When the Moon is in the Laffer Curve...

James Madison University philosophy professor Alan Jay Levinovitz might be a curious choice as an author of an Aeon article about why so much modern economic analysis -- to put it plainly -- goofs.

But as he notes in his article, his expertise in Chinese religion has given him a better than passing acquaintance with ancient Chinese astrology. And the way that discipline used rigorous attention to mathematical detail to obscure the wildest guesses calls to his mind the way a whole lot of modern economists seem to work as they try to predict things that will happen.

Economics is harder than some of the so-called "soft sciences" because it deals with things that can actually be measured. If a sociologist wants to measure what kind of family arrangement is best for children, there are a whole lot of fuzzy terms that she will have to contend with to get any kind of result. And even if she can come to some agreed-upon definition of her terms, she's still stuck with the reality that people are different and that a family situation onerous to some might be perfect for others. Certain situations are obviously bad, such as ones where children are neglected or abused. But what about, say, a family of athletes surrounding a child not all that strong, fast or coordinated? Would that child fare "better" in a studious family? Does the difference of interests really make a family a better or worse fit for its members? Hard to say, but the sociologist who wants to make claims about such things has to figure out ways to say them.

Economics, on the other hand, uses real data and outcomes. Prices go up, prices go down. Hedge funds earn money, banks collapse. Actual results can be pointed out. But economics doesn't step all the way into the hard science category because it's difficult to repeat its experiments as exactly as a physicist or chemist can. The same conditions can be set up, but even if the experiment uses the exact same people there will be differences that could affect the outcome of the experiment. Water's always going to boil at 100º C, but the same person is not going to approach the same question about how to best spend $50 the exact same way.

At some basic levels, economics does seem to say things that are pretty solid. Price really is a function of supply and demand, and changes in any of those terms will affect the others in some generally predictable ways. More complex situations, though, are not nearly as clear-cut and really can't be treated as if they are. Too many variables can shift the outcomes so drastically that prediction really is off the table.

But the advent of super-fast number crunching means that seriously complex formulas can be created that are supposed to account for all of the variables.  Levinovitz points out that astrologers developed their own complex math to try to account for the different variables in their systems as well. They could run some blindingly complicated formulas and come up with precise answers. All the while they overlooked that while their math was nice and precise, it didn't make a lick of difference because the stars and heavenly bodies' only influence on human beings' daily lives is the miniscule amounts of radiation we receive from them and the tiny pull of gravity they may have. And those, governed by the inverse-sqare law, are so small they might as well not even exist. All the math in the world can't actually tell you how Mars being in your house made you lip off to the officer who was only going to give you a warning for your broken tail light. Mostly because Mars was in its orbit and nowhere near your house or your car.

Highly detailed and specific math formulas don't help whiz-bang economics do any better, Levinovitz says, because they're still dealing with the randomness of human nature and other factors that can't be predicted.

It's a great article. I kind of wish Levinovitz had pointed out something that could help rein in the ridiculous claims -- a recognition that they're guesses. Some of them are better guesses than others, but guesses are still guesses and sometimes they're wrong. The little of my science classes that I remember seems to have suggested to me that one of the things scientists did was to test their guesses about why things behaved the way they did. Good guesses got looked at more closely, and bad guesses got pushed aside. A bad guess dressed up in good math isn't any better of a guess than it was before it got all tarted up with formulae.

But getting a Ph.D. confess that he or she might be wrong about a prediction is probably harder than figuring out why Mars has it in for you. So change may be awhile coming.

(ETA: I should point out, that as someone who holds a master's degree, I am only slightly less likely than a Ph.D. to confess that I am wrong.)