Who Watches the Epidemiologists?

 

I read a take down of Emily Oster that I think is mostly hot garbage; a lot of guilt-by-association written by conflict theory authors who view everything through a conflict lens and have no understanding of how others might view things through a mistake theory lens. In essence, the authors see people like Oster as the enemy in their conflict. Oster doesn't view herself as being in a conflict, but as someone challenging ideas as she tries to better understand the world.

I also think it follows a Covid-inspired trend of the growing hostility between epidemiologists and economists, which I will return to.

But there was one critique I found meaningful. In one of her books, Oster cast doubt on studies that show the effects of alcohol on pregnant women. She found the studies lacking, which reminds me of the ACX post “The Phrase "No Evidence" Is A Red Flag For Bad Science Communication”. Or the Nassim Taleb idea that “absence of evidence is not evidence of absence.”

As much as I trust and appreciate Oster’s writing, there’s no way I would read that and then tell my pregnant wife it’s okay to drink. (Granted, I have not read Oster’s book and don’t know the context. It’s possible she’s speaking to pregnant alcoholics who are trying to wind down their drinking and are worried quitting cold turkey would cause serious bodily harm. In that case, the risk of a small amount of booze might be the less bad choice.)

Instead, the authors invoked the precautionary principle, a very Taleb-approved idea. We know alcohol is bad for you. We know that whatever a pregnant mom consumes will affect her fetus. Just because we don’t have good data to prove the harmful effects of alcohol doesn’t mean we should throw caution to the wind.

Why Economists

I find value in economists, not as generating science, leave that to the actual scientists, but in evaluating risks and tradeoffs. But alcohol in pregnancy is one area I would leave to the medical profession.

It reminds me of one of my favorite arguments against a Richard Thaler take. If a doctor tells a terminally-ill patient that a treatment has a 90 percent success rate, they are likely to pursue it. If they say it has a 10 percent mortality rate, they will decline. Therefore, Thaler says you should always frame it in the positive.

The argument against this idea is that doctors know what they are doing. If they know the rates but have other information that suggests it’s not in the best interest of their patient, let them frame it in a way they think is best for the patient. But this introduces a method of power and persuasion that requires a lot of trust in the medical profession. 

Understanding Conflict

I appreciate what people like Oster are doing and I think her popularity is more of a demand than a supply issue; people are seeking her out because they are losing trust in the science community. She is trying to give people a better understanding of their options so they can make informed choices.

I also understand why scientists don’t like her. They believe that people should not be making their own choices when it comes to pregnancy and Covid, they should listen to the medical and epidemiological experts. But I can’t help that this feels like a “who watches the watchers” scenario. 

I think pushback and good conflict is a good thing. I think people with funnel-approved credentials should weigh in and offer their perspectives rather than staying in their own lane (which, for economists, I guess means forecasting GDP and unemployment rates ¯\_(ツ)_/¯).

The other issue at hand is the battle between paternalism (represented by the medical community) and individualism (represented by the economists). Give people too much freedom to make their own choices and many people will make poor choices and be unable to distinguish charlatans from funnel-approved intellectuals. This becomes a society-wide problem when hospital have to turn new patients away because they are at capacity with covid-infected antivaxxers.

Give institutions too much power to make decisions for you and you risk corruption.

The Risk of Risk

There is also an asymmetry between economists and epidemiologists. The medical community is overly cautious because their message is binary, ie safe vs. unsafe. And if they tell people safe, and people die, they will lose credibility. So the cost of being wrong falls disproportionally on the medical community.

Economists are more nuanced and view safety as a spectrum. If economists say "here are the tradeoffs and this is how you can calculate your own risk" and people die, they can just fall back on saying they knew the risks when making their decision. They have no skin in the game. So the cost of being wrong is less severe.

So my heuristic is skin in the game. Both Oster and Bryan Caplan say schools are safe and they both have their own kids in school, so I trust them. If a governor says masks are important and we should all follow the science, and is then photographed at an indoor crowded event maskless, I will distrust him.

It's a tricky balance. You have to weigh the opinions of experts and major institutions versus some version of "doing my own research." You have to stay open to going against public opinion if it seems right. And I don't always know when going against the grain is right. I'm still trying to figure it out.