Jim revisited some of the lessons in his original book during a recent presentation at the Fortigent Winter Forum in Savannah, Georgia.
Some take-ways:
- To make inferences, investors should look at performance across longer time-frames. He recommends 20-years.
- History doesn't repeat, but it rhymes: "It rhymes because human beings are the agent of changing prices on the stock market.”
- The market is mean-reverting; stocks will both over-perform and under-perform their relevant benchmarks, but return to a long-term mean.
- Composite indicators of several valuation metrics work better than any single one because the efficacy of single metrics go in and out of favour over time.
- Value composites work in the long-term, but momentum works best in the short term.
- Momentum works best when adjusted for volatility (you don't say ;).
- Over the 20-year period from the March 2000 peak, U.S. large-cap stocks will mimic their performance after the 1929 peak, and deliver approximately 0.3% per year.
Let's take a moment to examine this last point. O'Shaugnessy expected returns to the S&P 500 to average 0.29% per year over the 20 years following the March 2000 peak. In fact, since that peak the S&P has delivered total returns including dividends of 1.57% per year. But there are still 8 years left before his forecast horizon expires in 2020.
Out of curiosity, we calculated the implied average stock returns from today through March 2020 that would be required to manifest O'Shaugnessy's well researched forecast.
(1 + 0.0157)^12 x R = (1 + 0.0029)^20
R = (1.0029^20)/(1.0157^12) - 1
R = -0.016, or -1.6%
By implication, O'Shaughnessy expects returns over the next 8 years to average negative 1.6% per year. Not a very compelling case for stocks. Of course, this jives with our own quantitative forecast, which you can find here.
How do you invest when everything's expensive? Adapt!
h/t Katie Southwick at AdvisorPerspectives.com