DoomYoshi wrote:I don't believe either author (the author of the book being reviewed or the reviewer) is trying to discredit economists as a profession. Rather, they are the attacking the assumption that economics is a hard science rather than a social science. Also, the idea that people are rational actors is ludicrous. People are 100% of the time acting against their own self-interest since they are 100% hurtling closer to death.
Let's unpack this in revese order. The fact that we live in a corrupt universe where death is inevitable is not our choice, so it says nothing about our rationality or irrationality. If we had a choice, I think most of us would choose to live in a more sane universe where entropy is either decreasing or at least stays constant. Alas, no choice was given. Point 3 down.
Point 2. We've been through this before. The idea that people are rational actors is only ludicrous because you assign this metaphysical meaning to "rational" which is something akin to "wise." That's a strawman. Chapter one of every social science textbook will tell you that we are not here to answer normative questions, only descriptive ones. When we say humans are behaving rationally, we are only saying that they are seeking the easiest way to get more of what they value at less cost. We don't have the power or the right to decide what they
should value. Whether they want more methamphetamine or more doctorates, our job is just to describe how they go about achieving those things. You think humans are crazy because more of them will pursue meth that a doctorate, and you're probably right, but it's outside the scope of scientific inquiry. You're trying to import this metaphysical definition into a place where it doesn't belong.
Okay, on to Point 1, which is the really interesting one. I think the conceit that physical scientists have that they are somehow better than social scientists is just based on the levels of certainly involved, which is a result of the sample sizes they are dealing with. When take flask of hydrogen peroxide and expose it to sunlight, you can say with a great deal of certainty that in time
t a certain percentage will break down into ordinary water and oxygen. You can say that with a high degree of certainty because you have a big flask with 6.022E23 molecules in it. But if I could somehow isolate just one molecule and set it aside, you could tell me absolutely nothing about whether
that single molecule would dissociate in the next hour or not.
Similarly, you cannot predict whether a single electron is going to be in a particular orbital. If you take a large number of atoms, you can say with a high degree of certainty that such-and-such a fraction of their electrons will be in the 1s orbital, and such-and-such a fraction will be in the d-orbital, and so on. But as the sample size dwindles, your level of certainty does too.
When economists deal with large numbers, they also achieve high levels of certainty. We can find with a high degree of certainty that if the price of cigarettes goes up 10%, then 2% of smokers will quit. That's okay when you're dealing with tens of millions of smokers, but many questions posed to economists deal with much smaller sample sizes. For many products there are 1-digit or 2-digit samples of producers. There are less than 20 major companies in auto manufacturing (ignoring small niche markets) yet we ask economists to predict how those 20 manufacturers are going to respond to a rise in steel prices, and we castigate them when they're wrong. Imagine asking a physicist to specifically predict the behaviour of a select group of 20 electrons!
7 billion people on the planet seems like an awful big number, but you would have to cube that number to get in range of Avogadro's Number. So I think a lot of this divide between the so-called hard sciences and the so-called soft sciences is just a difference in sample size. It's easy to get cocky when you can do an experiment in a controlled environment and with whatever sample size you need. It's a little harder when your environment is the uncontrolled real world and your sample sizes are limited to small numbers of actors in particular markets.