Imprecise Return Predictions Are Better Than No Return Predictions At All

Beyond Buy-and-Hold #95

Valuation-Informed Indexing offers investors higher stock returns at lower risk, relying on research-based predictions of long-term returns to work its magic. One complaint sometimes made by critics of the new strategy is: ?Those predictions are too darn imprecise!? It?s not an entirely unwarranted criticism.

The Stock-Return Predictor performs a regression analysis of the 140 years of return data available to us to tell how a broad stock index will perform over the next 10 years starting from any possible valuation level. Purchase stocks when they are selling at fair-value prices and you are looking at an annualized 10-year return of something between a negative 0.4 percent real and a positive 11.6 percent real.

That covers a lot of territory! We all would be thrilled with an annualized 10-year return of over 10 percent. We all would be greatly disappointed with an annualized 10-year return of a negative number. So maybe all this historical-data mumbo jumbo is not telling us anything of significance. The calculator is reporting that we might see a return we will like and then again we might not. Big whoop!

I think it?s a Big Whoop. It?s a limited sort of Big Whoop, I acknowledge that much. It would be better if we could predict returns with greater precision. But what we are able to do today (as a result of the last 30 years of academic research) is a big step up from what we were able to do yesterday. It?s so big a step up that it is my view that no one should buy stocks without first checking the P/E10 level that applies and what it signifies.

Here are my six rejoinders to the argument that the predictions of long-term returns that we are able to make today are not sufficiently precise to be of value to investors:

1) Imprecise Return Predictions Are Better Than No Return Predictions At All.

Buy-and-Holders don?t look at return predictions when setting their stock allocations. They invest blindly. How is that a good idea? When I put my retirement money at risk, I want to get my hands on as much relevant information as possible. Imprecise predictions certainly add value. So why not make use of them?

2) Longer-Term Predictions Are More Precise.

The 10-year predictions extend over a range of 12 percentage points of return. But we are able to make effective predictions that extend over a range of only 8 percentage points of return if we look 20 years out. And we are able to make effective predictions that extend over a range of only four percentage points if we go 30 years out. I focus on the 10-year predictions because most of us have a hard time focusing on time-periods farther out than that. But more precise predictions are available for those able to do so.

3) The Predictions Become More Precise If You Dismiss Outlier Possibilities.

The data shows that it is possible that an index fund purchased when stocks are selling at fair-value prices will provide a negative return of 0.4 percent real. However, the full truth is that there is only a 5 percent chance that we will see a return so poor. Consider only the most likely possibilities and you can make more precise statements. When stocks are selling at fair-value prices, there is only a 20 percent chance that the annualized 10-year return will be less than 2.61 percent real or greater than 8.61 real. That?s a range of only 6 percentage points.

4) Predictions Are Most Valuable When Prices Are At Extremes And the Entire Range of Possibilities Tells the Same Story at Such Times.

Most investors were afraid to invest in stocks in 1982. But the data-based predictions showed that the worst possible annualized return for stocks purchased at the price that applied at the time was 8.5 percent real. What difference did it make that there was a small possibility that the return might be as high as 20.5 percent real? Knowing that the worst-case scenario was so appealing told you all you needed to know to make a smart allocation choice. The same was so in 2000, when stocks were priced so high that there was only a 20 percent chance that stocks would offer an annualized 10-year return of better than 1.9 percent (and there was a small chance that the return could be as bad as a negative 7 percent real).

5) Even Imprecise Predictions Prep You Emotionally For What Is Going to Happen to Your Money in Days to Come.

Say that you determined that looking at return predictions is not for you because the 10-year numbers are too imprecise. You still would have benefited from examining what the P/E10 level that applied in 2000 told us about how stocks would be performing for the next 10 years. Investors who were familiar with those numbers were not even a little bit surprised by the price crash of 2008. Investing blindly sets you up for emotional upsets and emotional upsets set you up for panicky choices.

6) The Greatest Value of the Predictions Lies in the Comparisons They Make Possible.

We didn?t know precisely how stocks were going to perform from 1982 forward or from 2000 forward. But we sure knew that the odds of seeing good long-term performance were a whole big bunch better in 1982 than they were in 2000. That?s the value of the predictions. They provide a rough idea of how risky stocks are at any given time. That?s what you need to know to get your stock allocation right.

Rob Bennett often writes about indexing without the emotional baggage. His bio is here. For background on the Big Fail of Buy-and-Hold and on the need to move to Valuation-Informed Indexing, please check out the ?About? page at the ?A Rich Life? blog.

( Photo from Flickr by Helico )

Leave a reply