Daniel Carpenter on Smart Regulation

Conversations with Tyler is produced by the Mercatus Center at George Mason University, bridging the gap between academic ideas and real-world problems. Learn more at mercatus.org. And for more conversations, including videos, transcripts, and upcoming dates, visit conversationswithtyler.com. Hello everyone, and welcome back to Conversations with Tyler. Today I am here with Daniel Carpenter. He is one of the world's leading experts on regulation. He is perhaps the leading academic expert on the FDA. He has a new book coming out this May called Democracy by Petition: Popular Politics and Transformation, 1790 to 1870, and he is in the Department of Government at Harvard University. Daniel, welcome.

Thanks for having me, Tyler Cowen.

Now, we will get to the AstraZeneca vaccine, but first let me ask about a few other topics. The Pendleton Civil Service Act of 1883, right? The goal was to replace hires by patronage with hires by merit. How should we be changing that act today?

Oh, good question. I think I would begin by saying that I'm not sure we need examinations, which, by the way, still exist in many areas of the civil service. That we should have perhaps minimum criteria, but the criteria should not be reduced to, say, a 4-year college degree. I do believe that the Pendleton Act created the basis for the decline of patronage, and it took a while. The Pendleton Act depended heavily upon presidential enforcement, and it was actually Theodore Roosevelt 2 decades after the act, more than 2 decades after the act, who really implemented the pivotal administrative decisions there. But I would say if we were trying to do this today, I would actually reduce the number of political appointees at the top levels of the civil service. Um, it's actually relevant to some recent politics regarding the FDA. Surely the leaders of an administrative agency executive branch or so-called independent commission, are appointed by the president. That's an Article II power. They should continue to be appointed by the president. It's all of the other folks, communications, undersecretaries, and so forth, that I think I'm not sure about the added value of those appointments, and I think we would do better by restricting their number.

But the regular appointees, say the G-14s, just people in the agency, How should the standards for their hiring and firing be different, or are we at some kind of actual optimum?

Oh no, I don't think we're anywhere near an actual optimum. I think it's worth pointing out, by the way, that the federal civil service, which we're talking about here, is quite different from the state and local level civil service. One of the things that merit reform did is it got rid of one form of politics by reducing the link between partisanship, party organization, and appointment at the middle and lowest levels, but it introduced another. It accentuated and facilitated the growth of employee— public employee labor unions. And I have a mixed judgment on those. I don't think you can have a conversation about policing like we're having in this country right now without calling into question the value of some of those unions and the degree to which they protect their members from egregious actions. Obviously, there's a huge literature there as well on teachers' unions and other public employees' unions. At the federal level, I actually think the federal employee unions are pretty weak, and part of that is just basic political science. Just about nowhere other than a few western states, and, you know, I guess you'd probably say parts of Virginia now, are federal employee unions politically strong enough to make a difference in the kind of electoral calculations of a congressional majority. So I don't know that we've reached an optimum there. I do think termination at some level should be easier for basic non-performance. I do think using career incentives among a whole bunch of positions in the civil service, uh, should be easier to do. But I would say at the federal level, you've got a situation where, again, I'm not gonna say optimum, If anything, I believe the last presidential administration showed us the value of having an independent civil service, one that cannot be used by a party or a leader who verges on authoritarian energies, while at the same time avoiding some of the problems that one sees in the worst case of public employee union abuses, such as we see in some police departments.

If one studies the history of postal supply and regulation, As you have done, postal supply. What lessons can we learn about regulating the internet today?

Oh, goodness. Good question.

Because the Postal Service was a kind of early internet, right? Remarkably cheap and fast by the standards of its time. Everyone had a voice. It was hard to censor.

Yeah, I mean, look, there are whole sorts of folks you could talk to about, you know, net neutrality and its value. I would say that our political culture in the history of the United States. And much of our economy and much of our education system presumed an open access, I mean, minimal cost for sending letters with all these cross-subsidies. So the letters subsidized the newspapers, which was crucial because the newspapers kind of created a much more informed society. It was a platform for advertising, things like that. So I don't want to go— I don't want to make a simplistic mapping from that to, say, net neutrality, because there are a whole bunch of other issues involved there. But I do think the degree to which providers can close off access or make access to this large externality much more costly is a grave concern. And for what it's worth, at some level, yes, we're there with the internet. Now, and at some level it's replaced the post office, but as I recently argued in the Boston Review, not entirely. There are still a large number of magazines, newspaper subscriptions, for that matter now pharmacy activities and others that rely heavily upon the postal system and probably will for the foreseeable future.

In the 19th century, how was the postal service central to the progressive reform movement?

I think it was seen as the— both the exemplar of what a potentially open-access public service could be in a way that nobody had really thought about any other function of the federal government being so universal. I mean, even consider the Army, which we would today say is, okay, how can you define a state without its military and its army? But as we know, Not really until the Progressive Era was the United States Army unified, and the Civil War was fought through militias. Any number of other services were, of course, decentralized to the state level, sometimes to the local. And so Progressives saw at one moment in the Postal Service, well, here's actually something that the national government can do, and it can provide this platform, this platform for, you know, a democratic republic, open deliberation, markets and a much more universalized cultural dialogue. I don't want to say singular cultural dialogue— pluralism was there— but a much more globalized dialogue. That said, I think the other thing they saw in the Postal Service, and this is something that Theodore Roosevelt saw in the Postal Service, was something that was being eaten from within by the patronage system. And it was in part the folks in the Pendleton Act, but I would argue even more the presidents who kind of really pushed civil service reform forward, forward with their administrative decisions, who said, if it's going to happen anywhere, it has to happen in the Postal Service, in part because there's a big part of our economy as well as our society that depends upon it. And Roosevelt, before and after he made these very controversial moves to cover in, as it said, thousands of postmaster positions under the merit system, actually received a lot of support from shipping industries and sort of medium-sized, smaller-sized retailers.

If we just outright privatized the entire postal system in America today, what is the most significant change that would result?

I think you'd be cutting off vast sections. I mean, truly privatized without a universal service standard, right?

Correct. No price regulation, nothing.

Yeah, we're not doing Ma Bell, you know, we're not doing— truly privatized, I think basically you would accentuate and accelerate the bicoastal concentration of culture, wealth, talent, and economic innovation in the United States.

Isn't that a good thing though, right, to induce people to move to higher productivity regions?

Uh, agreed. Um, I think there's all sorts of reasons to think that, you know, depopulation of some rural areas might be potentially a good thing. But I also think that there are ways to do that. We certainly have, you know, lots of those pressures right now. And there are also, I think, ways to do that that still— I mean, the concern I have is that, you know, you're cutting off a vast number, I mean, tens of millions of people from a part of their culture, again, that they still use.

But those people can use Federal Express today, right? Or UPS.

Yeah, but I don't think that's going to be, at least not until we get different models, I don't think that's going to be a viable model for express scripts. Express Scripts uses the postal system for good reason. They don't use Federal Express.

What kinds of records should the Postal Service keep about itself?

Great question. Well, there's a whole set of things that they don't since the Griswold decision and since the kind of First Amendment decisions. They don't keep as much records of kind of what goes through the mail. They can't prohibit things like pornography. Contraception. Well, I guess it depends on what you mean by itself. I would start with the idea that I think basic privacy restrictions, which governed the postal system as much through norm as by law in the 19th century and early 20th century, should govern the system. So, you know, it's a crime if I were to walk past your mailbox and open your letter. I'm committing a federal crime. But there were also norms that seals were not to be broken, things like that. So I do think whichever way the Postal Service goes, and it's quite possible that you could imagine an electronic platform for the U.S. postal system, I think basic privacy restrictions have to be guaranteed. I also think, actually, in some respects, I think we need to know a fair amount about what postal workers do without, say, calling for Amazon-like tracking. But if we think that postal workers are misplacing ballots, or not providing birth control pills or something like that, then, you know, we should probably have some way of picking up on that kind of nefarious behavior.

If we consider regulation in general at the state and local level, should we think of that as simply a one-way ratchet that basically just goes up over time and hardly ever goes down, and eventually gets to be too complex, too cumbersome, too restrictive?

No, I think that narrative is pretty easily falsified in the United States, as well as some other countries, definitely as a one-way ratchet. I mean, if we want to talk about pure monotonicity, take into account first that the degree to which we see price regulation, which was exactly what the Chicago School and other capture theorists were so concerned about so long ago, we've seen a whole range of areas where the government regulates prices far less— governments, I should say, regulate prices far less. We see far less regulation of antitrust M&A than we used to. I actually think that's a problem. And I'm happy to say, I think that case against price regulation, especially micro-level regulation of prices, is generally a strong one, the case against price regulation. We've seen a fair degree of deregulation in environmental spaces. And I would argue, depending on which part of the FDA you want to look at, even in medicines approval and things like that, we've seen some deregulation.

This is state and local level, right? If you look at, say, freedom to build, many more cities are putting restrictions on building than are liberalizing. And you might expect over time the entire Southeast will become a bit like, say, the California cities or Cambridge, Mass., for that matter, where they just don't want that much new to be built. Isn't that a one-way street, in essence? Yeah.

So I think, I think this gets to some of the differences between an economist and a political scientist. I, I see what you're talking about as a kind of regulation. Although to me, I see it much more as housing policy. And I think there are folks who say, well, let's try to construct a unified theory that puts the EPA installing smokestack scrubbers, mergers and acquisitions regulation, zoning regulations, because it's called a regulation, and say some set of price constraints and puts them into a unified theory. I think the point that you're talking to right now, whether it's— I think there's a development in local level valuation of housing in American culture that is leading many governments, I agree, to crack down on, on building, uh, you know, happens in my own town to make building new construction harder. And I could see that as ratcheting one way, except of course now let's talk about some of those very rural areas that we just mentioned a few minutes ago that are depopulating. It would be interesting— I don't know the answer to this question— to see whether there's more freedom to build in areas like that. My guess is where you're going to see less freedom to build, if you want to use that metaphor, is in precisely the kinds of places on the coasts and maybe in certain aspects of the Southeast, places that are driven in part by racial and ethnic discrimination, but also by wealth discrimination, that are going to crack down much more heavily upon new building, especially new residential building.

Which are the most and least captured federal agencies in the technical use of that term?

Oh, good question. So I don't, I don't have a rank ordering for you from which I can draw the top and the bottom. I think knowing some people who've been involved in the space And seeing also what's gone on worldwide, I would say one agency that has been captured or plausibly captured would be the Nuclear Regulatory Commission. And I think there's a complicated story there. The story in France, where they also have, you know, nuclear regulation, is that so many of the folks who work at that agency come through the country's formidable technical schools And they're trained in a certain kind of way that leads them to be much more likely to see nuclear power as a good thing. For what it's worth, I happen to be a believer in nuclear power as part of a bundle of non-carbon-based energy sources. That said, I think it's very plausible that due to kind of professional training, things like that, nuclear regulation, may be a plausible case of capture, both here and again in other countries.

And who has captured it in the US?

I think it would be some of the very power companies and interests there again.

But isn't that the case where the state and local governments are so restrictive, right? You cannot really get a new nuclear power plant built today. That limits the supply. That's the one-way street we were talking about earlier. And it's just harder and harder to access what seems to be a pretty good solution for a lot of climate change problems.

I think that's right. But let's, let's keep in mind that that restriction of supply raises the price in some equilibrium sense for those who supply nuclear power. And so it's not that I think— I think you're right that it's state and local governments. It's a certain degree of, of, you know, nimbyism on nuclear power plants. But I would also suggest that there are— it's not like you have a really robust coalition pushing for stronger, or shall I say, more laissez-faire attitudes on whether to build more nuclear power plants or not. Let me just flip it by one. I would say for the moment, probably the least captured agency, and I wouldn't say the least, an agency that is newer and less captured would probably be the Consumer Financial Protection Bureau. That's not to say that I believe in every decision it makes, I do think there's something to the way an agency gets set up, that when it's new, assuming it's set up in a certain way, it's got certain kind of staffers, it's got merit service protection, things like that, that it is less likely to be captured. And this is at some level a difference between the early political science theories of capture and the Stiglarian theory. Stigler— I'm about to participate in a 50th anniversary— it's another podcast with Luigi Zingales on the 50th anniversary of Stigler's theory of economic regulation. His theory was really much more about legislative design, and in much the ways that you and I have been talking about. But before him, there were two political scientists who were thinking in similar but not identical terms. One was my late colleague Sam Huntington, who actually wrote on the ICC, and the other was Marver Bernstein. And Marver Bernstein, who was at one point Dean of the Woodrow Wilson School, he talked about the life cycle of agencies. And so capture wasn't something that was designed kind of, you know, de novo into an agency. It was something that evolved in an agency that wasn't originally captured, but became so. I'm not saying that's a better model necessarily, but clearly sometimes that happens when capture by design does not.

If your least captured agency is a very new agency, doesn't that make you out to be a lot more sympathetic to the capture theory than your published work might make it sound?

Well, you have to remember my published work in the book with David Moss suggests that capture— the problem with capture is it's too sexy a term. David and I wrote during the midst of the financial crisis, when every time somebody didn't like the decision of a regulatory agency, a pundit an academic, the term capture would be thrown around. And you can think that an agency is doing a good job, a bad job, efficient, non-efficient, and capture can be potentially orthogonal to that. That's the first thing I would say. The second, and we really got hit, I'd arguably say more from the left than from the right on this, is we actually argued it's kind of hard to prove capture. You have to have a counterfactual. Of what an agency that wouldn't be captured would, in the same situation, ceteris paribus, be doing. And a lot of the existing empirical studies— and this goes back, you know, generations, and I'm, I'm indicting a set of things in political science as well as economics— suffer from the usual omitted variable bias problems. And some of them, I think, were making the wrong inferences from the evidence that exists, such as noticing, for instance, that in certain kinds of patterns of regulatory decision-making, larger and older firms did better, enjoyed, say, quicker or more favorable decisions than did smaller and newer firms. And as I, you know, wrote in a— was a mathematical essay published almost 20 years ago— that can be true even when the underlying regulator, or the regulator in question, has zero implicit or explicit preference over the kind of firm it's regulating. It has everything to do with kind of not even risk avoidance, but uncertainty avoidance.

How can we improve the process for public comments on regulations or pending regulations?

This is something I'm actually working on a bit right now. Let me say two things. The first is there's a paper that your audience should check out by my student Brian Liebgauber. He's a collaborator of mine too.— and what he did— excuse the digression, but I think it's a very important paper. So if you look at the Volcker Rule, for instance, it was a multi-agency rule issued under Dodd-Frank. Long before the comments rolled in, there were a set of hundreds of meetings with the agency that happened even before the notice of proposed rulemaking. And essentially, the architecture that we have in the Administrative Procedures Act doesn't regulate those meetings at all. And a legal scholar named Kimberly Kraviec— I believe she's at Duke, or she was when she wrote the article— Notice that the vast majority of these meetings, and I just go to the article, but I'm going to say 90% were with either large bank holding companies, so not the smaller regional banks, Fifth Third, things like that. We're talking JP Morgan, Goldman Sachs, Citi, law firms representing those companies like Sullivan Cromwell, and associations representing those companies. We get into a whole question about why the hybrid representation in another But that was one thing. And so pre-rulemaking, pre-proposed rulemaking, earliest agenda-setting meetings are dominated by industrial interests. The second, and this was an asset pricing study by Professor Liebgauber, who's now at UCSD, showed that the banks that met, though, with the Fed, and he only looked at the Fed, and then he used a very small window for when the proposed rules were announced. They witnessed and enjoyed a statistically significant and financially significant continual abnormal positive return on their stocks the moment the proposed rule was announced. Not so much the final rule because that comes much later, but the first. He totaled up, I think, $40 to $70 billion of positive abnormal returns just from those meetings with the Fed, just in the first hour, which is what an intraday study has to do after the proposed rule was announced.

So you're telling me this is good? Yes.

I'm not so sure. I mean, yes and no.

It's recapitalizing banks. It's making them more valuable. It means the trading isn't all that risky if the market—

So let's go back to your original question, though, Tyler, because your original question was about procedure. And to me, I guess I'd like to see open access. So I'm not by any means saying, hey, we need to get rid of these meetings. I do think we need to embed consumer and non-industry perspectives more. And what that might mean is inducing rule writers to more regularly meet with advisory committees, consumer, labor, other groups, folks like the group at Better Markets, Americans for Financial Reform, things like that. I'm not saying I agree with all of these folks, but I do think that right now all of our attention on notice and comment rulemaking, and for that matter, cost-benefit analysis as concomitant to it, misses the fact that there are people who are close observers of this space, and now a very empirically sophisticated observer in this space, Brian Liebgaber, who believe that most of the action is actually occurring before notice and comment starts.

If we take, say, nuclear power regulation, do you think that Giving all these groups you mentioned a bigger role in nuclear power regulation would improve the outcome or make it worse? It seems to me it would make it worse.

Well, why do you think that?

Citizens groups mostly don't want nuclear power. Homeowners organize. Try to build a nuclear power plant in Fairfax County, you have zero chance.

Asking whether I know whether the policy would be better or worse is a tougher call, but it might operate in some unpredictable ways. So nuclear power regulation, you might get Steven Pinker-like voices. He's a person who— Harvard scholar who wrote kind of a pro-nuclear power op-ed in the New York Times recently. You might get some other voices at the table, less voices, less organized, particularly if you embedded them in. And there are some— there's a great chapter in our book, Preventing Regulatory Capture, by Daniel Schwartz, who's shown that at least in insurance regulation in Minnesota, that Minnesota has been able to do this. Whether it's made folks better off or worse off, that's much harder to tell. But I would argue, in fact, that if you proactively embed voices, interests that wouldn't otherwise be represented, that you'll actually open up the conversation. My guess, if I had to think, would be the folks in Fairfax County, the NIMBY folks, the anti-nuclear power folks, they're already pretty well represented in this process.

Is the impact of FOIA laws overrated or underrated?

Depending on how enforced, I would say underrated. They have had one cost, which is it used to be the case— I don't want to say it's a cost, but they've changed an equilibrium in one sense, which is that used to be the case that kind of, uh, reporters had to cultivate a relationship with an agency insider, a bureaucrat, in order to get information out of that person. And what that led to sometimes was, of course, well, we didn't see as much, but it also meant that the reporter became better informed about what was going on in the agency. And there are stories you can tell, you know, reporters and mid-level agency folks who would talk on a semi-regular basis for years. Now, they kind of don't need to have that relationship to get the information they want. And so they just file a FOIA request and they get it. Now, that said, and again, going from the experience of the last few years, going from what I regard as a critical worry for our globe, for our country, kind of creeping authoritarianism, I think FOIA is— and strongly enforced, because it was just weakened by the Supreme Court— is one of the most important laws that we have.

In the area of regulation, the impact of ethics laws, overrated or underrated?

Just clarification, what do you mean by ethics laws, like revolving door?

That would be one example, but restrictions on people who work, say, in federal agencies. If you work at a financial regulator, you can't just let the people at Goldman Sachs take you out for a fancy dinner, right? That's an ethics law. We have lots of them.

I see, yeah. So I'm going to say again, it would be easy for me to say underrated or overrated. I'm going to say, you know, it really does depend. The revolving door constraints might be overrated. I think I worry a little bit less about people moving back and forth than others do. It turns out there's some evidence that banks and other firms, companies that want talent are most likely to get the regulators who are toughest earlier on, and so I'm not sure it leads to the usual kinds of stories. In terms of small favors, let me say that I think that's underrated. I think the psychological literature has demonstrated that small gifts are sometimes the most powerful ones.

Okay, now the FDA. You've written yourself that the main function of the FDA is often about information. Rather than safety. If that's the case, why don't we have what my colleague Robin Hanson called the banned product store? That is, the FDA would announce what it has approved, but, uh, say treatments or drugs or vaccines that had not approved, within reason, would be available for purchase. You just would know you were taking your chances. Why not do that? Have the FDA provide the information only?

Yeah, I think, uh, I would say two things. Number one, It's a brute mechanism, but the mechanism of what I would call approval regulation, which is basically the FDA as a veto player granting access to profitable markets, creates very, very powerful incentives for R&D, and those are incentives that we would otherwise not have. My worry is that in the Robin Hanson case— and by the way, I haven't read this particular case, but I've seen some, you know, FDA is Good Housekeeping seal of approval. We basically, you know, there are different ways of generating information. We want the good stuff. We want the RCTs. We want credible information. I'm happy to talk about Bayesian clinical trials, so-called real-world evidence. That's a term that's thrown around a little too much without definition. But if we're thinking about ways that, say, a pharmaceutical company can inform its potential consumers, and I mean consumers at all levels, so payers, physicians, nurse practitioners, pharmacies, you know, folks who decide on formularies, and then relevant to vaccination, the people who get the shots in their arms or whether or not they adhere to the drug. Well, they've got other ways to do that, and that's advertising. And so you could have a world in which it's kind of like a Lemons equilibrium, in which essentially you would have lots of bad products outcompeting good ones.

Why would they outcompete the good ones? So the UK approved the Pfizer vaccine. The US hadn't approved it yet. I would have the option of going to a special store and trying it. Right. We would have saved thousands of lives.

But that's, that's kind of a partial equilibrium story. I'm talking about a world where these institutions didn't exist, where neither the UK nor the US had anything like the FDA, anything like an approval regulator.

No, we have an FDA, but it issues judgments that are like a Good Housekeeping seal of approval.

Well, yeah, but then there's no— then there's no incentive or very weak incentives for anybody to do randomized clinical trials, which I would argue is one of the most important scientific revolutions of the 20th century.

You want to get the seal of approval, right? That's one reason to do a trial.

But why if you don't— if you can just go and sell it offline and then advertise your way into the marketplace? This is exactly the way the patent medicine industry worked.

Well, look at Tesla, all the R&D they've done to create a better version of their car, which they sell for a higher price. Let me ask you a different question.

But I think— Let me just stop you there, because I think there's an important difference. And I think, you know, you're an economist. I'm not professionally trained. My wife is, spent a lot of time around economists. I think the biggest difference here is that cars are experience goods, maybe even inspection goods, and drugs are credence goods. We can't really know whether or not, even if we were cured, whether the drug was the cause. That gets into a lot, and I've written about these kinds of issues and other things. So I don't usually buy the model of cars or other kinds of experience goods as mapped onto drugs, because I think the information structure of those markets is quite different. I think that the role of beliefs is far, far more important. Collective beliefs, if you will, Bayesian priors about whether or not such things work is absolutely central in deeply endogenous to the value of the product itself.

A doctor today can attempt a new method of surgery without getting FDA approval, right? Should we change that? And it seems to work fine, right? And there's plenty of RCTs for surgery.

Yes, although keep in mind that if the doctor rolls out a new medical device, which leads to a lot of surgical innovation, that device has to be approved by the FDA. There are, of course, you know, other kinds of bodies, and it's not just the tort system, but medical professional societies, things like that, that regulate that space very well. So when I say very well, very strongly, I'm not saying that's an optimum policy, but if anything, I'd say far more than in the case of drugs. In the case of surgery, it's the tort system that regulates that. And, uh, well, pick your poison. I'm not sure the tort system is— I mean, you could say that's a system that works well, Really? I mean, do we know that, you know, surgical innovation has provided that much more in terms of health benefits than pharmaceutical innovation? I haven't seen the aggregate accounting, but if it has, the form of regulation there is deeply decentralized, subject to all forms of bias, and that's juries. It's at the state and local level, not the federal level, right?

Take off-label prescriptions for drugs, right, which are a highly significant phenomenon, as you know. That's not subject to FDA approval. Should we change that? In essence, a significant percentage of total drug usage is not regulated directly by the FDA for the purposes it's being used for.

Right, right. Excellent. No, good, good question. You know, I've thought about this a little bit. I think there's a lot of room for flexibility in the current system. And one of the things that I try to say, I'm not, you know, I'm happy to criticize the FDA on any number of things.— and I'm sure we'll get a chance in what time we have today. But there's actually been a fair degree of flexibility. Maybe it's come slower than you might hope for in cancer, but cancer is kind of, you know, sui generis in many respects. Adherence rates are very high because a lot of the drugs are given just directly in the clinic. It's a community of researchers that— it's essentially highly closed. I mean, there are some, you know, doctors here and there who do things, but governed immensely by ASCO. There's a kind of a community of practice there. And so what the FDA does in a lot of these cases is allows for a molecule to be approved— say cisplatin would be an example, the first platinum-based chemotherapy— and then basically knows— I think the FDA knows that when a lot of these molecules are out there, there's going to be off-label prescribing. That's not to say, however, that off-label prescribing is not regulated. Whether something is on the label or not is still a significant decision. And I've written about the case of Bevacizumab for breast cancer. And that was a pretty significant decision by the FDA and one that a number of industrial interests fought against pretty hard, in part because they knew that money was on the line.

If we think about testing for COVID-19, You know, starting maybe with Paul Romer and Glenn Weil, who have been on this podcast, Michael Mina, a variety of commentators, Alex Tabarrok, now Ezra Klein, all have believed, and in some cases shown using measurements, mathematics, statistics, that if much earlier COVID-19 testing had been available, although imperfect, that the United States would have been much better off. And the FDA only now this week in the end of March is allowing such testing. Doesn't that verify the old-style Sam Peltzman argument that the FDA is too risk-averse, too slow? And when the FDA finally did let some of that testing through, you had to do it with a prescription, which seems makes no sense under any worldview. Right. So was Sam Peltzman right about the FDA?

No. So let me just say I agree with the criticism of the FDA on testing, and I Pretty sure I mentioned this to a reporter, you know, back in the spring. I don't know whether they ever— I've said this to a number of folks. For what it's worth, again, keep in mind that that's actually not the same division that approves drugs, which, by the way, is not the same division that Sam Peltzman critiqued. So I don't think it verifies Sam Peltzman. And it's not clear to me.

But the general view that the FDA is too slow, right? Whether it's Peltzman or not.

Yeah, but I hear you. But actually, Peltzman, for what it's worth, the original article was more about sort of number of new chemical entity introductions, and it was really much more about the regulation of the research process. And people like to say, well, is it FDA too slow, FDA too fast? A lot of that just came from Congress, from legislation, including the 1962 Kefauver-Harris Drug Amendments. But, you know, in this case, I think there's a very plausible case. In the case of testing, yes, I think there's a very plausible case that the FDA was too slow. Exactly what the public health, did that cause more deaths than, say, a culture that didn't take COVID seriously, people not believing COVID was a problem, vaccine hesitancy in the United States, which I think is pretty significant. You know, that's much harder for me to, or I think for anybody, to judge what the marginal contribution, but I have no problem in saying that, yes, I think that that was an issue. But look, we can always find a single example of where an agency was too slow or too conservative, or— and I think this is the other way— stuck to its scientific procedures too much without adaptability, which I think is one of the cases, one of the problems with the FDA and testing, and not think about all the other cases in which, no, the FDA was right to abide by a certain procedural conservatism.

But shouldn't there be a button within the FDA that can be pushed where the FDA goes into a kind of wartime mode. Let me give you another example. The AstraZeneca vaccine was approved by over 15 nations, including Canada and the United Kingdom. FDA still has done nothing. You might think now, much later, well, only approve it for men who are over 50 due to some data on thrombosis. But basically, it seems many responsible countries approved AstraZeneca. FDA did nothing. Had we had that vaccine earlier, would have saved thousands of lives. FDA didn't seem to care. Why isn't that an example again of the FDA being too slow and too conservative?

So I, I actually think the AstraZeneca vaccine is more complicated than the story that you're saying. So in part, it's what Dr. Fauci, now our national healthcare celebrity, but deservedly so, called an unforced error. So you should check out the Karen Landman op-ed at the Times back on March 23rd. As she shows, AstraZeneca has had managerial issues all along with production, with testing, early problems in data reporting, manufacturing problems in Belgium. Then there were study design and data reporting issues that were called by the company's own DSMB, the Data Safety Monitoring Board. So there are a whole bunch of folks— you could say, oh well, there's a whole bunch of folks who've approved it. Of course, as we know, in many cases in Europe, there are people who don't want to take it in part because of its bad reputation.

But that comes from the regulators itself, right? Macron badmouthing AstraZeneca.

I don't— I disagree with that. I disagree with that.

Back and forth on who should take it below 60, above 60.

I disagree with that strongly, Tyler. Yeah, I disagree with that strongly. No, again, AstraZeneca has had credibility issues all along. So the idea that vaccine hesitancy is just some regulatory construct or some byproduct, I think is number one, demonstrably false. And number two, I think it misses the big problem that corporate reputation matters. And AstraZeneca, I could grant that, yeah, maybe we should, you know, roll AstraZeneca out there. About the thousands of lives counterfactual, I'm going to disagree because it's not clear to me, and you might see this problem with Johnson Johnson, I clearly see it in a lot of conversations It'd be interesting to see survey data about whether people trust certain brands more than others. And part of it is vaccine nationalism. You know, the French don't trust the British. The British trust what's endorsed by their National Health Service, but also their own company far more, things like that.

Look, it's the same AstraZeneca company. So the British saved many thousands of lives by being prompt with AstraZeneca. Surely we don't think that is a mistake. We've had a supply constraint on the vaccine side in the United States, maybe up until now. We could have done the same had we approved AstraZeneca. So the company made big mistakes, but why penalize the people who can get the vaccine for that when, again, you could save thousands of lives? I just don't get it.

I just— to me, actually, I'm not sure that the AstraZeneca vaccine would have been taken in the United States. So you referenced the Ezra Klein story, and I gave a talk recently to a group in Israel and Europe on Regulation, where I said the big bullet we dodged in the past year was one in which there was all this pressure in summer 2020 on the FDA to quicken a whole set of therapies for coronavirus and COVID-19. And this obviously started with hydroxychloroquine, which we— I think the jury is in on that now, not effective. Convalescent plasma, probably very effective. But its rollout certainly didn't help. And what you see in at least two different surveys over the summer of 2020 is pronounced drop in expressed willingness to be vaccinated, anywhere from 10 to 15 percentage points, and deeply concentrated among the young. And the mathematical models would suggest that's exactly the population you want to have vaccinated sooner or later. Why? They're more mobile. They're much more likely to be vectors. So I've done some back-of-the-envelope calculations. I shared them with Ezra. I've also presented them to this group. And if you just, you know, we start to think about what's the social value of good vaccination just for the United States, right? And we think anywhere in the area of $10 to $20 trillion. I know some economists who think it's much higher. But now we think, okay, what if the vaccine— because you could talk about could have said this with vaccine, just about any vaccine. Pfizer's readout from its data wasn't until November, but you could have said, well, why didn't we approve that vaccine in May or June? And you would have had, I think, a social catastrophe, tens of millions of people not taking the vaccine, very consistent with those polls that right after the convalescent plasma debacle showed a 10 to 15 percentage point decline and possibly greater in vaccine confidence, and you would have lost thousands and thousands of lives that way. At the end of the day, vaccines are only as good— especially a treatment with positive externalities— vaccines are only as good as the set of population that takes them. And again, I emphasize, it's the young that you need for herd immunity. The marginal benefit of vaccinating the young is higher for herd immunity.

If we look at actual behavior, if you take the United Kingdom, which was the first to approve Pfizer, was quick with AstraZeneca, has a government that in some ways has a reputation for being reckless, right? They have arguably weaker approval procedures than the FDA. They've been one of the countries with the least vaccine hesitancy. They got vaccines out there. The results already are shown to be good. More people want the vaccine. They see that their uncle is not keeling over dead. Why isn't that a better theory of public opinion, that if you get people good vaccines quickly that make them healthier, more people are going to want vaccines?

So two things. Number one, I think that the UK is again kind of sui generis. In part, they trust their own company and they trust their National Health Service. They've got a very good public system, which if we recreate it in the United States would really dramatically transform that. Look, the correlation between vaccine hesitancy factors as expressed in public opinion and vaccination, first dose vaccination of eligible population in the United States is 0.6. That's an extremely high correlation. So rather than just use a story, let's actually look at data. And the data suggests, in fact, that public opinion on vaccine confidence is a very, very strong predictor of actual vaccinations in the United States. The other example I would say is part of my theory of the FDA is let's divorce it to some degree from vaccination more generally. Let's look more broadly at what happens. It's not that just that people said last summer, "Oh gosh, we don't trust Trump." It's that we don't trust certain kinds of procedural irregularities that are going on at the FDA, which of course are very deeply Trump-related. But I'll give you another example. It's a completely different example in a different world, but I think it's important. It's a tragic story. There's this deadly disease, Duchenne's multiple dystrophy, for which was developed a really important drug called eteplersen or Exondys, and it was approved under a very— and I say it's an important drug, it's an important drug therapy, it's actually still not known whether it works in many respects— but the FDA approval procedure was, shall we say, different and marked by much more controversy than usual. The advisory committee voted against it, the division director voted against it, Janet Woodcock, CDER director, voted for it. I believe it was then Commissioner Robert Califf who approved it over and above the other objections. And that's fine. He had his right to do that. But it was clear from everybody watching that this was not the usual approval process. So you say, nope, what's the problem? Like, now we've got this drug for a deadly disease that's on the market. Well, the problem is lots of insurers now won't cover it. Still won't. And the New York Times ran a story about this, and they got one exec to say, um, and I'm paraphrasing, you can look up the story, but I know that I just looked up a whole bunch of Blue Cross Blue Shield groups, still won't cover Exondis. They basically said, we don't know whether this works. And these are not folks— this is not vaccine hesitancy where you've got all the kind of standard cycle. These are informed purchasers, right? And so part of what happens when you have procedural irregularity or, you know, corners cut or bad evidence underlying approval decisions is you get, in some cases, what might be rational hesitancy on the part of treatments. I think we can say right now, with the benefit of hindsight, hey, look, we've got these great mRNA vaccines, we've got some other vaccines out there. Sure, if we had done things differently, that's right. But, of course, that's not what I would regard as a proper rational and Bayesian way to think about it. The question is, what decision should we have made at the time given the information we had? You might be saying, well, I want a value function that places heavy, heavy emphasis on the option value, that things could just go right here. I'm very much open to that. But what I want also is a world where the unknown, deeply unknown product studied or being potentially approved gets studied massively thoroughly and where the studies continue well after approval. And that's, by the way, something else that you miss with the Good Housekeeping model. You don't have any incentive for sort of continual study of things that are sometimes, say, like something like Prozac or a statin, are taken for 5, 10, 15, 20, 30 years of a person's life.

Now, you can often get better data when you do a human challenge trial, right? So do you favor those? Should the federal government have started those in, say, May?

Yeah, I can, I can see an argument for that. I don't have any— yeah, a challenge trial.

But that seems like a huge gain, right? It's not a small thing. Hardly any public health people came out for them. Derek Lau said he wasn't on board. Would have saved hundreds of thousands of lives if our public health establishment—

oh, I'm not sure about that causal estimate.

They'll say the vaccines would have been ready by October, right? The whole winter wave, we would have had significant protection against. And yet hardly any public health experts will come out for this.

So again, I'm going to put out a different counterfactual, Tyler. I wonder again, and a part of this is inseparable from the fact that we just had a really, really divisive election. But I actually think one big feature of the increased confidence in the vaccines is that they were approved after the election. I'm not saying they should have been, but I'm saying if you were to rush things out in September and October, which was precisely when vaccine confidence was at its nadir, it was actually much higher during the summer beforehand, but it had sunk again 10 to 15 percentage points. You would have had, you would have had a vaccine that tens of millions of people were not taking.

How should we improve our system of drug labels? Does anyone read them? Do they give real information? Can we make that better?

Actually, so there's an argument. So I think one of the findings out there is that black box labeling definitely changes behavior, sometimes for the worse. So there's an argument out there that when you added suicidality to SSRIs, that one of the results was that people were— psychiatrists and primary care physicians were prescribing them less commonly to teenagers, adolescents, and others. And that we had increased suicides as a result. I mean, so I think we need to think very carefully about— there ought to be— to call it a cost-benefit analysis, I think, is both right and wrong. I think there needs to be a prospective consideration of what the costs are of putting— the health costs of putting something on a black box label that would do it. The other thing I would say is the usual argument is that, what, you know, doctors only, what, 50 to 60% of the time read the label. There are some people who say, well, then the label beyond the black box really isn't effective. I actually agree. It's not that I'm generally a glass half full guy. I just happen to think that actually 50 to 60% of people reading the actual label is, is actually something that we should be pleasantly surprised about.

Should we allow for placebo prescriptions? For depression, for back pain, it seems they might work. We do already in a sense, right?

Yeah, we kind of already do. I mean, vitamin D was that way for a long time. Right. And so that's one of the, you know, one of the things that, again, you know, there's where we get into a different kind of form of regulator. You know, in part it's the FDA, but in part it's state medical societies and practice boards.

If major pharmaceutical companies want to run drug trials in poor countries, often it's Africa, but it could be Latin America, other parts of the world outside of the scrutiny of U.S. regulators. Should we allow that? Should we encourage it? Should we stop it?

I think it should be allowed, but I also think that the FDA is, or any other regulator, say it's Britain's regulator, uh, which by the way, I don't think is that, uh, laissez-faire. I think compared to Europe, if anything, European pharmaceutical regulation from which Britain has just only recently broken has long been at least as procedurally conservative as the US, say since the 1990s. I think it's entirely fine for some of these wealthier countries to say, well, we don't trust those trials, and we're not going to rely upon those trials alone in, say, a pivotal phase 3 or a really important phase 2 trial. You get into problems with what's the quality of a country's institutions, and this is something we usually think about from the standpoint of, say, somebody like, you know, Robert Putnam or Douglas North or Margaret Levy. But I think institutions and the credibility of scientific institutions is also important. Back when I wrote the book on the FDA, I spent a little bit of time in India and thinking about its issues in regulating drugs. And of course, there's a lot of drug manufacturing in India, a lot of generics manufacturing and so forth. I don't know the stats offhand, but, you know, the FDA has any number of times had to crack down on generic drug manufacturers in India, there have been a whole host of quality problems there. So if we think that a country has weak scientific and, and kind of, you know, credibility-based institutions, then I'm fine with allowing the trials there. But I'm also fine with regulators in advanced, so-called advanced industrial countries saying, yeah, we're not going to allow those trials, or we're not going to accept those trials as pivotal for approval. Approval.

Okay, we now get to your new book. Again, I'll repeat the title, Democracy by Petition: Popular Politics and Transformation, 1790 to 1870. It helped me think about a whole host of issues anew. First question, how and why were petitions key to the women's rights movement in the 19th century?

Great question. So I think if you think about the international women's rights movement, not just the United States, but also New Zealand, Australia, Canada, even France, Britain. In part, women saw themselves signing these petitions not just individually but collectively. They saw the weight of numbers, and they saw that especially white women, as they learned, actually, at some level, Native American women in the United States were petitioning in ways that were more innovative even compared to white women. But as they saw this, what I call the accountability of voice, the idea that their voices could matter, not only in the arguments they put forth, but also in the weight of the numbers. I think that became a much more powerful motivator for women's rights movements to make bolder and bolder claims for the vote.

Why was early 19th century French Canada so important for the history of petitions? This is the Hardcore Podcast, in case you hadn't known.

No, and by the way, your questions have been challenging and thought-provoking all along. This has been great. So yeah, I think actually, you know, we tend to think, oh, you know, what are the big petitioning societies? You know, the United States, Britain, Anglo-American societies, and in part, French Canada is that because it's still technically a colony. But I would say two things happened. At some level, it was kind of a Jacksonian revolution before Jackson. 1822 and even 1827 and 1828, before Jackson was president, you get what would be called kind of a producer's revolution in favor of parliamentary responsibility and parliamentary sovereignty. And at some level you could say, weren't they just doing what had already happened in the United States with, you know, the American Revolution? And I would say no. And let me be clear, what I'm talking about here is arguably, you know, the largest on a per capita basis petition of the Atlantic world of its time, a single petition with over 90,000 names attached, excuse me, 87,000 names, roughly 90,000 names attached that was signed in what is today Quebec. And with which the French Canadians did several things. First off, they fought off colonial fusion. If you ask yourself, why does French Canada still French Canada? And why is Canada bilingual? Kind of have to point to these petition campaigns as one of the reasons that colonial fusion was fought off. The other thing that they did, which was not easy to do at the time, was to depose a colonial governor. Even in British North America, the 13 colonies that eventually became the United States, getting a colonial governor deposed was much easier through elite mechanisms, going through the colonial office, going to the crown, than by massive protests. And that's exactly what they did with deposing Governor Dalhousie in the United States. Finally, as I mentioned, and there are other historians who have really, I think, cemented this story better than I have, Malcolm Chase, who's arguably the historian of the Chartists. Working-class reform in Britain depended heavily upon lessons that the Chartists and other working-class agitators in Britain learned from the French Canadians. They called them our brothers in oppression in the 1830s., and a whole host of things that happen as the suffrages open, as a whole bunch of changes happen in 19th century Britain, were taken with at least partial inspiration from French Canadians, the patriote movement, as I described them, and these petitions.

Now, let me ask you an analytical question, and please try to give me an answer for both the middle of the 19th century and then for today. So there's a bunch of different hypotheses you might have about petitions, that they strengthen the relative power of what is truly popular, or you might think it strengthens the relative power of groups who organize well, or maybe groups who are good on the media, groups that are good at fundraising. On net, overall, on average, for each of those two periods, what do you analytically see as the major effect of petitions apart from any particular case?

Yeah, excellent question, Tyler. So I think the 19th century is easier to answer because that's what I just wrote about. What I see is, particularly during this crucial period, say the 1820s to the 1850s, in the United States, Canada, parts of Mexico, people who didn't already have organizations, didn't already have parties, which of course in the United States especially were really important technological innovation, didn't already have lobbies, didn't already have interest groups, were able to successfully harness the petition to make their voices better heard, to set the agenda. I think agenda setting is really important when we think about democratization and also to organize. I mean, what I argue in the book is that everything from anti-slavery societies to new parties such as Canadian Reform Parties, the American Whigs, and others drew much of their energy and structure from early petitioning campaigns. Look, we're in a different world now, especially in the United States, where I don't think petitions have quite the societal and historical imprint that they used to. There's a whole bunch of reasons for that, but I want to answer your question. I would say now, in part, we— well, we don't have— let me just first bracket this by saying we don't have anything like the legislative petition process that we used to have in the 19th century. And by the way, Canada still does, and Canada still gets, I think, 1,500 paper petitions per year. Now, we could get into a whole argument about whether that changes anything. I think what you're going to see with petitioning nowadays, especially of the electronic variety, occasionally of the plebiscitary variety like the initiative and referendum, is that select groups and interests that know how to make something go viral— and virality, if I can use that term was there in the 19th century, now, but you have to make it get hot right away. I'm reminded of that, there was an electronic petition signed about 10 years ago contesting new card fees at Bank of America. I think it was by somebody in Chevy Chase. I remember the story was in the— and immediately, within days, hundreds of thousands of signatures, Bank of America backed down. I don't know, by the way, whether that's good or bad economic policy. Probably an effective petition from the standpoint of having changed things. And can you imagine any other institution, the courts, parties, through which that would have happened? Nowadays, we could say social media. Somebody could jump on the social media and say, "Hey, look at this awful card statement I just got," and then you get a bunch of likes. So it's some kind of aggregation mechanism like that that can do that, but I don't think we have the kinds of institutions that we used to, in part because we don't have a petition institution that structures robust public debate as much as we used to, and I think that's a problem.

If I look at state-level referenda, I tend to think it's a bad idea to make them too easy, California being Exhibit A. Do you agree with that view or differ from it? And if you agree with it, doesn't it mean you now think petitions actually have become on that somewhat harmful.

Yeah, so I do agree with that. I teach a course at Harvard called Res Publica, or it used to be called What Is a Republic, and I argue that I think, uh, it's problematic when not just, say, a policy, but any constitutional amendment, as has occurred in several states, can occur with just 50 plus epsilon percentage of the vote of the people who happen to show up on a given day. We know that some of these are small turnout elections. Let me just say that's about initiative and referendum. I'm not going to wax much more generally about, say, school financing or bond issuances, things like that. But at the level of California, yeah, I do have problems with those institutions in part because it weakens the separation of powers, it weakens deliberation. One advantage of representative government is that there are certain deliberation and investigation that go on. It weakens compromise, right? It's either we're going to have this ballot proposition or we're gonna have the status quo, and nobody gets to go into the ballot and say, "Oh, let's think about a third option between the two that might be better than both." The ballot and the plebiscite doesn't give you that option. I distinguish for that reason between plebiscitary petitions and what I call directed petitions. I also worry about plebiscitary petitions. And I don't think they have the same value, institutional robustness, and benefits that directed petitions, which are the kind of petitions that I write about in "Democracy by Petition," do. And so in part, the answer to the question is yes, I, you know, agree that I'm not that high on the initiative and referendum process. Ohio, by the way, if I can bracket, has a policy that might be a little better, which is to say, All right, we're gonna let the legislature pass a law. So we're gonna respect the Republican small r process, separated powers, things like that. We're gonna allow the people to veto it, but then the agenda is then set by the Republican process, mixed regime and so forth, not untrammeled, purely discretionary lawmaking power or constitutional amendment power, again, given to that 50 plus epsilon percentage points of people on a given day. And then, what I will just say again is, I think, in part, petitioning, I don't know if it's net harmful, it's definitely become less beneficial, in part, because of the way that plebiscites have transformed the petition.

What would be an example of a reform you would like to see that would make current American government more democratic, at any level?

Well, let me extend part of the logic of my book a bit. I actually think it wouldn't necessarily be such a bad idea— if legislatures were to reintroduce— I'm not saying this would be the top-ranked reform, but I don't think it would be such a bad idea if legislatures at all levels of government had to reintroduce petition days, such as governed the United States until the 1950s, and many state legislatures. And what it would mean is that— I think part of the problem that we're— you know, we've got institutions that are efficient at some level, If we want to take a mass of millions, hundreds of millions of people and condense their preferences down into a single platform and a single label, parties do that well. The problem is, is that so much gets left out of the process, to say nothing of the kinds of polarization that can end up really kind of breaking a society into two, which I do think we're at risk of in the United States, and of course, which has happened before. And so it would basically be a way to make not just the United States more democratic, but actually to make the democratic republic more deliberative. And it would be inefficient, I'll be honest, you know, and at some level you could say, well, yeah, wouldn't legislators, you know, be better off reading scholarship from economists and doctors and things like that and getting presentations from people in the know? You could argue, yeah, maybe they would, but I actually think it would potentially increase social trust. And by the way, New Zealand and Canada do this now.

Right?

Of course, you can always say, well, yeah, New Zealand and Canada do everything better. Maybe they do, maybe they don't, but it's not unknown for this to occur in the modern day. So again, giving, you know, the people a sense of being heard. And then I would say, arguably, potentially opening up the legislative agenda, because I think there's a degree to which, you know, the closure of agendas, which makes parties efficient, often prevents some of the best ideas from, you know, reaching the table of policy.

Last two questions. First, can you tell us a bit about your work with Native Americans and Indigenous peoples more generally?

Yeah, so I grew up in a very small town in Michigan. My wife also grew up in a very small town. She actually grew up on a farm right next to an Indian reservation, and we had one near us as well. I didn't think about it for a long time, but when I got digging into this petitions project, I spent a lot of time in the archives. As I did in the FDA book. And as I began to look into the petitions, I just saw more and more petitions from Native Americans. And that's part of the reason that I emphasize them as much as I do in the book, is that on a per capita basis, maybe the French Canadians aside, few people are petitioning as much as they are. And then that got me into thinking also, why don't political scientists study Native Americans? I mean, there's millions of them, especially you start to take a look at Alaska, First Nations, things like that. You've got all this variation in tribal governments. So part of it was an intellectual interest. And then without getting into a long personal history, I'm not Indigenous myself, but I've got a number of family members who are on different, both sides. In part, it was hearing just different stories around dinner tables, family reunions, things like that, that made me start to think that these are things that we should probably be studying. The work I do now is in part work on Native history, allowing these people to understand— not allowing them to understand, they understand it pretty well— working with them to better understand their histories of petitioning, advocacy, things like that, lobbying. And in some cases, like one of the courses I'm teaching right now, part of what we do is work with particular tribal communities, uh, nations that are dealing with some pretty significant issues like suicide and suicide prevention at the Fort Peck Reservation, or traditional food gathering practices and the way they're regulated by state governments and the kind of maze that different Native Americans have to navigate as they, you know, practice their traditional cultures.

Last question: How and why do you enjoy fly fishing?

Thanks. I think maybe you'll disagree, but I think most of your academic visitors would agree that sooner or later you need something where when you do it, you can't think about work. Not that you just don't, but you almost can't because of the absorption into a task. I disagree with those who say that fly fishing is some kind of Zen practice. I mean, I've spent enough time, you know, freezing on a stream fishing for steelhead in the Great Lakes or the Pacific Northwest, or fumbling with my knots and things like that. There are moments that are I don't even know what Zen is because I don't practice it, but there are moments that are definitely contemplative and lovely, but it's, it's absorbing, and it's absorbing in, um, some very few beautiful places. I've got a brook trout right behind me in a picture. Uh, it's one that my students see in all my lectures. I happen to think that brook trout are some of the most beautiful creatures on the planet, proof for the existence of a divine, and I like spending time around brook trout.

Daniel Carpenter, thank you very much.

Thanks so much, Tyler. It was an honor.

Thanks for listening to Conversations with Tyler. You can subscribe to the podcast in iTunes, Stitcher, or your favorite podcast app. And if you like this podcast, please consider rating it on iTunes and leaving a review. This helps other people find the show.