Seven Principles for Dealing with Disorder

Now that we're all on the same page...

Over at Slow Boring, friend of the ’stack Matt Yglesias has gotten on the anti-disorder bandwagon:

Somewhere on the road from Barack Obama and John Kerry getting endorsed by national police unions in 2004 and 2008 to the present day, the Democratic Party has become ambivalent about the idea of punishing people who break the rules, to the point that the party says we need to accept disorderly and dysfunctional public spaces.

The starting point of that process is totally sensible: Democrats are the party of people with humane instincts, and there was a lot of interest in trying to find ways to make the American criminal justice system less cruel. That’s a reasonable and important problem to work on, and it’s something I’ve written about over the years (most recently in May). But it turns out, like many policy problems, somewhat difficult. And in too many cases, the fallback option has become just not enforcing the rules.

This is not politically viable. But it’s also worth saying that while humane impulses are good, letting public spaces go off the rails is a kind of false humanitarianism. Most low-income people are not criminals, and it’s precisely the poorest and most vulnerable people who most need things like public spaces and public transit and affordable housing and libraries, and they need these things to be actually good.

Matt joins a growing chorus of center-left commentators who recognize that poor governance in blue cities explains why those jurisdictions turned towards Trump in the 2024 election. And they recognize that that starts with Democrats’ abdication of responsibility for enforcing norms and rules around disorderly behavior, which has bred a surge in said behavior in recent years.

I’m obviously to Matt’s right. But I’m always happy to work with people who want to agree with me on the public policy issues I care about. I am … skeptical that the Democrats are going to “be the party of public order.” At a surface level this is about the much-maligned Groups and their dominance over Democratic politics. But those groups, and especially the criminal justice “reform” groups, have purchase in the Democratic coalition for a reason—a basic alignment on values, even if there’s a disagreement about how much they should talk about or emphasize those values. And I do not think that that is so easily changeable, Bill Clinton’s temporary success notwithstanding.

But if there are Democrats who want to try to get tougher on crime, I’m all for it! So with this bipartisan groundswell of interest, I thought I might lay out some principles for how to think about managing disorder. I’ve written about much of this before, but consider this post a handy reference guide.

1. Don’t Apologize for Enforcing the Rules

As Matt notes, the urban turn away from disorder enforcement is a response to a specific line of argument. That goes, roughly:

• disorderly behavior in any individual case is not, fundamentally, a big deal;

• enforcement against disorderly behavior involves at least the threat of violence and incarceration;

• the cost of the latter outweighs the cost of the former, particularly because:

• these costs are disproportionately born by the most disadvantaged;

• therefore, don’t enforce against disorderly behavior.

I’ve talked about this before, but there are two arguments for why this is wrong. One is the classic broken windows argument: in the unit case, disorder doesn’t really matter. But repeated instances of disorderly behavior have spiraling effects, yielding diseconomies of scale. As Wilson and Kelling put it, “arresting a single drunk or a single vagrant who has harmed no identifiable person seems unjust, and in a sense it is. But failing to do anything about a score of drunks or a hundred vagrants may destroy an entire community.”

But the other is the argument that tolerating antisocial behavior is intrinsically unjust. The cost-benefit calculation above only makes sense if the interests of the antisocial count equally to everyone else’s. But they don’t. We live in a society, and in shared communities, with rules about how to use public spaces. Everyone has to follow them. It’s not unfair to the rulebreaker to enforce a rule; it’s unfair to those who follow the rules when you don’t. And that should be our normative starting point, not the one that prioritizes the well-being of the guy who tags the subway or pees on the platform.

2. What’s Important is the Dose

Does this mean that we should be maxmimally harsh on disorderly behavior? Should every guy who pees on the platform do life? No, and that’s not really the broken windows prescription. Bill Bratton—the NYPD Commissioner who pioneered broken windows policing in practice—put it quite well:

Unfortunately, people have equated "zero tolerance" with the so-called broken-windows issue. Those are two totally separate concepts. Broken-windows enforcement is really about controlling behavior to such an extent you change it: If you deal with the little things, you can keep them from going into the big things. Zero tolerance implies zealotry. It's oppressive. And it's not achievable. You're never going to be in a position to eliminate all crime.

Frankly, I think you see this kind of confusion between zero tolerance and broken windows even from liberals who were actual advocates of broken windows. But their criticisms almost always miss the point, which is that zero tolerance is confusing inputs—do more arrests, no matter the context—for outcomes—is the jurisdiction becoming more or less orderly?

When I write about problem-oriented strategies, I am trying to encourage people to take the latter view. After all, enforcement is costly—in money, officer time, and in risks that something goes wrong. That doesn’t mean don’t do it, but it means you have to be smart about how you do it. As Bratton put it, “what's important is the dose. You might have the greatest doctor in the world dealing with your cancer, but if he ODs you on chemotherapy, he's gonna kill you.”

3. Focus on the Risk of Consequences

The required dose of chemotherapy, of course, is proportional to the size of the cancer. When the tumor is big, you need a lot of it. Once it’s shrunk down, you need quite a bit less. At a certain point, the costs of the treatment outweigh the costs of the disease.

The same can be said of public disorder problems. When they get smaller, you can dedicate proportionally fewer resources to their suppression. This is not, though, a precisely linear relationship. That’s because if you dedicate too-few resources to them—as many cities did from 2020 to 2024—then they can become much bigger.

What you want to do is focus on deploying your resources such that the risk of consequences is held constant. This is a point I (like Matt) am cribbing from Mark Kleiman (although Kleiman gets the insight from others). Deterrence is a function of swiftness, certainty, and severity. Particularly if you want to keep your deterrence proportional to the offense, you want to amp up certainty and swiftness: the frequency with which a crime will result in apprehension, or the risk of consequences. You do that by dedicating large bursts of police resources to big problems, with the goal of “tipping” them down to a smaller size. Once they’re tipped, it requires relatively fewer resources to keep them at that small size.¹

Matt has an argument in his piece about cigarettes—we no longer really enforce smoking bans, everyone just kind of follows them. As I’ve written, that took a huge amount of effort. You had to up the consequences until enough people were deterred that the situation “tipped.” And of course, most smokers are not compulsively anti-social in the way many criminals are. For them, you need enforcement.

4. Address Disorder Generators

One of the key insights of criminology is that crime is highly concentrated in place. A small number of areas have features that are particularly conducive to criminal behavior. The same thing is often true of disorder. Poorly managed spaces create disorder. The problem is often a specific bar, or a specific house where people are running drugs or women, or a specific park where people like to use.

If you can deal with these “disorder generators,” you can have an outsized impact on the total level of disorder. Strategies include cleaning up, greening, improving lines of sight, and increasing surveillance temporarily to attract a more active crowd that can serve as eyes on the street. Again, you’re trying to shift a ratio, in this case the ratio of pro-social to anti-social behavior. You’re fixing broken windows.

Often these problem require some police support, but can mostly be dealt with through civil remedies. Citizens’ groups often have private rights of action under local or state law they can exercise to abate public nuisances. Public health authorities can compel the owner of a delinquent property to shape up or face monetary consequences. The cops are usually a backstop to such efforts, preventing violent retaliation. But the legal system and non-police actors are out front.

5. Treat Different Offenders Differently

Many low-level offenses are driven by high-frequency recidivists: guys who commit the same crime over and over again. When we reduce penalties for such offenses, we are more lenient towards infrequent offenders, like the high schooler who steals an eraser in a fit of pique. But we also become more lenient towards frequent offenders, like the guy who walks into the same store multiple times a day and steals everything.

Almost everyone agrees that the former does not deserve serious punishment—that’s why the criminal justice reform movement uses them as poster children. But almost everyone also will agree that the latter deserves, indeed requires, meaningful consequences. And people sour on reform laws when such consequences aren’t imposed—the reason New York’s bail reform has become so unpopular is that people find it intrinsically offensive when the 60-time offender is let out yet again.

One solution is to undo the reforms. But the more reasoned solution is for laws and law enforcement to treat different offenders differently. Laws should enable discrimination based on criminal history. And judges and prosecutors should take advantage of that capacity to discriminate.

Advocates would and will respond to arguments like this by saying that this is unfair: that criminal history enhancements deny people second chances, and that they trap people in the cycle of crime. The implicit claim here is that criminals aren’t really responsible for their actions; they’re just victims of the system. If Democrats like Matt want any hope of making their party tougher on crime, they need to tell these people to shove it. They are objectively pro-criminal sympathists of the worst sort. Stop having time for them.

6. Just Because Someone Can’t Control Themselves Doesn’t Mean Do Nothing

One response to point 1 (“don’t apologize for enforcing the rules”) is that some people can’t help but not comply with the rules. Maybe they are too poor not to sleep outside; maybe they are too mentally ill not to rant and rave in public. How can we justify laws, the argument goes, that hurt the least among us?

We should first acknowledge that many people who engage in disorderly behavior are doing so voluntarily—the guy who blasts his music on the subway is choosing to do so. That said, this is a perfectly reasonable point—people who can’t help their actions deserve to be treated differently from those who can. (See point 5.) It is hard to justify incarcerating someone just for sleeping rough.

But that doesn’t mean we can just do nothing—which is how many jurisdictions have responded to these challenges. Disorder generated by the homeless and the seriously mentally ill is disorder, and it still can have serious consequences. And if people are actually incapable of acting as citizens—of respecting the shared rules of public life—then the state is justified in trying to curb their behavior, even if not through explicitly punitive methods.

This is why the “tolerant containment” strategy for dealing with serious mental illness is so harmful. Mental health is a part of the public safety and order conversation, and providing more resources for those systems is vital. But vital, too, is no longer insisting a humane society just lets people live in their own filth. It doesn’t and shouldn’t.

7. Carceral Urbanism is Good, Actually

If you’ve ever advocated any of the foregoing views on the internet, you’ve probably been accused of being a “carceral urbanist,” a catch-all slur used by progressives who don’t want more housing think that such talk is a pretext for mass social control.

The phrase is much abused, but I think it’s a useful one. It has its roots in a Foucaultian² critique of the decarceration and deinstitutionalization movements of the 1960s and 1970s, which strove to replace control within the walls of an institution with controls out in the community. In particular, the phrase “carceral urbanism” invokes the idea that the built environment itself is designed to keep people partially controlled—that the city itself is an extension of the prison.

Like many Foucaultian arguments, I find myself responding here “yes, that’s true, and it’s good.” This is because, like many critiques, it suffers from a failure to evaluate the counterfactual. Semi-coercive design choices—like benches that stop people from sleeping or anti-teen sirens—are unpleasant. But they are less unpleasant than aggressive enforcement. And to the extent that public and private actors are willing to be good carceral urbanists—to design the built environment to deter disorder—they will be able to depend less on the formal mechanisms of coercion.

How do you do this? I like this Urban Institute guide, from a time when the Urban Institute was not terminally left-brained. Keep property clean; keep track of the chronicly disorderly; Impose curfews and sales limits; make spaces slightly less pleasant to be in.

Of course, such carceral urbanism is often only necessary insofar as there is not enough enforcement. I don’t like that I can’t just hang out in Starbucks anymore—and it’s bad for the whole city that I can’t. But it’s part of a strategy for disorder suppression.

More to the point, much like with the people who argue against frequent offender laws, you cannot keep listening to those who tell you such interventions are bad and mean for theory-brained reasons. You need to show them the door.

This gets at a broader point: it is no longer the 1970s. We know how to make cities clean and safe. The strategies are pretty straightforward. The impediments today are not knowledge, but a) resources and b) ideology. We have to be willing to do more to build a better, more effective crime control apparatus. But we also need to stop taking seriously arguments against its reasonable use. Those lead to profound and unforced social errors.

1. This paragraph is, more or less, a summary of the argument of Kleiman’s When Brute Force Fails.

2. The adjectival form of Foucault, as in French philosopher and social critic Michel Foucault.