully appreciate the impact of extreme correlations on portfolio efficiency—in particular, on exposure to loss. 

Correlations are the most unstable of the three components of strategic asset allocation. This is unlike standard deviation which is a more stable component.

So what is a strategic asset allocator to do?

Start by dropping the use of optimizers (and monte carlo simulators) if it’s impossible to forecast correlations.

Second, treat asset allocation for what it is. A balance between return and risk.

Third, remember it’s more of an art than a science which requires a little more reasoning and a little less MPT by defining risk as loss (not volatility).

Finally, among other things in life gaining and sustaining wealth requires prudence beyond all other measures.

Learning is to the studious, and riches to the careful, as well as power to the bold, and Heaven to the virtuous

Benjamin Franklin, The Way To Wealth, 1758