Uncategorized

Why Stock Prices Follow a Random-Walk Pattern in Some Time Frames and Not in Others

Valuation-Informed Indexing #250

by Rob Bennett

The book A Random Walk Down Wall Street contains a passage discussing research that purports to show that there is no such thing as “hot hands” for basketball players. When we watch a game and see a particular player sink basket after basket, we hope that he will get the ball again late in the game because we believe that the odds of him sinking a basket are greater than they would be in ordinary circumstances. The book argues that that’s not so.

The point that the book is making in citing this research is an important one. The point is that we fool ourselves into thinking that the stock market has hot streaks just as we fool ourselves into thinking that basketball players have hot streaks. What we seem to see with our eyes is not necessarily what is really going on.

A basketball player whose skill level permits him to sink x percentage of his shots will as a matter of random chance have time-periods in which he will sink a higher percentage of shots as well as time-periods in which he will sink a lower percentage of shots. The book’s argument is that the hot hands that we see in real life are what we should expect to see as the result of random chance. Players have good nights and bad nights. But there is no reason to believe that a particular player has a better chance of sinking a basket late in a game just because he has had a hot hand all night. His shooting percentage has been better that night not because he possessed any greater ability that night but only because random chance caused more baskets; random chance could begin to cut in the opposite direction at any moment.

I have never been persuaded by that argument. I see the logic of it. And I have never felt that I could deny the argument because I respect the science behind it. But my experience playing sports tells me that there really is such a thing as the hot hand. I have had times when I felt highly confident that I could hit a baseball or throw a bowling ball and then did just what I anticipated I could do. And I have had times when I did not feel such confidence and the results reflected that feeling.

Of course, such feelings are subjective. So, while I have never been persuaded of the argument advanced in the book, I have never felt sure that the argument is wrong either. I now feel that I can at least put forward a theory as to why the statistics show what the book asserts that they show with hot hands remaining a reality.

Say that a significant factor in whether a basketball falls through the hoop or not is the level of confidence felt by the player taking the shot. And say that confidence grows when a high percentage of shots taken over a period of time goes in. Having shots go in would increase the shooter’s confidence. So it would be natural to expect more shots to go in during the following time-period.

Hot hands are the product of increased player confidence. That means that they are real. You want the player with the hot hand taking a shot in the final seconds of the fame because he is the more confident player and thus the player more likely to sink the shot.

Why, then, do hot hands turn up no more often than you would expect as the result of pure chance, as the research shows?

It could be that, as a player’s confidence grows, his expectations of himself grow too. As expectations grow, the player becomes more vulnerable to the sinking feeling that follows when his expectations are not met. The player with the hot hand enjoys it for a time but the hot-hand phenomenon is doomed from the start. A hot-hand produces greater expectations, which produces a sinking feeling, which overcomes the confidence that caused the hot hand.

Hot hands not only disappear over time. The player comes to experience a time when he shoots worse than he did with his normal hand. Confidence leads to high expectations which lead to shattered confidence, which leads to a cold hand.

I am not saying that this is so for certain. I am saying that it is possibly so. It is an alternate explanation of the data cited in A Random Walk Down Wall Street, an explanation that suggests that hot hands are real.

I believe that something of this nature drives changes in stock market prices.

A Random Walk Down Wall Street argues that price changes are random. This is why we are told so often by Buy-and-Holders not to time the market. Random prices cannot be predicted. So timing is a bad idea.

But this claim has been suspect for many years. It’s true that short-term price changes follow the pattern of a random walk. But Yale Economics Professor Robert Shiller showed in 1981 than long-term prices (prices that appear 10 years out or 15 years out or 20 years out) are strongly correlated with valuations. If the Buy-and-Hold investing strategy recommended in A Random Walk Down Wall Street were valid, stock prices should fall in a random walk both in the short-term and in the long term. Why don’t they?

I believe in an explanation similar to the one I employed to explain why basketball players develop hot hands. Prices fall in the pattern of a random walk in the short term because they are determined by investor emotion and investor emotion cannot generally be predicted. However, at some point prices move higher than they should and the improper price jumps supply positive feedback for further improper price jumps. Like a basketball player enjoying a hot hand, the good results that investors see cause them to come to have increased confidence, which in turn causes even more good results and even more confidence.

Eventually a time comes when expectations of good results in the future have risen so high that they cannot be met. Confidence does not then return to normal levels but to sub-normal levels. Bull markets don’t produce normal markets. Bull market produce bear markets.

Buy-and-Holders look at short-term results and distant long-term results and see confirmation of their belief in an efficient market. But long-term results (results seen at 10 years out and 15 years out and 20 years out) tell a very different story. Those results show that, while short-term timing truly is a bad idea, long-term timing (changing one’s stock allocation in response to big valuation shifts with an understanding that benefits may not be seen for as long as 10 years), ALWAYS works and is ALWAYS 100 percent required for investors seeking to keep their risk profiles roughly stable.

Rob Bennett’s bio is here.