Modelling Choice and Mistakes
“Man is not a lightning calculator of pleasure and pain, but is instead caught between alluring hopes and haunting fears.” – Peter Boettke
If you want to claim the mantle of methodological individualism, you need to understand how people actually behave. A priori reasoning is not sufficient – you need to look out the window from time to time. Humans are animals. We have emotions, and limited capacity for thought. That’s not to say that humans are stupid, we’re not, but we are not computers, and we are not gods (altho a few lessons of econ 101 would go a long way to helping the denizens of Olympus get along with one another). Economists simply need to be honest about that, and not pretend like it’s a “gotcha” whenever people aren’t perfect. The Econ 101 model of human beings is not realistic. Get over it. The physics 101 model of motion isn’t realistic either; there’s friction and relativity and a million other factors, but simplification is necessary in order to understand the basic framework, for any science. On one hand, you can ask what is the simplest model of human beings which takes you to the conclusions of free trade and supply and demand, etc. On the other, we can ask what do we actually know about what people want and how do they act. These are very different questions with very different answers.
I’ve been reading “Thinking, Fast and Slow“, which is an excellent introduction to economic psychology by Daniel Kahneman. It covers most of the unusual patterns of economic behavior which have been discovered in the waves of behavioral economics research of the past few decades. It is a useful endeavor to improve people’s abilities to fulfill their goals, and part of that is highlighting common mistakes they may make. But showing that people make mistakes doesn’t imply that the whole decision making framework of economics should be thrown out. For starters, most of the experiments get the maximand wrong. People don’t try to maximize money, necessarily. They try to maximize how they feel. So if a gamble has positive expected value, but makes them feel bad, they will avoid it, and there is nothing wrong or fallacious about their action. Money only matters to the degree that it can improve your ability to satisfy other goals, such as meaning and happiness.
Finally, as a methodological note, if reality and your model diverge, your model is wrong. Obviously, in some sense this is tautologically true, but it’s more of a attitudinal statement. A lot of times economists seem to imply that because their model doesn’t fit how people actually behave, that must mean that the people’s actions are wrong. There is usually a good reason why people act the way they do, if you look hard enough. Possibly the two best economists who use this approach are Coase and Ostrom, who begin by looking at real world behavior first, and building the model second.
I meant to have one more point, which I forgot to add. Societies don’t act like individuals. If individuals make a variety of errors, that does not imply that societies as a whole do.
Aggregates – Errors may balance out. Some people get it wrong in one direction and others err in the other direction.
Survivorship bias – Those who make mistakes exit the market/activity and those who are left do not make the mistake. Firms which are more profitable get bigger and those who are not get smaller and go out of business.
Learning – Experiments typically only run the game once, leaving people to fail and never improve. In reality, most situations you face, you face several times and have an opportunity to get better.