Predicting Stock Returns for Fun and Profit

Beyond Buy-and-Hold #46

I say that stock returns are highly predictable. Most investment experts and most ordinary investors disagree.

We all love the returns provided by stocks. The one thing we don?t like about them is that they are risky. Risk is uncertainty. To predict stock returns is to know in advance what is going to happen to them. To know something in advance does away with uncertainty. So if we really could predict stock returns, we would do away with stock risk. That would be a pretty darn exciting development, would it not?

My aim with the words to follow is to distinguish two kinds of predictions and thereby explain how I can say both that next year?s stock return cannot be known in advance (I believe that it is this reality that causes most experts and most investors to believe that stock returns are not predictable) and yet that stock returns remain highly predictable all the same.

Does anyone know with a high degree of confidence who is going to win the Republican presidential nomination for 2012? Lots of people act like they do. There are hundreds of articles dealing with some aspect of this question posted every day to the internet. The reality is that no one knows. The reality is that this sort of thing cannot be predicted.


There are too many variables affecting the outcome of the various campaigns. Candidates will commit gaffes not anticipated today. There will be economic and foreign policy developments that will tilt things in favor of some candidates and that will do harm to the chances of others. Some candidates will gain momentum and rise to levels of support unimaginable for them today. Others will react to pressure poorly, turning what might have been small stumbles into ongoing declines.

Smart people can guess one or two or three developments. But the smartest political observers in the world cannot guess right on enough of the factors that affect the result to know in advance with a high degree of confidence who is going to win the nomination.

Next year?s stock prices are unpredictable for the same reason. There are hundreds of factors that affect short-term stock returns. Gas prices matter. Terrorist attacks matter. Who wins the Republican presidential nomination matters. Interest rates matter. All sorts of things matter. So I agree with you that next year?s stock prices are not predictable.

The reason why I take a different view about long-term stock prices (the prices that will apply 10 years from now) is that predicting long-term stock prices calls for the use of an entirely different sort of skill. I?ve been predicting long-term stock prices since 1996 and I?ve never gotten one wrong yet. Historical stock-return data is available on the internet going back to 1870. I?ve made year-by-year predictions from 1870 forward and, again, I?ve never gotten one wrong. You could do the same if you cared to. It?s not even a little bit difficult.

Long-term stock prices are not determined by unpredictable day-to-day events. They are determined by mathematics. Numbers are highly predictable. You can check them over and over and over again and they always turn out the same. I am more of a words guy than a numbers guy. But this is one quality of numbers that I greatly appreciate. Numbers, calculated properly, do not surprise us. Numbers are predictable.

Say that instead of asking you to predict who will win the Republican presidential nomination I asked you to predict whether the casinos in Atlantic City will turn a profit on their slot machines next year. Do you think you could offer a prediction with a good chance of coming through for you?

Casinos always come out ahead. Why? The odds are in their favor. You can beat a slot machine if you only pull the lever one time or two times or three times. Walk away after winning jackpot and you will remain a winner.

But you cannot remain a winner if you continue to pull the lever. It is a mathematical impossibility that you could do so. Pull the lever 10,000 times and you will walk away with the word ?Loser? stamped on your forehead. I can predict that. It is a certainty. It never turns out any other way. It never can.

That?s how it is with the stock market too.

You can?t predict what will happen in six months because six months is not enough pulls of the lever to make stock investing predictable. But you can predict what will happen in 10 years. At that point there have been enough pulls of the lever to insure that the price that applies reflects the economic realities.

The economic realities say that the U.S. economy is sufficiently productive that stock prices are going to increase by about 6.5 percent real per year. If you buy at a price far above that, you are going to see losses as the power of mathematics works to pull stock prices down to where they belong. If you buy at a price far below that, you are going to see huge gains as the power of mathematics works to pull stock prices up to where they belong. There?s no getting around this basic truth of stock investing. There is not one case in the historical record in which things did not work out this way.

Stock returns are predictable. Stock risk is optional. Those of us willing to look at the price at which stocks are selling before putting our money on the table thereby do away with most of the risk of stock investing.

I think that?s very cool. I hope that over time I will be able to get millions of middle-class investors interested in the idea of investing in this far more profitable and far less stressful way.

Rob Bennett challenges the conventional money management advice at his web site. Rob?s bio is here.

( Photo from Flickr by Helico )

Leave a reply