Wednesday, July 7, 2010

Metro 1, Econ 0

There's an old joke among free-market types where one economist sees a $20 on the ground and delightedly bends over to pick it up. The other economist scolds him for the effort, saying, "If there really were a $20 on the ground, someone would have picked it up by now. It must be an illusion."

Today, I head to the Metro for the evening commute. There are two turnstiles to enter and one to exit. Yet there is a huge line of people behind one of the enter turnstiles while the other is vacant.

I reason: if the other enter turnstile was working, someone would be using it. So I write it off as broken.

After I've waited behind 8 or 9 people for the one good turnstile, someone walks by and gets through the "broken" one without incident. I follow behind, through the turnstile that's been there all along.

So much for rational markets ...


Adam Gurri said...

OK, forget rational expectations for a moment; let's go adaptive expectations.

We use other people's behavior as a way to economize on the amount of information we have to gather. When we see a long line, as you did, many people naturally think the others are broken. Checking for yourself has the risk of losing your place in line and it turning out that the others are, indeed, broken.

So the first person to test the others is taking a risk, and one that everyone else who stays in the line can free-ride off of. If it's broken, they stay in line and lose nothing. If it's fixed, they can switch lines and benefit from the fact that the person was willing to test it.

I think when such lines form it's the freerider problem that really gets it to bottleneck.

BAM, how's that for a response? :D

Greg Finley said...

Great response! Though I did not have time to consider all the implications of the free-rider problem.

There's been times when there are two ATMs, and the other customers in line insist that one is broken. I admit that I don't test it out solely so I didn't look dumb in front of complete strangers (or even offend them).