25 June 2015

The Stock Market as a Place Where Learning is Punished

One popular theory about the stock market is that its movement is random. This theory states that today’s prices reflect everything we now know. Only new information changes today’s prices. We don’t know if the new information will be good news or bad so we don’t know if prices are about to go up or down.

Maybe, though, there is a different explanation for stock movement.

Consider the possibility that the stock market is a place where - paradoxically – lessons are only beneficial up until the time they are learned. Once lessons are learned, they no longer apply.

That’s probably confusing but bear with me.

What happens during a bull market? First, a few investors who have purchased stocks after the bust experience great gains. Other people soon learn that you can miss out on big returns by not buying stocks, so they start to buy. Eventually, many people have bought stock.

Lesson that drives people’s behavior during a bull market?
If you buy stocks, you will make a lot of money.

What happens during a bear market? First, many investors who bought stocks at the peak lose a lot of money. Other people learn that stocks are a dangerous investment and best avoided.

Lesson that drives people’s behavior during a bear market?
If you buy stocks, you can lose a lot of money.

It is only once the lesson is learned that it no longer applies. Once people learn to be cautious, there is no reason to be cautious.  Once people see the benefit of risk, it is best to avoid risk. (Think about the 2008 financial crisis.)

It’s not that today’s information is reflected in today’s stock prices. If that were true, new information would change prices only incrementally and stock markets would not be so volatile. Instead, today’s prices reflect yesterday’s lessons learned. For a while. Once this lesson has spread to a critical mass of investors, though, it becomes obsolete. We reach a tipping point and at that point the lesson no longer applies.

This paradox of learning is not just a collective issue. It applies to individuals.

About 15 years ago, I bought a stock that I knew was high-risk and high-return. I thought that it had enormous potential but also knew it was really vulnerable. I told the kids that we would take a vacation on its value in a year. If it fell in value, we'd go camping for a weekend, If it took off like I thought it could, we would spend weeks in Europe. Well, it doubled. Then tripled. And then collapsed. The company went bankrupt and the value of my stock was not halved or reduced by 90%. It was zero.

Later, when I bought another stock that I thought had tremendous - but uncertain - potential, I used the lesson I'd learned.  The stock doubled. Then tripled. At that point, I cashed out my initial investment AND the amount by which it had doubled, leaving me with just a third of the shares I had initially purchased. Pretty smart, right? As of today, though, that stock is  up 15,585% from when I bought it and its management has announced that it will soon do a 7 for 1 split for this stock. (Yes. This is Netflix.) The lesson I learned from the earlier stock didn't apply to the next stock. Or more accurately, once the lesson was learned and changed my behavior, it was no longer a good lesson.

I’ve learned other lessons. I waited for Google’s seemingly inflated stock price to fall after the hype around its IPO in 2004. It rose steadily after its IPO and has never returned to that initial price since. I missed out on that ride by refusing to pay what I thought was a temporary blip in price.

So I learned the lesson that when a stock goes public with great potential it will open high and the keep rising just in time to apply it to a stock that opened high and then fell. Because of the lesson learned from Google, I plugged my nose and bought Lending Club at what seemed like a high price. Lending Club fell about a third from that initial, inflated price. Presumably other people had learned the lesson from IPOs like Google, a lesson that didn’t apply once people used it as a basis for paying what seemed like too much.

The next time I then decided to wait for the hype to fade, the high initial price of the IPO just kept climbing, rising 25% before I finally bought it. 

You might argue that stock movements are random and you might be right. But it might also be that the stock market is one place that punishes learning by changing in response to what we've learned the instant we've learned it.

You might even say that the real lesson learned is to do exactly the opposite of what worked last time but beyond the obvious problems (how do you actually define the last time? What is the opposite?), to the extent that this is actually learned, it will already be reflected in today's price and no longer be an applicable lesson. 

It is only the lessons not yet learned that work, which may be one reason that hedge fund managers like Jim Simons at Renaissance Technologies Hedge Fund have done so well by creating algorithms that detect patterns too subtle for us to learn. Even Simons couldn't explain why his algorithms found the relationships it did: he only knew these relationships (e.g., the link between yesterday's hog bellies price and tomorrow's value of the yuan) existed. Simons personally made $6 billion in income in just a few years using algorithms that broke the code on these obscure relationships that - apparently - he could never articulate in simple English. Why? Because in the stock market, it is only what hasn't yet been learned that is worth knowing.

No comments: