The valuation metric (P/E10 — the price of an index over the average of its last 10 years of earnings) used in The Stock-Return Predictor to identify the most likely long-term return is research-tested. It has worked well for the entire 140 years of U.S. stock-return history available to us. For example, research by Wade Pfau, Associate Professor of Economics at the National Graduate Institute for Policy Studies in Tokyo, Japan, reports that Valuation-Informed Indexing beat Buy-and-Hold in 102 of the 110 rolling 30-year time-periods now in the historical record.
Please understand, though, that while we now have the tools we need to make statistically valid predictions of the 10-year return of index funds, we are not able to make precise predictions. We can identify a range of possible returns and assign rough probabilities to each point on the spectrum of possibilities. But when the most likely 10-year return is 6 percent real, there is a 20 percent chance that the actual return will be less than 3 percent and a 20 percent chance that the actual return will be greater than 9 percent real.
Moreover, the investor needs to hold the stocks he buys for the 10 years that need to pass before effective predicting becomes possible. In those years, anything can happen. Stocks can perform well for one or two or three years starting from a time of insanely high valuations (this happened in 1997 and 1998 and 1999). Or they can perform poorly for one or two or three years starting from a time of insanely low valuations.
Shifting Probabilities in Your Favor
The smart Valuation-Informed Indexer prepares not only for the most likely outcome but for all other realistic possibilities. And the smart Valuation-Informed Indexer takes into consideration the emotional hit he will feel if he shifts to a low stock allocation because prices are high and stocks perform well for a few years or if he shifts to a high stock allocation because prices are low and stocks perform poorly for a few years.
He does this by avoiding dramatic valuation shifts. It is generally better to maintain at least a small stock allocation (perhaps 20 percent or 30 percent) even when stocks are selling at the insanely high prices that applied from 1996 through 2008. And it is generally better to avoid going with a stock allocation above 90 percent even when stocks are selling at prices so low that it is impossible to imagine a bad long-term outcome (perhaps your imagination is too limited!).
The idea behind Valuation-Informed Indexing is to shift the probabilities in your favor. Start thinking that our ability to predict stock returns is greater than it is and you will almost surely hurt yourself. Moderate allocation shifts always work. Extreme valuation shifts rarely do.
What Is Fair Value?
A P/E10 value of 14 is fair value for the U.S. market. About the lowest P/E10 value we ever see is 7 (half of fair value). A P/E10 value of 28 (double fair value) is insanely dangerous. Every time we have gone above 24 we have seen a stock crash and an economic crisis in the following years. The only time prior to the 1990s that we went to a P/E10 value of 33 we brought on the Great Depression. In the late 1990s, the P/E10 value rose to 44.
The idea behind the Valuation-Informed Indexing approach is to keep your risk profile roughly constant. You must be willing to change your stock allocation occasionally to be able to do this. But there is no need for frequent allocation shifts. An allocation shift is needed once every eight years or so on average.
When the P/E10 value is 7, it is virtually impossible for it to drop any lower. Stock prices must eventually return to fair value for the market to continue to function (the primary purpose of a market is to set prices properly). If the P/E10 value remains at 7, you would earn a return of 6.5 percent real for any money invested in stocks at that price. If the P/E10 value moves in the direction of fair value, your return would be better than that. So it is almost impossible to imagine a way you could lose on stocks purchased at so low a price.
The converse is true when stocks are selling at the sorts of prices that have applied from 1996 forward (except for a few months in early 2009, when we dropped down to moderate price levels for a short time). When stocks are selling at a P/E10 level of 24 or higher, the best possibility is that they might remain at that P/E10 value for a few years and that you would earn a return of 6.5 percent real. The more likely possibility is that the price will crash within few years and you will see big losses. The long-term probabilities are very much against investors buying stocks at such high prices.
Asset Allocation Strategy
The most simple strategy is probably the one that calls for a stock allocation of 30 percent stocks at times when the P/E10 value is above 21, 60 percent stocks when the P/E10 value is between 12 and 21, and 90 percent stocks when the P/E10 value is below 12. Pfau tested a 30/60/90 stock allocation strategy.
Valuation-Informed Indexing increases your long-term return dramatically. It is a virtual certainty that the Valuation-Informed Indexer will go ahead of the Buy-and-Holder sooner or later because Buy-and-Holders get killed in crashes and we all are certain to experience crashes at some point in our investing lifetimes. Once the Valuation-Informed Indexer goes ahead, it is almost impossible for the Buy-and-Holder to catch up because both strategies call for high stock allocations at times when stocks are priced to provide good returns. And, once the Valuation-Informed Indexer gains an edge, the compounding returns phenomenon continually increases the value of that differential over the remaining years of the investor’s lifetime.
But the primary purpose of strategy is not to increase returns. The primary aim is to limit risk. Please examine the discussion of how Valuation-Informed Indexing limits risk put forward by Pfau in this Bogleheads Forum thread (the highlight is the chart set forth in the post put forward on February 28 at 9:45 pm).
What do you think of valuation-informed indexing?