S8E5: Christopher Robertson

This is Larry Lessig. This is the fifth episode of season 8 of the podcast Another Way, produced by EqualCitizens.us. In this season, we've been considering a case that we are helping to litigate that will end super PACs. Sounds crazy, I know, but I am, even though as cynical as anyone else, I am increasingly confident that this Supreme Court will apply its principles consistently in a way that will lead it to conclude that super PACs can be limited, not by overturning Citizens United, not by going back on anything they've ever said, indeed by embracing the logic of Citizens United and applying that logic to a question that Citizens United never considered. We talked about that extensively in the third episode with Courtney Hostetler. And in this episode, we're gonna talk about another part of that original Supreme Court test, because in Buckley v. Valeo, which the last episode suggested has no foundation in the original meaning of the Constitution, but we're gonna put that aside for now. In that opinion, the Supreme Court said that what the states and Congress can regulate is to avoid what they said is corruption. And by corruption, they meant quid pro quo corruption, this for that corruption, or the appearance of corruption. Now, when you say appearance of corruption, That raises the question, what is corruption? So we're going to think about this in the narrowest way possible, the appearance of quid pro quo corruption. Now, that's a standard that many people poo-poo as being too vague. But on the episode today, Chris Robertson, who is a law professor but also has a PhD in serious science that helps him be able to analyze these questions in a serious and scientific way, Lawrence Lessig developed a strategy for trying to measure appearance of quid pro quo corruption, to find a way to say whether a people are likely to see a system as corrupt based on the amount of money you're allowed to contribute to candidates in that system. It's really an elegant set of experiments that he pulled together, and he submitted them in a submission in the district court to lay out an understanding of the appearance of corruption that will give the court a way to make sense of that vague term and to see why rules like the one Maine passed, 600,000 Mainers passed in the 2024 initiative, actually make sense of the standard of the appearance of corruption. So stay tuned for that. Chris, thanks so much for coming to talk about, uh, what you contributed to our case. Um, but I want to just put this in a little bit of context because you and I have known each other for a long time, and you've been doing work in the context of understanding the effect of influence for a long time. So why don't you just introduce the genre of work that you've become, I think, of the world expert in thinking about this particular kind of problem?

Well, I was trained as a philosopher, and a lot of the foundational questions about epistemology and how do we know what we know and who we can trust and what we can rely on. And a fundamental premise is that there's very little that we know directly. Most of what we know, we know through others, through their representations, and especially through their expertise. And so you think about these chains of epistemic reliance is what I call call them where, for example, in biomedical science, a patient relies on the doctor, but the doctor relies on the scientific literature and also on the Food and Drug Administration. And if any of those steps can be distorted by conflicting interests, and for example, the physician can have a patent on a medicine or can have a side payment from the drug industry.. Or the FDA can be funded by user fees that are structured in a way that create incentives to approve drugs rather than reject them, and to do so quickly. So understanding how each step in that chain of reliance shapes the advice that you get is really important then to really our theory of knowledge, because all of our knowledge, almost all of our knowledge, is indirect in that way. And so I've also been interested then as I came to law school and got trained to think about institutions and incentives and durability of what are the solutions to those sorts of problems. And noticed that sometimes a bias is literally a psychological one where someone intends to do one thing, but their subconscious causes them to do something differently. Other times we can have a corrupting influence that's perfectly rational, right? That someone's just doing what the system is designed to make them do, even if that design isn't well aligned with public interest or the patient's interest. So a lot of the work that we engaged in over the past, when you were leading the Institutional Corruption Lab at the Safra Center, the Edmund J. Safra Center, here at Harvard, involved solutions to bias or institutional corruption. And one of them that I was very attracted to and focused on at the time was blinding. Can we design institutions or decision processes where people will have relationships or incentives, but if we can carefully design the decision space so they're not aware of them, and when will that work, when will it not work, right? It's a technology for for knowledge production, a technology for reliance, that'll work in some contexts and not others. So that was a big path of my work in that space. Beyond that, more recently, I've worked on all sorts of things about, a lot of it court-related. How do you decide whether to trust the government's witness in a criminal prosecution, where the government's witness has gotten a plea bargain that's essentially conditional on him singing the tune of the government. More recently, I'm working on a project around PDUFA, the Prescription Drug User Fee Act, that funds FDA in this way I alluded to. So the problems, I mean, as you've seen, are fractal. We see them recurring in sort of everywhere we look.

Right, and so one of the big questions is whether, as the user or the voter or the consumer, you have a reason to trust the representations that are being made I remember at the Safra Center, there was some really great research done with doctors where they would give doctors descriptions of research. And in the descriptions of the research, they would specify whether it was, you know, kind of the gold standard random blind trials. And then they would also mention how they were funded. So it was either NIH funded or it was funded by a drug company or it was funded, didn't mention how it was funded. And what the study demonstrated overwhelmingly was that doctors discounted representations from the research if they knew that it was funded by the drug company. So the mere fact of the funding by the drug company was enough to lead the doctors to be suspicious, not to have confidence in that research. And from my perspective, you would say to the people who are funding the research, geez, you're just throwing money down the drain because you're doing all this stuff to like stand up a result that people are not believing. And that's a context where a blinding structure could be really helpful for the pharmaceutical companies, right?

Yeah, for your listeners, I'll mention that's a 2012 study in the New England Journal of Medicine with Kesselheim and Rose that we did. And let me just add though that that mechanism of disclosure, classic solution to institutional corruption, —of disclosing industry funding really depend on the users here, the doctors, which we were studying real physicians. They have to be in a position to get that information. But the status quo at that time, and to some degree still today, is journals weren't putting that information on their abstracts. Yeah. Right. So it was really a proof of concept study to show if you do put it on the abstracts, your users, your readers, will discount the abstracts accordingly. The second thing I'll say is that the scientific literature is filled with industry-funded studies, right? The research shows they are biased in the way the statistics are performed, in the way the results are reported, et cetera. When the physicians were discounting their credibility of those studies, it was actually a rational, reliable sort of discount. They weren't just being prejudiced against industry-funded studies. Studies. So we followed up on that research with a hackathon, also hosted by the Institutional Corruption Lab, where we showed how a browser extension could be overlaid over websites like PubMed, and it would go fetch the disclosures from the backend of an article and append them to the appearance of the abstract. And we showed that that was perfectly feasible. Fortunately, a few years later, the PubMed, this National Library of Medicine, implemented that on the front end, so you didn't need a browser extension, at least for certain journals, those abstract disclosures now appear, so.

So it was all of that work that made it seem so obvious to me that I needed to try to bring you in to the case that we are litigating about super PACs, because that case involves a decision by the Supreme Court, 50 years old, Buckley v. Valeo. And in the Supreme Court's divining of what the First Amendment standard will be to judge campaign finance laws, and I mean divining in a pejorative sense because one part of our case is arguing this is all made up relative to the original meaning of the First Amendment. But anyway, 50 years ago, the Supreme Court was totally cool with just making up standards of law. So here they are announcing this new standard, and the standard is based in corruption. And I gathered a couple quotes from that opinion to just make clear their really rich conception of corruption. I think people think of it in a very simplified way, but it was actually much richer. So what the court says in Buckley is, it said, the primary interest served —by the limitations in the statute, and indeed by the act as a whole—is the prevention of corruption and the appearance of corruption spawned by the real or imagined coercive influence of large financial contributions on candidates' positions. I go on to say, to the extent that large contributions are given to secure a political quid pro quo from current and potential officeholders, the integrity of our system of representative democracy is undermined. Although the scope of such pernicious practices can never be reliably ascertained, the deeply disturbing examples surfacing after the 1972 election— that's Richard Nixon— demonstrate that the problem is not an illusory one. And then goes on, of all, almost equal concern as the danger of actual quid pro quo arrangements is the impact of the appearance of corruption stemming from public awareness of the opportunities for abuse inherent in a regime of large individual financial contributions. Here, as there, Congress could legitimately conclude that the avoidance of the appearance of improper influence, quote, "is also critical if confidence in the system of representative government is not to be eroded to a disastrous extent." And then finally, Congress was surely entitled to conclude That disclosure was only a partial measure— disclosure of contributions— and the contribution ceilings were a necessary legislative concomitant to deal with the reality or appearance of corruption inherent in a system permitting unlimited financial contributions, even when the identities of the contributors and the amount of their contributions are fully disclosed. Now, what I love about that concept is that it's both addressing the thing that we all agree is a problem, appearance, I mean the actual quid pro quo corruption, but it's also recognizing that the public rationally would believe that there's quid pro quo corruption going on in a world where there are no limits on the size of the contribution, 'cause you'd be sitting there thinking, yeah, yeah, so he spent $270 million to try to get that guy elected, and he didn't get anything in return for that, or he didn't expect he'd get anything in return for that. And even if the person who spent $277 million didn't strike a quid pro quo, the point the court is making is that there's a corrosive impact on the public as a whole when the public rationally concludes that, of course, he must be getting something. So even if the contributor is completely, absolutely, rigidly not engaging in a quid pro quo, it still has this effect on the public, and that secondary effect is what they were concerned about. And so when I studied this really carefully, I thought the one person who could help us try to map that secondary effect, like, what can we even, can we measure it? Can we see how it grows or falls depending on the size of these contributions? I thought you were the perfect person. Person to bring in. So I was so grateful you agreed to help us, and you submitted a declaration in the district court. Uh, maybe that was part of the reason why she was willing to conclude that there was, of course, a risk of quid pro quo corruption with these large contributions. Um, but she didn't really address exactly what you had been talking about. So I was really eager we have a chance to unpack it here in this conversation. So help us understand how you set up a study like the one that you did, and like what, what we should be able to conclude from the study you set up.

Well, sure. Well, I'm honored to be called on, but I will say that there is a broad literature of people in the political science field in particular studying perceptions of trust in government and all sorts of related constructs. And so, but one of the things I think we wanted to be very careful about is making sure that we are answering the question that the court is asking, right? Because saying that, you know, for example, trust in government varies or doesn't vary, may not be what the court accepts as perceived quid pro quo corruption. And so, and even the word corruption, I mean, you've advanced and other scholars advance other conceptions of construction, corruption that are maybe more capacious, but for this purpose, we sort of limited ourselves to the, even the more recent cases, which have cited smaller and smaller portions of that quote. To say, you know, the perceived risk of quid pro quo corruption is what we want to measure. And so the first thing we did, of course, is, you know, look at the literature, and it turns out no one has really studied that as directly as we should in order to speak to the court process. And so the next thing we needed to do, though, is to be able to construct a test that has a real counterfactual. We wanted to understand how different levels of contributions affect perceptions of the risk of quid pro quo corruption, and in particular, whether a cap affects that risk. And so each of those, to get to evidence about what difference it makes, you have to measure the with and without the counterfactual reasoning. And so the social science methodology that lets you do that is a vignette experiment, where you first of all recruit a broad sample of respondents, and we weighted in our recruitment by age and gender and race, and to get a decent cross-section of America from 49 different states. And then for this litigation, since we're based in Maine, we oversampled Mainers to get 115 or so of them as well. We asked them in this survey experiment to imagine that they're in a situation. The survey had two experiments built into it. The first one systematically manipulated the amount of a campaign contribution, and we went from $5 all the way up to $50 million in logarithmic scale, 5, 50, 50, 50. $500, $5,000, $50,000, et cetera. And we just asked them in a short paragraph, imagine a political official running for office receives a contribution from a committee supporting his campaign in the amount of X, how likely do you think he would sell official action in exchange for that? And so the outcome was measured on a Likert scale, 7 points, very likely to not very likely. And we're able to see how the rate of perceived selling risk varies with the amount.

So, but when you're doing this, I mean, if you just think that you're just giving people a whole bunch of numbers, it doesn't seem very convincing. But the actual way you set it up is much more elegant, right?

It's a between-subjects experiment, so each respondent only gets one of those values, right? Partly that's so they don't go bored with numb with too long of an experience, but also to maintain the validity. We don't really want them to know what number we care about. We want to keep them blinded to our research hypotheses. For them, it's just a one-shot. Some respondent got a $500,000 gift that they're assessing, and so they measure it on this 1 to 7 scale. One of the things we found is that consistent with the prior literature, people think quid pro quo risk is very prominent. Even at the lowest levels, under $5,000, we found 30 to 40% of the people saying, "Yeah, he could sell his vote for that." What we did find is that above that base, as the amounts of the country— $5,000. Above $5,000, we had several levels below it, but above that, the risk starts getting higher and higher up to the 70 and 80%. You get supermajorities of respondents saying, "Wow, you can give him $50 million.", that looks like it could create a risk of quid pro quo, and we see that at $5,500,000 all the way down to— and so we see that variation, that differentiation between small gifts and large gifts. Large gifts to a committee create a much bigger perceived risk. So that's a pretty straightforward first finding from our science is that there is a correlation with the amount, and there's a discontinuity above $5,000.

So that's the first experiment. Right, and so that experiment gives you a reason to believe that the concern that they were talking about in Buckley is really created by systems that remove all caps and allow $270 million contributions to committees that are supporting a particular candidate.

Yeah, I mean, I think, I don't think this is a, you know, mind-blowing finding. It's common sense, but sometimes our job is to remind the court of the sense that we already have, right? So in the second experiment, we wanted to drill down more into the actual legalisms of who can give what where. That first experiment was sort of naive in that way. It didn't talk about the law of bribery or this kind of committee or that kind of committee or federal limits. It was kind of naive, and we think that's probably the right approach for a lot of low-information voters who are more concerned with the price of gas and what they're having for dinner tonight. But nonetheless, to speak to a judge and a bunch of lawyers, we wanted to drill down more carefully, and so we presented a longer vignette that laid out an entire campaign finance regime in about 10 points about there's direct contributions, and then there's committees, and there's et cetera. But ultimately, we randomized in that second experiment whether this regime had a cap on contributions to— at $5,000 to super PACs or independent expenditure committees. And so just to emphasize, only in the second experiment did we introduce the $5,000 level. We didn't want to drive responses in the first experiment, so this is distinct. Now, our question is, in this regime, with all these different policy levers, how likely do you think quid pro quo corruption is? We actually asked that question in 3 different ways, and here again, we found in the conditions where we randomized them to get the policy regime with that feature, the risk of quid pro quo corruption was dramatically lower. I mean, here it was like— 10% with the cap, but it went up by 3x to close to 30% without the cap. So statistically significant, substantial, a 3x is a huge ratio, and so that was the second experiment.

But just to make sure, because this is really important to connect it to the case, Maine has passed a law. The law has capped contributions to these independent expenditure committees. If Mainers behave the way this experiment suggests, that suggests that Mainers will have less suspicion that there's quid pro quo corruption going on in the campaign finance system because of that cap. And without that cap, they would have 3 times the suspicion that quid pro quo corruption was going on than they would with the cap. So Maine's objective, or the people of Maine's objective, to try to produce a regime that gives us more confidence in our political system is borne out by the results of your study. Yeah, and again, it's not mind-blowing.

The Mainers enacted this law presumably because they had a view in this same ballpark, right, or they wouldn't have voted for it in the first place. And so I think we just, we find that same sensibility. And I'll say that, I'll just remind that our study included a sample of Mainers, and we got that same statistically significant effect in the subsample of Mainers. So it's not peculiar or something to them. We also collected qualitative data. We asked people, you know, what do they think would happen in the system? Why do they think it's important? And when you look at those responses, people in their own words, you see exactly these sorts of themes. They talk about corruption, they talk about buying and selling votes, they talk about temptation even among good people, as well as other conceptions of of corruption. Like, they also mentioned that in this system, the wealthy people would have a lot more access, and they would have a lot more power, and they could flood the airwaves. There are other effects as well, of course, and we asked a bunch of other questions and saw just generally good government perceptions are promoted by this sort of reform.

Yeah, now the court has put off the table these other conceptions, like court has expressly said the mere fact that you get more access doesn't mean it's, quote, corrupt. But what's important is that you're able to distinguish this and show that we're not talking about access. It's not a perception of access. It is the view that there's likelihood of quid pro quo corruption going on given the presence or absence of a limit.

Yeah, I mean, if this court wants to keep our doctrine that narrow, I can answer that question. But I do think it was worth seeing that— Yeah. There's— this also answers broader questions.

Yeah, I mean, I think that one big question that, you know, I feel increasingly animated to press is, you know, by what right does the court narrow the scope of the kind of corruption that we're allowed to try to police in the context of our political system, right? You know, it'd be one thing if we the people passed an amendment that said the only regime that we permit under this amendment is to fight quid pro quo in the narrowest possible sense, corruption. But we never said that. Yeah, we the people never did anything like that. This is all made up. It's just this doctrine that comes out of the head of these justices who feel like they need to build this elaborate regime to basically shut down any of these reforms. But the point is, even with their regime, even with their extreme regime, what you've given them is a tangible way to measure what they were describing in this kind of abstract way. Because the problem with the way they described it before is that most people who were trying to push away the appearance of corruption thing sort of talked about it as if it was just a kind of fantasy, like, or phantom out there like this. You know, I mean, it's either corruption or it's not corruption. But the point is, when you see that what they were rightly worried about is the public believing that the whole system was rotten. And so why should I participate in a rotten system? Their perception is the rotten system is a function of these rules. And so if you can craft rules that may lead people to believe it's not a rotten system, then they're gonna be more willing to be part of the system. They'll be more willing to believe that even though they don't have, you know, a trillion dollars, there's a reason for them to give their $100 or to show up and vote. I mean, so it's really critical and it's right at the core of what Buckley was worried about. So have you had pushback or have you had people—

Well, this was litigation, so of course I sat for a deposition and there was exposing an opposing expert on the other side. And so we, you know, exchanged reports back and forth. And by the way, when I talked about my prior work around institutional corruption and bias, I mean, I focused a lot on expert witnesses in the past, so I was very sensitive to make sure that I wasn't just becoming a hired gun, but could let the science speak for itself. And so the expert on the other side is a well-respected political scientist who's published a book on perceptions of government trust and polling. But you can see in our battle of reports, the biggest problem with that research he was relying on where most people when polled remain skeptical, remain cynical is really the word, about government. And so if you just ask them, "Do you think campaign finance reform is gonna help?" And they're like, "I don't know." On the other hand, again, why would they pass the law if they don't think it's gonna help? I think the referenda initiatives actually speak for themselves in that regard. Primo and other regression scholars, scholars who take big surveys and try to compare them to state law changes, they also say, "Well, in these years, some states passed campaign finance reforms defined broadly, and then in polls conducted in subsequent years, we don't see a big increase in trust for government. And there's a million problems with that sort of research. It's really hard to get— A, it's not asking our precise question about quid pro quo, but it's not even measuring this particular reform the way we did here. And there's real problems with statistical power because you're trying to measure a diffuse concept of trust in government, which can be affected by all sorts of things, right? I mean, if you think— Senators are too old or come from a different class or not, you know, whatever your objections to why you might not trust government, campaign finance is of course not the only one. And so it's really hard from a social science perspective to measure a specific construct with a general measure. And so he basically gets a null result, but the confidence intervals are too wide to really reject the sorts of problems that we observe here.

So this is new research. It's really important to understand the legal context. What we're really hopeful for in the First Circuit is that they look at the research and they, you know, along the way to like reaching the right conclusion about whether Citizens United means that super PACs are constitutionally compelled, which obviously we think the answer to that question is no, they can help fill out the jurisprudence by giving the Supreme Court a clear way to understand what appearance of corruption, appearance of quid pro quo corruption should mean. And the measure you've given is a beautiful way to measure it. Now, I will confess, I was a little— it felt a little bit too good. Because it turned out $5,000 was a magic number, right? I mean, we— the initiative had $5,000 just because that's the same number at the federal level. There was no, like, study to figure out what the number should be. But the idea that it turns out $5,000 is a magical number for people and that you begin to see the curve shifting dramatically at that line is quite significant. I mean, it makes it really important that we understand why this number makes sense.

Yeah, it may partly be that enough Americans implicitly are aware of those caps already, and so they start, you know, they've internalized them. It may also be that the team that passed this, that wrote this initiative in the first place, you know, were aware of both of those factors. They're like, I know, they know the law, but they also know the common sense of Mainers. And so we picked up on, I think, that background phenomena. Even though 5,000's obviously an arbitrary number and we would probably prefer it to be lower, I don't know, maybe higher, but—

We prefer what the people of Maine want. They should just be allowed to have what they asked for.

And that's a classic, you know, 1L law school problem with line drawing, you know?

There's always the Goldilocks issue. Right. Well, Chris, I'm really grateful that you did this work. I'm really, really happy the work came out the way it did and that we had a chance to present it. And we're gonna have a really fantastic Court of Appeals have a chance to consider it. We don't know the panel yet, but at the end of the month, July 29th, we will have arguments and it's up from there. But—

Well, happy to be part of it. Thank you for having me.

Great. That was the 5th episode of the podcast Another Way. As I said in the first podcast, there's a chance we will have one more episode depending on whether I can twist a lawyer's arm. That episode will unpack exactly why the recent Supreme Court case about campaign finance, whether party committees can coordinate or not, um, is perfectly consistent with the theory we've advanced about why regulations on contributions are different from regulations on spending and why the argument we're making in our case about the permissible regulations on contributions are constitutional under the analysis the court has given us in Buckley and in Citizens United and every other Supreme Court case. So, Stay tuned. That'll pop up on your feed if indeed I get the— or I'm successful in twisting this lawyer's arm. Meanwhile, thanks so much for listening. Grateful to Josh Elstroh of Elstroh Productions for producing these episodes. You can follow the case at equalcitizens.us and learn about progress after the Court of Appeals hears arguments. We'll be sending out updates and feedback and reflections after the argument on the 29th of July. Thanks for listening. This is Larry Lessig.