Advisor Perspectives welcomes guest contributions. The views presented here do not necessarily represent those of Advisor Perspectives.
This is a follow-up to my September 1 article, and it explores the issue of how best to use the Shiller P/Es in setting asset allocation strategies. Reader Bryan Zink, CPA of Tampa, Florida asked “how the Shiller P/Es might be used to beat the 11.09% stock market average—for example, utilizing a portfolio that is 100% stocks until the P/E reaches a certain level, and then switching some percentage to bonds."
The purpose of the original article was to highlight the value of using Shiller P/Es because of their historic relationship to long-term stock market returns. I used an example to show how the P/Es might be incorporated in an asset allocation strategy, but I did not evaluate specific strategies. Here I extend my earlier work to show results for a selection of alternative strategies.
I want to offer a caution at the outset about the limitations of this kind of analysis. I'm using historic annual data from 1928-2008, and while 81 years may seem like a long time, it's not a lot of data points. There may be issues of statistical significance and related "data mining" effects—if you try a bunch of different things on a limited set of data, you'll likely uncover some significant-looking relationships that just happened by chance. Also, of course, there's no guarantee that the future will be like the past, so even having loads of historic data may not be enough to point the way forward.
With those cautions in mind, let's proceed. All of the conclusions in this article relate to the data shown in Chart 1 below. This chart shows the results for seven different asset allocation strategies which involve using Shiller P/Es to annually rebalance investment portfolios. For each strategy, the chart shows the following:
- ROR: Average annualized rate of return for overlapping 10-year time periods beginning in 1928.
- 10 Yr. Std. Dev.: Standard deviation of the ROR's measured over rolling 10-year periods.
- One Yr. Std. Dev.: Standard deviation of one-year returns. (Clients may be concerned about how much annual volatility they will experience.)
- Added Return: Difference between the expected ROR for the strategy vs. the expected ROR from holding an all-Treasury-bond portfolio.
- Added Risk: Increase in standard deviation compared to the all-Treasury-bond portfolio.
- Efficiency: Calculated as the Added Return minus 2X the Added Risk—based on a 95% confidence interval and measures the amount that downside risk is improved versus an all-bond strategy.
First let's focus on the all-bond and all-stock portfolios. The bond and stock average returns are 5.16% and 10.81% respectively. These numbers are lower than the 5.44% (bond) and 11.09% (stock) returns from my earlier article, which were arithmetic average returns. These are geometric returns based on combining 10 years of annual returns. The goal is to move the returns up close to the all-stock level, but keep the variability of returns as close as possible to the all-bond standard deviation.
