Tuesday, May 17, 2016

Someone Else Wrote It Better: On the Magic of Numbers and Subservience

I had planned to pen an overly-long blog post on this, but then I stumbled across an article here that makes the point I was getting after far more eloquently and with significantly more documentation. It's super and interesting and very much worth your time. So here's a short blog post getting at why I think it's so interesting and worth your time:

Not too long ago I was at a sociology conference on a panel with a well-known (well, for a sociologist) old timer making an argument about why economics is more influential than sociology (as evidenced by the number of times either is cited in the New York Times, a terrible metric for measuring such things, but that's a different argument). To drastically over-simplify his argument, he pointed out that while pretty much all sociologists are liberals (as measured by their affiliation to or ideological closeness to the Democratic party, a similarly terrible metric for measuring such things), while economics departments have a decent number of Republicans (though they'll still heavily outnumbered by Democratic-identifying peoples). As such, sociologists come across as partisan and biased, while economists appear to be objective and balanced, with a much greater diversity of thought.

To put it politely, I think this argument is dumb and wrong for many, many reasons, but two in particular stand out: 1) "right" and "popular" are often very different things!, and B) a major reason economists are granted more mainstream coverage than sociologists is that they've been better at cloaking themselves in the mantle of "pure" science. Mainly because they use more math than sociologists do (as a whole; plenty of sociologists rely solely on math in their research).

And this largely comes down to the fact that: A) math eduction in America is terrible, so B) most Americans don't understand basic math, so C) anything that has even slightly complicated math in it impresses most Americans, as they confuse complexity for meaningful knowledge or understanding.

You can see this in political horse race coverage: while all pundits are generally wrong in their predictions (and hilariously so!), those that use math, like 538, are seen as not just speculating, but scientifically predicting outcomes. Of course, 538 is just as hilariously wrong as everyone else, but look at all the complicated models they use! So complex! So many different letters and numbers! As such, 538 tends to be seen as sober political scientists rationally examining empirical evidence, other than just another set of political pundits spinning shit out of their ass.

And this is more-or-less how the field of economics has bamboozled so many, or in the words of this philosopher (which, again, you should really just go read): "world history tells a story of mathematical models masquerading as science and a public eager to buy them, mistaking elegant equations for empirical accuracy."

In this fascinating piece, we learn that apparently astrologers used to use this very same model of complicated mathematical bullshit signifying nothing to spin themselves as a legitimate scientific enterprise. At the turn of the previous century, many top leaders in politics and business used astrologers to "scientifically" plan their investments, production rates, etc. To make another snotty point about how much I disagreed with that presentation mentioned above, I bet astrologers outnumbered sociologists in the pages of the NYT during that period as well, but I also don't see that as a terribly big problem.

But in addition to the fact that this math-humping "imbues economic theory with unearned empirical authority," this article does touch upon the other problem I noted in that argument when it quotes a tenured economics prof on what influences the models he develops:
‘In economics and finance, if I’m trying to decide whether I’m going to write something favourable or unfavourable to bankers, well, if it’s favourable that might get me a dinner in Manhattan with movers and shakers,’ Pfleiderer said to me. ‘I’ve written articles that wouldn’t curry favour with bankers but I did that when I had tenure.’
And this, to me, is the real heart of the matter: basically, whatever a sociologist writes is not going to win them much political esteem or favor in the business world. However, an economist can add several hundred thousand a year to their bank account if their models just happen to be pleasing to those with a lot of money. Now, I'm not saying they simply invent complicated mathematical models that have no connection to empirical reality solely to please those in positions of wealth and influence, making them little more than modern court jesters, but...well, I guess I'm more or less saying that. But Dr. Levinovitz says it much better, so go read that.

1 comment:

revdj said...

A lot to read there! But what I gather, in essence, is that you are saying is that mathematicians should be better funded. I agree.