October 20, 2002
From the desk of Jane Galt:D-Squared (who has fled my
D-Squared (who has fled my comments section for some reason, and is greatly missed) has a nice item on the nobel prizewinning experimental economics work.
One of the great problems for libertarians is that the longer we study economic theory, the clearer it becomes that people are apparently incapable of rationally maximizing expected value in many cases. This is due to two things: first, we don't have access to all the information we need, and second, we don't always make decisions rationally. The great contribution of Kahneman and Tversky (whose work I highly recommend if you can get your hands on it; they're not so impenetrable as to be incomprehensible to the layman), was to test out the ways in which people made bad decisions, in order to try to reveal the processes that were causing them to engage in non-value-maximizing behavior.
Now, many on the left who had never heard of these folks are triumphantly saying "See! We told you so!" Not so fast, l'il ranger. First of all, it's not like professional economists are unaware of the work. They're working on incorporating it into theory. In some ways, some of the decisions which might appear sub-rational simply manifest themselves as preferences in the larger model which have already been accounted for: risk aversion, for example. Second of all, the great lesson of studying either regulation or people who try to make their living gaming markets, such as financial professionals, is that the technocrats are subject to the same decision-making problems as the mass of consumers. That means that while it certainly means we need a more complex model for markets than we're currently using, it also means that government regulation is not the automatic solution to the problem.
Take one phenomenon documented by Kahneman and Tversky: loss aversion. This is the interesting discovery that people fear loss more than they value gain.
Take one game, where people can either take a $50 sure gain, or a $100 possible payoff, with a 50% chance of getting the cash, and a 50% chance of getting nothing. Most people will take the sure thing.
Now, take a game where people can either take a $50 sure loss, or a $100 possible loss, with a 50% chance of losing nothing, and a 50% chance of losing the whole $100. Most people will choose to gamble. This tells us that people evaluate expected losses differently from the way that they evaluate expected gains: they will assume more risk in order to avoid a loss.
Now, let's look at a government agency such as the FDA. The FDA's ostensible mandate is to maximize health and safety for everyone. In a rational world, this would involve doing a cost-benefit analysis of the expected gains, weighing it against the expected side effects, and approving drugs whose expected benefit outweighed the expected risk.
Is this how the FDA actually works? Of course not.
Remember Thalidomide? Approved in Europe, not here. Not, mind you, because the FDA was prescient; they were incompetent, and couldn't get the paperwork together. This stalled the release long enough for the horrible teratogenic effects to become known. The FDA was a hero. And ever since then, they have been ever-more inclined to err on the side of not approving. I was shocked to hear a pharma person of my acquaintance say that he didn't think Prozac could be approved now -- too many known interactions with other drugs, side effects, what have you. I know that, for example, aspirin probably couldn't get approved now, but Prozac hasn't been around that long. That indicates a dramatic rise in risk aversion in the agency in a relatively short time.
Is this because the FDA people are mean, and want us all to die? Because they are stupid, and do not know how to weigh benefits against risk? Because they're bureaucrats, and just don't give a hoot?
No, it's because they're weighing risks far more heavily than gains. And in some part, this is a rational reaction to the fact that we are weighing risks far more than gains. If the FDA approves Thalidomide, a lot of people are going to get yelled at and probably lose their jobs. If they don't approve Prozac, are they going to get a horde of depressed people marching on Washington to roust them from their sinecures? No, the depressed people are going to be home in bed, eating ice cream and pondering whether two weeks is too long to go without showering. And most of us would probably say "better safe than sorry". The failures loom much larger to us than the successes. They mirror that.
So, no, this is not an excuse for the uninformed to decide economics is all bunk. You have to study for years before you're allowed to come to that conclusion, leave your PhD program, and move to Nova Scotia to become a lobsterman. Trust me. After you've done a couple of years in a department full of economists, you'll start to appreciate the lobsters more.
At any rate, it's interesting, interesting stuff, and we'll be thinking about it for years to come.
Update
Juan Non-Volokh has more on why proving that people don't always have perfect information, or behave perfectly rationally, does not thereby bolster the cause for government intervention.
Posted by Jane Galt at October 20, 2002 1:44 PM | TrackBack | Technorati inbound links