Monday, August 27, 2012
I Am Mad At Economists (Part 1 of 2)
There's a whole bunch of things I'm not mad at economists about. Let's start there.
Sometimes people get mad over the whole "economics, it isn't a science" thing, but I'm not mad about that. Against economics being a real science it's sometimes said 1) that economists are bad at making predictions and 2) that you can't study people with scientific methods because their behavior is non-deterministic and unpredictable.
But with respect to 1), it's not obvious to me whether lame prediction-making means not-a-science or whether it means a science still in its infancy encountering massively complex phenomena. It's hard to predict the weather too, but that doesn't mean meteorology isn't a science. It just studies something super-complicated.
With respect to 2), it's also not obvious what the significance is of people's behavior being "unpredictable," because things can be unpredictable in the literal sense while still being deterministic and describable by rules. The brilliant eighteenth-century philosopher David Hume said that the world of human behavior was just as deterministic as the natural world, it's just that it was much more complicated. Yes, it's true that there are exceptions to every rule about human behavior. But that doesn't necessarily mean that behavior isn't described by rules. It could just mean the rules are really really complicated.
So that's not what's bothering me.
Other times people get upset about the way mathematics and formal methods are used in economics, but even though I think they might be onto something, I'm not mad about that either. Against the use of mathematics and formal methods, it's sometimes said 1) that we have no reason to think that obscure theorems in infinite mathematics would apply to real world phenomena, and 2) that formal reasoning in economics requires us to make obviously false simplifying assumptions, like that people are "rational" and have "perfect information" when they decide what transactions to engage in.
But with respect to 1), it's easy to say "obscure," but it's worth noting that economics seems to use standard mathematics from the well-established areas of real analysis and differential equations. These are the same theorems that we use when we design bridges, make airplanes fly, and calculate how to send the Curiosity Rover over to Mars. Even if there is something wrong with using the same mathematics in economics as one does everywhere else, it's hard for me to get mad about it. The math is just sitting there. It worked before. Why not give it a shot?
And with respect to 2), lots of, maybe almost all of, applied mathematical reasoning requires the use of obviously false simplifying assumptions. Natural scientists routinely assume things like frictionless planes, infinitely deep oceans and the like. Some of these idealizations work well and some don't; that's why people talk about the difference between having a good model of a situation and having a crappy one. In a way, the whole thing about math is that it's abstract and idealized. To apply it, you almost always have to make obviously false simplifying assumptions. As in 1), this methodology might be show, in the end, to be mistaken for the given domain, but its use doesn't make me angry. Again, what could be more natural than trying what worked before?
So this isn't what's bothering me either.
But now we are getting close to the heart of the matter. Because the first of two things that really makes me mad is the wildly overblown confidence with which the results are applied.
Look, if you know your science isn't so good at making predictions, and studies something very complicated, and if you know that you're using the tools that happened to be lying around the math and physics department and weren't really designed for the job, and if you know that you have to make obviously false simplifying assumptions to use those tools, wouldn't you proceed with caution?
Yet this is the opposite of what happens. The IMF strong-arms countries into adopting specific policies. Economists in the US, whatever they are recommending, pronounce their policy recommendations with total confidence. Graphs are produced alongside massively complex equations to show that certain short- and long-term results will follow from certain ways of doing things.
This would be like trying to send a manned mission to Mars before the effects of friction and so on were understood. Sure, you can use Newton's laws to calculate basic stuff like where the planets will be, if you make certain simplifying assumptions -- such as assuming there is no friction. And you could produce graphs alongside massively complex equations to show how your getting-to-Mars plan is based on the science of gravitation and planetary motion.
But you'd have almost no chance of success and every chance of disaster. Even if the graphs are good and represent something real, you'll never make it work. Because you just don't know enough about how your idealizing assumptions affect your outcome. You don't know enough about whether you have a good model. It's not enough to have a model. You have to have reasons for believing your model is a good one, an apt one for the thing you are trying to do, one that takes into account the things that need to be taken into account in the given situation.
We know roughly what it would be to proceed with caution in this domain, and policy makers don't do it. If you were going to proceed with caution, you might start by identifying the things you want most to avoid, and by distinguishing those from the things you'd like that aren't totally necessary, and then you might aim directly at protecting the former while you tinker with the latter. I take it that in modern economics, the things we'd most like to avoid are people's deaths from hunger and preventable diseases, closely followed by people's lives being completely ruined or rendered completely awful because they have to work 16 hours a day to make ends meet. And then after that would be people having big houses and fun gadgets, doing molecular gastronomy, and sending rovers to Mars.
I'm not saying these have to be the only considerations. Just that these are among the things you'd consider if you knew there was a lot you didn't know and you were trying to proceed with caution with economic thinking. You might also, contra the IMF, grant that other people might have different and equally good ideas about how to proceed, and that they have a prima facie right to put those ideas into practice.
Next week, part 2: the other thing I am mad about, the ethics bait-and-switch.