Robert Shiller has discovered a backward looking metric, which he calls Excess CAPE Yield (ECY), that shows a correlation with future returns. He discusses it in a blog post (www.project-syndicate.org/commentary/making-sense-of-soaring-stock-prices-by-robert-j-shiller-et-al-2020-11), but doesn’t go into that much detail. He does share the underlying data on-line.
Essentially, ECY measures how desirable stocks are compared to bonds, in terms of expected yields. Larger numbers are better for stocks, and ECY, as of March 2021, is 3.02%. I set out to determine if 3.02% is a “good” number or not.
More specifically, ECY = 1/CAPE – (GS10 – inflation)
CAPE = Cyclically adjusted price earnings ratio = S&P real price / (average real earnings from last ten years)
GS10 is the ten year US treasury yield.
Inflation is measured over the last ten years.
See docs.google.com/spreadsheets/d/126BtxbKgmsIQrTPLMqQ0bZmkDU7unx1s3q9O2AzgBU4/edit?usp=sharing -> Charts. The first figure shows ECY on the x-axis, and subsequent 10 year annualized excess return on the y-axis. Visually, there is a correlation that starts about 1915. The next two plots show the correlation more explicitly, first for 1881 – 2011, and then for 1915 – 2011. The second plot is cleaner.
I then did some back testing and experimenting, which is in the Data tab of the google doc. I found the highest return with the rule that stocks are more desirable if ECY from 3 months ago > 2.0%.
If I move a maximum of 1/12th of my portfolio each month (turning over the whole portfolio in a year) I get an annualized return of 8.73% from 1915 to 2021, compared to 6.78% for stocks alone.
- In January, ECY was 3.77%, which is significantly greater than 2.0% so now is not the time to switch to bonds.
- It is rare to get a signal to switch to bonds, and the signal takes a long time to solidify, but that signal should not be ignored.
- Past performance is no guarantee of future results. All back testing requires diligence about when and how past rules may break down in the future.
- I don’t take into account taxes or other transaction costs.
- I used 10-year treasuries and the S&P 500 as my benchmarks because they are both available all the way back to 1871, not because I think they are optimal investments.
Disclaimer: This information is only for educational purposes. Do not make any investment decisions based on the information in this article. Do you own due diligence or consult your financial professional before making any investment decision.