Steve Teles ranks among America’s leading academic experts on the application of conservative ideas to problems of governance. He chronicled the rise of the conservative legal movement in a 2010 book. This spring, he and co-author David Dagan have released a new study: Prison Break: Why Conservatives Turned Against Mass Incarceration.

Not a conservative himself, Teles writes as a sympathetic outsider, always looking for ways to bridge ideological gaps in the service of better policy. With the Cato Institute’s W. Brink Lindsey, Teles has long co-chaired and co-hosted a monthly seminar that steps outside the usual perimeters to bring together not only conservatives and liberals—but also libertarians and socialists—for discussions that are simultaneously unusually frank and unusually practical. I’ve been an occasional participant myself, so I have to disclose that I’ve received four or five free meals from Teles’s funders. Until they start serving better wine, however, I consider my intellectual independence uncorrupted.

I interviewed Teles by email in late June.

David Frum: Let’s start with the thesis embedded in your subtitle. Is it indeed true that conservatives have turned against mass incarceration? I can think of a few who haven’t.

Steve Teles: It is certainly true that there are a few conservatives who are still holding out against the reform spirit in the GOP, most notably Senator Tom Cotton of Arkansas. And in general, Republican legislators who have been in office a long time and have a record of supporting measures like mandatory minimums are going to be less eager to reverse field than those with a fresh record. There are also states where the GOP leadership is still pretty hostile to reform, like Virginia—in purple states the temptation is very strong to cling to the old criminal-justice orthodoxy. But the list of people who have, at least minimally, signed on, is much more impressive. Readers can consult the list of the signatories to the “Right on Crime” statement of principles.  You’ll see a lot of names of prominent Republicans, including many—like Jeb Bush, Newt Gingrich, Asa Hutchison, and Ed Meese—who were important figures in building our system of mass incarceration in the first place. And the most important evidence is the list of bright red states like Texas, South Carolina, Mississippi, and Georgia that have passed major reforms over the last few years. I think it’s fair to say that reform is much closer to being GOP orthodoxy now—especially in the states—than the old tough on crime stance is.

Frum: One of your defining intellectual interests is the interplay between policy elites and real-world political outcomes. The connection between elites and policy is always tightest when policy can be advanced through the courts. But criminal justice is more swayed by public opinion than, say, tort law: That’s why the U.S. still has the death penalty, for example. Would you explain how the “right on crime” movement gained traction—and what limits may be imposed by the conservative rank-and-file? It’s noteworthy in this context that Right on Crime was a cherished ideal of the Koch funding network, and if 2016 has demonstrated anything, it’s the limit on that network’s real-world political power.

Teles: Well, first of all, the relationship between public pressure and action in criminal justice is very complicated. I’d put it this way—starting in the 1970s and going well into the early 1990s, any regular working politician would have gotten very strong signals from their voters that something needed to be done. Crime became what political scientists call a valence issue—you pretty well had to be on the side of aggressive action. Now, how that actually cashed out in terms of actual policy was not something that you could read directly out of public opinion. Politicians shaped that response, for good and for ill—especially the competitive pressure between politicians. Take something as simple as the death penalty. Sure it’s the case that voters usually had pretty specific preferences on that. But on whether there should be mandatory-minimum penalties for particular offenses? That’s a different kettle of fish. The important point, however, is that public opinion mattered a lot on the way up by creating pressure on politicians to act, and to be seen to be acting, and by creating the structural conditions for competitive dynamics between parties that pushed up severity.

The openness of conservatives to rethinking criminal justice is, to a significant degree, a function of the declining salience of the issue. Voters since the late 1990s simply haven’t cared about it as much, as the great crime decline started to register. Voters will still tell you in polls that they think that our criminal laws aren’t severe enough, but they also don’t care about it as much. And that lack of strong concern creates space for politicians to move without fear of reprisal, and to be more entrepreneurial in their framing of the issue. I’d also note that the death penalty, in recent years, has been in retreat in many states, which is both a response to the same declining fear of crime and the reframing of the issue as being about “actual innocence.”

The involvement of the Koch network is certainly interesting, but what’s most significant about this movement is that all of the actions we document in the book preceded Koch getting really aggressively involved. Right on Crime, Prison Fellowship, and the other actors in this story got funding from other sources—including some funders who were quite lefty, as well as those in the center like the Pew Charitable Trusts. Up until recently, conservative funders have not been that interested in the issue, which I know has been a matter of frustration for organizations on the right that wanted to get involved but couldn’t find the cash. If Koch makes a difference going forward, it’ll be because it gives those organizations (especially state think tanks) the resources to do what they wanted to do already. It won’t be because the Koch name makes such a big difference.

Frum: There’s evidence that crime rates—especially homicide—began rising again in 2014 in major metropolitan areas. The Black Lives Matter movement seems to be having a racially polarizing effect on attitudes to police: more expressions of mistrust by African-American communities, some rallying to the police among non-African Americans. Is it possible that the trend you describe in Prison Break was the product of very specific circumstances that have already come to an end?

Teles: We speculate on that in the conclusion of the book, which was written a few months ago, but by which time we could already see some of the factors you’re discussing. We honestly think there’s reason to worry. Declining fear of crime absolutely played a role in the shift in conservative positions on the sentencing, re-entry and the like. If voters start clearly expressing that they’ve gotten worried again, and that the issue is a priority for them, the room for politicians on the right to experiment will start to shrink, and the temptation to use the issue for political gain will increase. I have no question about that. That is a reason why it is so vital that we do whatever we can to keep crime rates down.

Part of that is putting resources into establishing more productive relationships between communities where crime is highest and those that police them—there is just no way to sustainably keep crime down without such relationships, which give police the consent to do what they need to do. There are also a lot of things that we should have been investing in years ago that are proven to keep crime down, especially intensive parole and probation (the sort of things that Mark Kleiman has been writing about). So my first answer is to say that there are reasons for concern, but that a sustained spike in crime rates is not inevitable, and my friends on the center-left need to prioritize crime control. But there is space for agency here—we are not at the whim of structural forces over which we have no control.

Secondly, I think that the momentum for reform where it has already started is very strong. In states like Georgia and Texas, they are already on their third or fourth round of reform. In Georgia there is an official commission that keeps reloading new reforms every legislative session, and so in places like that, reform is sort of hard-wired. The politicians there have learned that making changes doesn’t lead to crime spikes that they will be blamed for, and the Republican establishment has pretty well put all its cards on reform. It’s very hard for them to turn back. I’m more worried at the federal level, and in states that haven’t already started down this road. In those places, there are in fact strong incentives to at least keep your powder dry. But even in those places, there’s just so much more infrastructure on the right pushing reform than their was just a few years ago, and so prosecutors and other pro-severity interests don’t have an open field to themselves. So I’m more pessimistic than I used to be, but still think there’s a lot of momentum.

If I’m worried about anything, it’s Trump. He’s like a throwback to New York in the 1980s, as my friend Rick Perlstein has argued very persuasively. The Right on Crime movement depends upon, in some important ways, the transformation of the Republican Party into a more consistently anti-statist party in the wake of the Tea Party, combined with the role that evangelical leaders have played in encouraging an emphasis on second chances and forgiveness. Neither of those changes in conservatism is characteristic of the conservatism of Trump. I could imagine him going all-in on a back-to-the-80s, Charles Bronson-ish approach to crime, and if he’s able to rebrand the Republican Party in that way, that would be very troublesome.

Frum: Well, yeah. Isn’t the lesson of Trump’s rise that the conceptualization of the Republican Party as an “anti-statist” party was at best inexact? Given your interest in elite politics, and the influence of ideas on politics, maybe there’s another way to look at what happened in 2009-2015. A populist protest movement took form, reacting as much or more against President Obama’s perceived threat to existing distributions of state resources as to deficits and debts. That movement often used libertarian-sounding language (“protect the Constitution!”) because it was the only language the protesters knew. Lacking charismatic leadership of their own at that time, the populist anti-Obama protesters were vulnerable to elite capture by well-resourced groups with a highly developed anti-statist ideology. Especially at the state level, those elite groups could then have real impact. But they never represented their voters very well. And when charismatic leadership did materialize in the form of Donald Trump, those voters revealed how little deference libertarian policy elites really commanded.

Put it another way: Donald Trump talks a lot about crime. He tells audiences that crime is rising. He backs the police to the hilt. “I have to say that the police are absolutely mistreated and misunderstood.”

His solution isn’t crime control plus deincarceration. It’s firearms, so that “law-abiding” citizens can send “bullets flying in the other direction” when they come under criminal or terrorist attack.  And when put to the voting test, this approach drew rather more support from rank-and-file Republicans than Right on Crime’s.

Teles: I don’t really disagree with this take at all. Back in 2009, I published—in the first issue of National Affairs—an article called “The Eternal Return of Compassionate Conservatism.” I argued there that the effort to engage with minority voters and to try to find ways to engage with their concerns has a long history in the Republican party. To some degree it goes back to Nixon and “black capitalism,” re-emerges in the 80s—led by Jack Kemp—as “empowerment,” and comes back again under George W. Bush as compassionate conservatism. Each time, it’s an almost exclusively elite, mainly D.C.-based conservative initiative by those on the right who are genuinely embarrassed by the fact that the Republican party has so little to say to black voters in particular, and by think-tank types who think that they have a more effective agenda for them than the Democrats. You can see some similar dynamics around criminal-justice reform.

Each time, it became clear that the base of the Republican Party, and working politicians, didn’t really care about this—they weren’t particularly embarrassed and didn’t think of this as a priority. They were comfortable with where the Republican Party was. So I think it’s fair to say that the combination of anti-statism and outreach to black Americans is an elite phenomenon without much appeal to the Republican base, and a politician like Trump can absolutely drive a wedge there. That’s another reason why I emphasize how important it is that crime rates stay down—so long as there is not intense public engagement on the issue, it’s easier for politicians to look to elite conservatives for cues as to what position they should take on the issue. But if crime rates keep rising, working politicians connected to average voters will find the BLM criticisms of police in particular very hard to resist as a target for reaction. And to some degree those protesters—who I have some sympathy with—will provide them a very tempting target.

The difference, if there is one, between the movement to reform criminal justice and previous versions of compassionate conservatism is that this is a much more decentralized process. You don’t need full-spectrum congressional commitment for things to happen. And because criminal-justice reform holds out at least the promise of being budget-neutral—if not actually saving money, given how expensive incarceration is—it doesn’t have the same trade-offs as other efforts to reach out to black voters. But I completely agree that Trump has an approach to criminal justice that could very easily drive a wedge between Republican voters and the conservative policy infrastructure that has been pushing reform.

Frum: Your model of politics puts great emphasis on persuasion by and within elite factions. “Conservatives shifted position on criminal justice,” you write, “because, at least initially, a small number of [people fired by intense moral conviction] decided that the [Republican] party’s position was problematic, and were willing to use their personal credibility to persuade influential elites to rethink their position.” Have the events of 2016—from Trump to Brexit—caused you any second thoughts about this perspective?

Teles: This gets to a larger issue with our model. I think it’s right to say that ours is a “pre-Trump” model of the Republican Party. In that model, you’ve got two sorts of issues, those that are “anchored” by a major party-coalition member (social conservatives and abortion, global warming and the Koch network, etc.) and those that are “opportunistic” in the sense that they are rooted in conservative identity but not in a mobilized group. In our model, there’s really not much that you can do to shift party positions short of a cataclysmic change in the party coalition, of the kind that happened to the Democrats by the 1960s. On opportunistic issues—and that’s what crime was, there was no really deeply organized party coalition-member anchoring the position—change can happen when the “cue-makers” in the party signal to the rank and file that a new position is consistent with party orthodoxy. This gets a little bit to how we think partisanship works. Ordinary working politicians and even voters have an intense tribal commitment to the “brand” of conservatism, but that brand is flexible as to its content. The brand is filled in by the cue-makers in charge of the brand’s content. The actors who had the ability to serve as cue-makers have been pretty stable in the Republican Party over the last few decades. One way to think about what’s going on with Trump is that their power in the party has been shaken. The falcon can no longer hear the falconer, you might say. I don’t think this process is complete, but the ability of conservative elites to signal to their followers where orthodoxy is has been disturbed. That’s bad news for those trying to shift conservative orthodoxy—we may have moved from a period where, for good or ill, there was ideologically based authority, to one where even that authority is coming unglued. There were a lot of people who were disturbed by the challenge to expert authority by ideological authority, which was a big part of how the conservative movement changed American politics. We’ll see how well they like it when even that authority has lost its power.

Frum: Final question: The likelihood is that Donald Trump will lose, possibly very badly. The Republican Party and the conservative world will be plunged into intellectual and moral ferment, even crisis. How scalable are the lessons you offer in Prison Break? To what extent do they offer a guide to Republicans and conservatives who seek change? And what are the real-world limits you expect for them, given that the one thing that will remain true post-Trump is the revelation of the weak hold of the ideology of anti-statism on the Republican rank and file?

Teles: This is a very complicated question. And as a political scientist, the thing that strikes me is how under-determined the response to a very big Trump loss will be. If it’s very large, then it sends a signal beyond the personality of Trump, which is that you just can’t put together a winning presidential coalition on the basis of LePen-Americane. But if it is close, then it could be read the opposite way, which is that if you could just put a less obnoxious face on his coalition then it’s possible for the Republicans to succeed as a purely white-nationalist party. I have to think that if it really is a blow-out, then the authority of the Republicans who embraced Trump has to be weakened, that the NeverTrump people will have a strong claim to have warned their co-partisans that they were joining a sinking ship. But I think that one way to predict what is going on is that you’re going to have real factional conflict in the Republican Party in a way that you haven’t had in a couple of generations. The identity of the party is up for grabs in a way that it has not been in some time.

If it’s true that there’s going to be real factional conflict in the party, then one thing that suggests is that the power of party leaders is going to weaken, especially in Congress. As my friend Lee Drutman has argued, based on conditional party government theory in political science, if the Republicans are split factionally, the only way to govern Congress will be for party leaders to give up power over the agenda, and allow members to make coalitions they want—even with Democrats. That creates the opportunity for more strange-bedfellows coalitions, institutionally—party leaders won’t be in as strong a position to control the agenda. But simultaneously, the authority for party members to shift positions is going to be weakened, because they can’t just point to a consensus group of cue-makers as justification for changing position. So we may be going into a more anarchic sort of future, with shifting coalitions but also a lot of blame-avoidance dynamics among Republicans who are worried that whatever they do, they’re going to get punished and can’t hide behind the party power-brokers any more.

I do worry that if the Republicans go in a more white-nationalist direction—if that’s the interpretation they put on this election, that they can’t ignore the Trump voters and that those voters want statism, but statism used to serve them—then it’s harder to build coalitions of the kind discussed in the book that combine anti-statism of both parties. That makes cutting Pentagon spending, rethinking surveillance, and further criminal-justice reform harder. But it creates opportunities for other kinds of strange bedfellows coalitions, in particular on trade.

My bottom line is that it might be that the dynamics that created the process of criminal-justice reform on the right are Humpty Dumpty, and he can’t be put back together again.