Basic Numeracy and the Division of Labor
What percentage of the “elite” class in society (to include politicians, celebrities of all stripes, lawyers, high-ranking corporate managers, journalists, etc.) would you guess qualify as numerate? As in, being comfortable in analyzing some basic math and statistics, or being able to digest multiple inputs and perform a comprehensive cost-benefit analysis of a relatively complex issue, such as, say, whether or not entire nations should lock down to limit the spread of COVID?
From what I’ve seen, it looks like large swaths of this group aren’t even familiar with the concept of a cost-benefit analysis, much less capable of performing one. If I had to put a number on it, I’d say less than 10% of the elite class could handle something like this.
Many find that answer to be surprising. They claim that such people (well, aside from the celebrities) surely hold graduate or higher degrees, and have extensive experience in organizations where such things are, we assume, done on a regular basis. But I believe that sort of shock reaction greatly underestimates the extent to which a highly developed division of labor, as well as the condition of “learned helplessness” have infected most white-collar work within the US.
It’s very critical to understand that there basically are no generalists anymore. Literally nobody is paid to just be pretty smart and pretty numerate and have a pretty good understanding of cause/effect, cost/benefit, etc. The closest you'll find is maybe a management consultant. I've been pretty interested to hear what McKinsey might have to say on lockdown policy, but they've been eerily quiet. Probably because the whole mess is so politicized and they simply can't risk alienating potential future clients (of which various governments all around the globe are some of the most lucrative) by saying something like "it sure looks like those masks aren't working" even when it is painfully obvious to them that this is the case.
And sure – there exist websites where people do this sort of thing. Stuff like Slate Astral Star Codex or “Zvi”’s blog or whatever. But think about how niche those communities are. Caring extensively about the minutiae of arguing multiple iterations of a cost-benefit analysis is an extremely uncommon trait. Some of us do it for fun, as a personal hobby, because we happen to find it interesting. But basically none of us (not even the authors of these blogs) are really getting paid for it in a "day job" sense. And in Scott Alexander’s case, he's all but legally forbidden from engaging in that sort of behavior in his professional life (imagine him running his own research and trying to prescribe a drug the FDA won't approve, or conducting his own psychological experiment without permission from 5,000 IRB bureaucrats, etc.)
I suspect over 80% of lawyers and politicians and corporate managers could not quickly/accurately explain the concept of a standard deviation. Not off the top of their head at least. It's not because they "aren't smart" though. It's because they literally don't need to think about such things - pretty much ever. In the rare cases in which they might need something roughly approximating “statistical analysis” to be done, there will be someone else who not only can do it better/faster/cheaper than they can, they are probably specifically mandated to use such a person rather than attempt to learn and/or muddle through it themselves.
Think about it. If we’re talking about advanced math that requires raw computational power, that’s all done by computers. If it requires pretty advanced math but just a little bit of intuition and human direction, it has probably been outsourced to India or perhaps Central/Eastern Europe. By the time the various pieces of the analysis reach the US-based team, whose job is supposedly to interpret them and make a decision, the decision is obvious. The math suggests a certain number which means the idea is either good or bad. There’s no real room for much disagreement – unless you disagree with one of the premises of the team who did the analysis, but by what right would you do that? They’re the experts on the math – not you!
So what does management actually do? Managers have ideas, mostly from intuition, which may or may not be informed by experience, then commission various sorts of specialists to put together various sorts of data analyses that then "prove" the manager's intuition to be correct. Or, in the rare case when the idea is just so abjectly terrible and/or the organization actually does support some amount of pushback, they get analysis suggesting the idea is really totally terrible and they scrap it and move on to the next one. After some period of time, they’ll look back and see how it went. If it went well, they will credit their own brilliance and management skill. If it went poorly, they’ll search desperately for some sort of unpredictable/unanticipated external factor they can blame the failure on. Think of all the various companies that had a pretty bad 2020. Approximately 100% of the CEOs of these companies are going to claim the reason they had a pretty bad 2020 was because of COVID. That seems plausible enough – COVID and its associated lockdowns were very destructive to nearly all businesses, and nobody really saw it coming. Of course, in the alternate universe where there is no COVID, it wouldn’t have been the case that 100% of companies would have had a great and successful 2020, now would it? Presumably some of these companies made bad decisions and had a bad year for reasons unrelated to COVID, but nobody is going to admit that, and almost nobody is in a position to prove it wrong (perhaps a small handful of sell-side stock analysts could, if they felt like it, which they usually don’t).
So when we say that the elites who rule our society are functionally innumerate and incapable of the slightest bit of critical thinking – it isn’t meant to be a criticism of their intellect as such. It’s not even a personal attack at all, so much as it is the inevitable result of a society that has outsourced all critical thinking tasks to non-elites (and often non-citizens or even non-humans). The generalist is dead – therefore we are ruled by “expert consensus.” And isn’t that working out just splendidly?