When experts go astray

When experts go astray

Big topics are too important to leave to the insiders

Matthew Yglesias,  Jul 1, Slow Boring.com

Alina Chan, the Broad Institute microbiologist who’s one of the leading lab leak theorists in the scientific community, said on Twitter over the weekend that she sees more support for lab leak among people with more quantitative scientific backgrounds and less so among life sciences people.

Twitter avatar for @Ayjchan

Alina Chan 

@Ayjchan

One thing I've noticed is that experts in the hard sciences (mathematics, CS, physics) afaik are more likely to think that SARS-CoV-2 has lab origins.

But more experts in the life sciences afaik seem to think that a natural origin is more likely.

Could it be a matter of priors?

June 27th 2021

10 Retweets99 Likes

Lindsay Beyerstein, who tweets a lot about how she thinks lab leak people are annoying and wrong, summed that up with the quip “the more domain-relevant expertise people have, the less likely they are to take lab leak narratives seriously” followed by a sarcastic emoji.

I think that exchange does a good job of encapsulating a running dialogue we’ve been having about who to entrust with Covid-related questions. After all, two or three years ago, pandemics induced by novel bat coronaviruses were not a hot topic in American society. If you wanted to find someone who could speak with any knowledge at all on that subject, you basically had to find someone from inside a specialist community of practice that thinks about epidemiology or that specifically studies viruses.

Today things are very different, and lots of people who didn’t know what “a coronavirus” was in June of 2019 now know a fair amount about the SARS-Cov-2 virus and the disease, Covid-19, that it causes.

The Beyerstein Thesis is that notwithstanding the greatly increased level of interest in the subject and consequent broader set of people who know something about it, we should continue to defer to the community of specialist practice that we would have relied on two or three years ago. Something similar went on with the reaction to Dana Goldstein’s column on Brown University economist Emily Oster and the role she’s played in the reopening schools debate. I said that what I found so useful about Oster’s work is that even though (or perhaps because) she’s not a public health specialist, she tried to frame issues in terms of explicit cost-benefit rather than binary debates about “is X safe?” or “is Y safe?”

Josh Marshall sarcastically summed up my view with “It would be helpful if someone with relevant knowledge said what I want to hear. But if they won't do it, I'll take an economist!”

This again poses the same question: Do we need to defer to the community of practice at the CDC, or should we turn to an outsider? Not, I would add, to a person lacking relevant knowledge, but to a person whose relevant background is that she’s written several books where she synthesized medical research findings on issues related to pregnancy and child-rearing and tries to distill them into actionable advice based on credible quantitative analysis.

I would say that Covid has, over and over again, proven the weakness of credentialism and reliance on insiders and the incredible value of smart troublemakers from outside established communities.

Lab leak and the biology insiders

The big thing that I think you need to know about the lab leak theory is that while in national politics it tends to play as a proxy battle for issues related to China, in the scientific community, it’s a proxy for battles about laboratories.

In May 2014, Olga Khazan wrote an Atlantic article headlined “Scientists Are Creating New, Incurable Diseases in Labs: Is that reasonable?”

Her point, obviously, was that it’s not. And similarly, in February of 2019 Vox’s Kelsey Piper wrote “Biologists are trying to make bird flu easier to spread. Can we not?”

And very early in the pandemic, Piper followed up with a forceful piece not arguing in favor of a specific theory of Covid origins, but simply that the fact of the pandemic indicates once again that the cost of this style of research is too high.

This is, in other words, not a new controversy. In fact, in the late-Obama years, the skeptics successfully got a pause on gain of function research put into place. But then in December 2017, it got unpaused. Marc Lipsitch from the Chan School of Public Health at Harvard warned when the ban was lifted that “there have already been accidents involving pathogens. For example, in 2014, dozens of workers at a U.S. Centers for Disease Control and Prevention lab were accidentally exposed to anthrax that was improperly handled. Another accident like that—if it involved a virus that was both newly created and highly contagious—has the potential to jeopardize millions of people.”

But the key thing about this long-running argument is that the side backed by Lipsitch and Piper and Khazan lost the argument.

This was, obviously, not a very high-profile debate at the time. It was mostly waged by insiders, and the insiders decided that they like funding this style of research. The result is that most of the members of this community of practice have colleagues who do the controversial research or receive funding from people who also fund it. And the question on the table is basically “did this community of practice make the massive error that Lipsitch warned about and kill the millions of people he warned would die?”

The fact that most of the community is skeptical of this charge is a relevant piece of information for the public to have. But given the nature of the charge, I’m very skeptical that we should view it as dispositive, as there is a very natural and obvious form of bias at work here.

Economists vs. the economy

I’ve been a big fan of Oster’s work since long before the pandemic. But in general, I think economists have often been a good source of information during the public health crisis. Not because “subject matter experts” have no value, but because economists are accustomed to quantitative reasoning, explicit cost-benefit analysis, and thinking at the margin, all of which are relevant forms of expertise for dealing with something like this.

But as I’ve written before, I think economists underrate full employment, which in some sense is closer to their core area of expertise.

One thing about this, though, is that you’ll see a couple of things happening if you pop open the hood of academic economics. One is that on a bunch of big, dumb, obvious questions people have about the economy like “should we tax the rich more or should we tax them less?” there is absolutely no consensus. The other is that all recently trained economists have high math GRE scores and know way more statistics than I do. So if you give an economist a statistical analysis question, it may take her some time to do the research, but she’s going to come up with an answer. Whereas if you ask an economist about a big picture economics question he’ll give you an answer fast, but Greg Mankiw and Gabriel Zucman will give you totally different answers.

On the specifics of full employment, what happened after the 1970s inflation experience is that economists promulgated a view that democracy naturally tends toward inflation.

The theory is that in the short-term, elected officials always want to err on the side of less unemployment, even if that means more inflation. They say you can worry about the inflation later. But there’s a “time consistency problem” where once it’s later, you still want to defer addressing inflation until some further point down the road. The solution to this, according to economists, was to create central banks that are staffed by experts (i.e., economists), pay higher salaries than normal civil service agencies, and also receive unusual levels of insulation from the political process — to the point where the president criticizing their decisions is considered inappropriate.

Now as a political scientist could tell you, this is a pretty sweet deal for macroeconomists! Except the federal reserve banks also expand their coalition by employing lots of economists who don’t specialize in monetary issues to just do random research papers on this or that. And the Fed is fully self-funded (they print money), so they don’t need to ask Congress for the cash to do this research. It’s a sweet deal for everyone involved.

But just as biology lab insiders are reluctant to conclude that biology labs are running reckless risks, economists are reluctant to conclude that this set of special arrangements might have some downsides. I don’t think it’s a coincidence that Jay Powell is not an economist — he’s not intellectually, emotionally, or professionally invested in the community of practice that looks at the sluggish recovery from the Great Recession and says “yeah, we did a great job.”

The military-industrial complex

Of course in all these cases, whoever has the credentials will pound the table in favor of credentialism.

But the reason I immediately pivoted from public health to economics is that I think recently, left-wing people have specifically gotten invested in credentialism. But people on the left tend to be skeptical of economists, and I think it’s important to remind them that there are some good reasons for that. Sometimes in life, you want an academic with high math GRE scores to look at a question. But they’re not like priests of “The Economy” imbued with the knowledge of how to generate prosperity — and especially not for countries like the United States that operate mostly at the frontier of technical possibility and really need to get richer by inventing new stuff.

Beyond social sciences, however, the 800-pound gorilla of righteous progressive skepticism of insider knowledge is the military-industrial complex.

When I was in college, I thought that invading Iraq was a good idea. And I thought this was a very sophisticated opinion to have, informed by my assessment that the subject matter experts at Brookings, CSIS, etc. tended to be much more supportive of the war than, like, random philosophy professors.

But two things turn out to be the case:

The “foreign policy community” that exists in D.C. think tanks and revolves in and out of government service is a distinct community of expert knowledge that’s pretty sealed off from academic international relations scholarship in America’s political science departments. The academic experts had a very different view of this than the practitioners.

The national security experts in D.C. have tons of detailed subject matter knowledge. But they are also deeply enmeshed in the military-industrial complex and fundamentally committed to an unsound intellectual project of global military domination. They can give you highly accurate answers to questions like “name some towns on the Syria/Iraq border,” but they give you wildly wrongheaded answers to questions like “why should America care what happens on the Syria/Iraq border?”

This is something that leftists have known forever about American national security policy, and that I am embarrassed to admit I had to learn. That being said, I think it is important to actually learn the lesson, which is not that foreign policy think-tankers are somehow terrible people, but that communities of expert knowledge develop blind spots, especially in areas touching on their self-interest.

Life without expertise

Now it would be convenient if we could sweep all these problems aside and just ignore communities of specialized knowledge on the grounds that they are likely to be insular, self-interested, and conflicted.

But the fact is they also know lots of stuff and knowledge is important. There’s a huge risk of ending up where Donald Trump did and basically saying that you can’t trust anyone and so you should just engage in a lot of wishful thinking. That’s how you get “the virus will miraculously disappear by April” or “hydroxychloroquine is a magical cure.” Trump’s instinct for wishful thinking did sometimes lead him to the correct answer — he was right that the vaccine development timeline would be faster than a lot of observers expected — but Trump couldn’t use occasional flashes of correctness to guide policy. On vaccines, he made a lucky guess in the context of across-the-board wishful thinking, so he didn’t prioritize the right stuff.

Trump got very jazzed-up about reopening schools, which I think was mostly correct. But before the pandemic, his administration was super-enthusiastic about online-only charter schools whose results have been dismal. There’s no insight here, just a desire to stick it to unionized teachers.

You need some kind of disciplined approach to thinking about things, not just opportunism and ideology. But reflexive deference to communities of expertise isn’t it.

How’re we doing?

This is a pretty unsatisfying answer on my part, but I think a big part of the answer has to be that actual results matter.

If it had taken five years to return to full employment after the Great Recession rather than over 10, I wouldn’t be nearly so critical of the economics profession’s conventional wisdom about labor market matters. If Iraq had been a stirring success, we’d be having a very different conversation about the national security community today.

With that being said, two related things strike me about the public health community during the Covid-19 pandemic.

One is that on several early big issues like masks and travel bans, they seem to have made significant technical errors. Relatedly, the community, qua community, seems to me to have not really been clear or fair in its communications with the public. The dominant public health view since last summer has been that Covid mitigation isn’t good enough and we should be trying to genuinely suppress virus transmission. That’s a fine view — it’s one I argued for in spring 2020 myself — but it tends to ignore the fact that the “let the virus spread unless it’s threatening hospital capacity” approach that Republicans favor just is the “flatten the curve” strategy that the public health community was recommending in March 2020.

But this all leads to the second point, which is that there is a rival expert community — the community of Asian public health experts — who got the mask and the travel restriction calls right and then articulated a consistent suppression strategy.

I assume the people doing this work on both sides of the Pacific are composed of well-qualified experts who have relevant degrees and experience. But communities that work somewhat in isolation from one another can end up developing different views. Ed Yong’s most recent big pandemic story for the Atlantic seems to me to convey the idea that America’s public health community has decided the real problem here is with the American people who are just too darn individualistic. But I think the reason the American people ended up defaulting to a more individualistic view of the pandemic is that the early leadership from the U.S. public health community wasn’t very good.

We should change our approach to lab research

This whole thing started as being about the lab leak theory, but I don’t really have any new factual information to share. My basic analysis of the situation goes like this:

If you hear about a new virus, your presumption should be that it’s zoonotic, since we have lots of examples of that.

If over time you can’t find the animal host, that should shift your view — especially when the outbreak originates in a city that also has a big virology lab.

So the decisive question ends up being your background level of belief in the safety and wisdom of this genre of lab research.

In other words, people are fighting about the lab leak in part because they care about its consequences for virology research. But it’s also the case that the plausibility of a lab leak hinges almost entirely on your prior belief in the plausibility of safety mishaps at labs — so of course the views correlate very closely.

The thing the pandemic has really taught us, though, is something about the upside to gain of function research. Recall that the point of doing this research is supposed to be to help us develop treatments and vaccines faster. But our actual vaccine development had nothing to do with that line of research. Instead, mRNA technology let scientists design vaccines very, very quickly just based on studying the virus once it emerged. The entire problem then became one of doing the clinical trials and manufacturing.

If we want to do better with the next pandemic, what we need is much more work on vaccine manufacturing throughout the entire supply chain. We don’t need gain of function research at all. We need more vials and syringes. Obviously “the experts” won’t like that conclusion, because they like doing their work. But sometimes you need to tune the experts out.