Yesterday’s blog entry set forth the reactions of Steve LeCompte (author of the CXO Advisory Group blog) to The Stock-Return Predictor. John Walter Russell (co-developer of the Predictor) set forth his response to Steve’s comments in the Comments section of yesterday’s blog entry. I have set forth my response below (Steve’s words appear in italics).
Changes in market participation, regulations, financial services industry offerings and communications technology may have substantially modified the distribution of equity returns over the past century (e.g., decreasing volatility of returns).
Speaking strictly, the Predictor would be better termed a Reporter than a Predictor. The calculator merely reports how stocks will perform in the future assuming that they will perform in the future at least somewhat as they always have in the past. It essentially takes all of the historical data and sums it up in a few easy-to-compare numbers. There’s great value in knowing these numbers. But it is of course important to know the limitations of the methodology used to generate the numbers. Steve is pointing to one of the limitations.
It is entirely so that the Predictor does not take into account how the stock market has changed over the years. There are two reasons why I do not think that presents a big problem for the investors electing to make use of the tool.
One reason is that some of the changes that regularly take place will cause stocks to perform better and some will cause stocks to perform worse. There is a good chance that the changes will more or less even out over the long run. Perhaps not. But those who doubt that the pluses and negatives will more or less even out this time should bear in mind that it is very much the case that that is what has always happened in the past.
The Predictor assumes a far-out long-term return of about 6.5 percent real. There have been numerous occasions on which the majority of investors had come to believe that conditions had so changed that the number would permanently be a good bit higher or a good lower than 6.5 percent. The reality, however, is that we have always worked our way back to price levels where that is the number that applied. Thus, I think it is reasonable to use that number as a starting-point for an analysis of how things are likely to work out this time too.
The other reason why I believe that the possibility that things may turn out different this time is not so big a deal is that there is no law of the universe stopping the investor using the Predictor from incorporating her own beliefs about how stocks may perform differently in the future from how they ever have before into her own analysis. Say that you are trying to determine how long it will take you to drive from one location to another. Say that on ten previous drives it has always taken you two hours to complete the trip but that this time you know that there is work being done on the highway. You know that the trip is probably going to take longer than two hours. It’s not right to say that your prior drives provided you with no useful information. Knowing that the trip has always taken you two hours gives you a good starting-point for your analysis of how long it is likely to take to drive this distance under the new conditions.
So it is with the Predictor. Those who believe that stocks are now going to perform worse than they ever have in the past need to subtract a little something from the numbers generated by the calculator to obtain the numbers they will use in their own planning. Those who believe that stocks are now going to perform better than they ever have in the past need to add a little something to the numbers generated by the calculator to obtain the numbers they will use in their own planning. John and I designed the calculator to report the numbers that apply if stocks perform in the future somewhat as they always have in the past not in an effort to rule out the consideration of other scenarios, but only because there is no possible means of addressing every imaginable possibility and assuming that things will go more or less as they always have in the past is the most neutral option available to us.
Simplistically, there is always a reasonably diversified subset of the market with below-average PEs. If PE is reasonably prescriptive for future returns, then an investor could expect to outperform your calculator by continuously restricting investment to such a subset (e.g., “value” funds).
My e-mail to Steve directed him to the Predictor but did not make reference to the Valuation-Informed Indexing approach to investing. John and I certainly do not say that investors are limited to the returns generated by the Predictor. Those are the returns that are likely to apply for those following a Passive Indexing strategy. We strongly believe that far better returns are available to those who elect valuation-informed strategies. Valuation-Informed Indexing is one valuation-informed strategy; the approach described by Steve is another.
Whether an investor should adopt Valuation-Informed Indexing or an approach more along the lines of what Steve describes depends on her level of sophistication. I do not doubt that there are strategies that will provide better returns than Valuation-Informed Indexing. The benefit of Valuation-Informed Indexing is that it is as simple to implement as Passive Indexing.
A big part of the appeal of Passive Indexing is that it provides middle-class investors a way to invest without having to engage in lots of research aimed at determining which stocks or stock classes will do well. The big downside of Passive Indexing is that it is likely to lead to massive losses at times when stocks are as overpriced as they are today. Valuation-Informed Indexing provides all of the upside of Passive Indexing without forcing investors following it to endure the horrible downside associated with the more conventional indexing approach.
I expect that we will see exploration of many alternative ways to invest in a valuation-informed way in days to come. It is my hope that the Predictor and other calculators like it that I expect will be developed in days to come will help guide all of those explorations.
There are other asset classes besides U.S. stocks and bonds. These other classes may provide attractive (even conservative) alternatives with respect to return distributions.
Yes, there are many asset classes available to today’s investor. The reason why we contrast the prospects of a broad U.S. stock index with the prospects of Treasury Inflation-Protected Securities (TIPS) is that these asset classes stand at polar ends of the risk spectrum. Stocks are a high-risk asset class. TIPS come with a government guarantee and an inflation adjustment. Contrasting these two asset classes has helped us develop many powerful insights into how investing works in the real world.
For example. we have seen that at the top of the bubble there was a super-safe asset class that provided a likely 10-year annualized return a full five percentage points higher than the likely return for stocks. That reveals the often-heard claim that taking on added risk always leads to obtaining higher returns for the nonsense that it is. There are times when investors are compensated for taking on added risks and there are times when they are penalized for doing so; it depends on how irrational most investors have been in setting the price that applies for stock purchases at the time.
The Predictor does not have as direct an application for investors investing in asset classes other than stocks and TIPS. But it provides guidance for nearly all investors. Knowing whether stock prices are headed upward or downward over the long term, and the percentage odds of various returns along the spectrum of possibilities, opens up valuable insights to those considering investing in just about any asset class imaginable.
There is evidence that financial returns follow power law rather than normal distributions (as argued by Mandelbrot and Taleb — see http://www.cxoadvisory.com/blog/reviews/blog12-17-07/). Power law distributions are “wilder” than normal distributions and offer hardly any confidence for regression-type forecasting.
As John mentioned in his comments, we certainly do not rule out the possibility that investors will see results not yet experienced in the historical record. We cannot let the possibility that things may play out in ways never seen before stop us from doing the analyses needed to learn how to invest effectively in the event that the far more likely scenarios are the ones that play out. The calculator does not come with any guarantees. It offers insights to those interested in developing a sense of the future possibilities in the event that stocks perform in the future at least somewhat as they always have in the past. As Steve notes, however, Mandelbrot and Taleb have argued that it may not be entirely safe to go with that assumption.
The reality is that the historical data tells investors who bought stocks in the aftermath of the huge bull to expect exceedingly poor long-term returns. The most likely 10-year return starting from January 2000 was a negative number and the worst-case scenario was a 10-year annualized loss of over 7 percent real. Yowsa! My view is that a black-swan outcome would need to be pretty darn black to be worse than the worst-case scenario reported by the Predictor.
But, yes, it could happen, and it is a good thing to call attention to this possibility. I think it would be fair to say that anything that serves to deflate irrational exuberance is a good thing so long as we remain at the sorts of price levels that apply today!
Today’s Passion: A community member named Aaron tells us that the investing approach that we have developed through use of the Predictor “is what investing should be — calculated, deliberate, confident, informed and simple.”