Best of the Fool 2000
By Randy Befumo
March 15, 2000
The following article originally ran as a Fool on the Hill article on July 3, 1996.
Do you have a favorite behavioral science experiment? I do. And no, it is not Stanley Milgram's pioneering work in the 1960s. You know, the experiments where volunteers were told by authoritative men in white lab coats to administer painful electric shocks to strangers -- and they did. Nope, that's number two. My favorite behavioral science experiment actually involves rats and explains a lot about why people use bizarre investment approaches.
The experiments center on how long it takes to extinguish a learned response. A rat was placed in a cage with a pellet dispenser. When the rat pressed down on a lever, a food pellet would shoot out. Scientists outside of the cage could control when the pellet could come out, however, adjusting it so that the pellet would come out when the rat hit the bar two times, three times, etc. They could even adjust it so that the pellet of food came out randomly, without any regard to how many times the rat pressed the lever.
Now, the first part of the experiment involved teaching the rat that pressing on the bar made food come out of the dispenser. The rat was left alone in the cage and would occasionally stumble into the lever, causing a pellet of food to come down. A few run-ins with the bar and the rat would figure out the connection. This was one of the criteria for the experiment, as the rat had to have the learned response taught to it before it could be extinguished, or unlearned. After the response was learned, the real fun part of the experiment began.
Now, the scientists had quite a few rats in these contraptions at their disposal. Once each learned the response, the lever-push and food delivery relationship was adjusted. Some rats got a pellet for every push, some got one for every three pushes, some got the pellet for many, many pushes, and then some were just randomly rewarded. They left the rats to keep pushing the bar for food for a significant period of time, each with his or her own particular push-metric. Then one day, the scientists stopped the pellets from coming altogether.
This was the crux of the experiment. How long would the rat keep pushing on the bar before it gave up and realized that no more food was coming? How long would it take for the response to be unlearned? Well, most of the experiment is pretty intuitive. The fewer times the rat had to push the bar to get food, the quicker it figured out no more food was coming. The rat who only had to push once gave up first, then the rat who had to push two times, then the rat who had to push three times, and so on.
The relevant part to the world of investing, however, comes in with the rat who was rewarded randomly. This rat just kept pressing and pressing and pressing. Because it never learned any set schedule for reward, it took the longest to figure out that no more food was coming. Consequently, the rat who was intermittently rewarded developed the learned response that took the most time to extinguish.
What does this have to do with investing? It's pretty straightforward. Behaviors that are intermittently rewarded are the most persistent -- thus the success of the myopic science of technical analysis. Completely ignorant of the actual business dynamics represented by the company behind the ticker, the technical analyst relies on charts of past price movement that he swears foretell future stock appreciation. It's much like driving a car by looking in the rearview mirror. Nonetheless, technical analysts are occasionally right.
The problem is that they are right on an intermittent basis. Which means they develop a learned response that becomes darn near impossible to extinguish. All kinds of justifications, ranging from the sublime to the ridiculous, are given for why past price performance determines future stock appreciation -- in spite of the fact that Malkiel's tremendous book A Random Walk Down Wall Street clearly proved that no chart pattern had any predictive value above and beyond what was possible through pure chance. Even technical analysts will admit that their "science" is really more art than system, saying that experience informs the reading of charts.
Others might say that what is really going on is that they are still getting those pellets on an intermittent basis.
Specials Message Board