Riot Control

How to stop a mob

Etienne Girardet on Unsplash

A few days ago, I had a chat about the riots here with Franco-American political scientist Arun Kapil, whom you may remember from our French Election Twitter Summits. He kindly gave me permission to publish our exchange:

Arun: France is in a potentially dangerous place right now, and with a problem that has been festering for decades and appears insoluble.

Claire: Are you referring to the fact that the police executed a 17-year-old kid and no one believes this was just one bad apple? Or to the fact that large numbers of French youths think the appropriate way to respond to this is by sacking and pillaging their own neighborhoods?

Arun: The problem, of course, is the banlieue-ghettos, the people who live there, the police, politicians, and you name it. The police are a huge problem. It’s more than a bad apple or two.

Claire: When you say France is in a “potentially a dangerous place right now,” do you mean that these riots are potentially dangerous, or do you have something else in mind?

Arun: It’s dangerous because the situation could get out of control and with destruction on a level we have not seen.

Claire: They need to put overwhelming force on the streets.

Arun: Are you out of your mind?

Claire: No. It’s common sense.

Arun: Why not shoot to kill while you’re at it? Put the fear of life in those 14-year-old branleurs!¹

Claire: No, just restore order. People will stop rioting if there’s an adequate number of police (or the military) on the streets. The worst solution is what they’re doing now—enough force only to enrage and create more incidents in which cops and rioters are injured or killed.

Arun: What is needed to stop the rioting is someone or some force that will turn down the temperature. Problem is, that person or force appears not to exist. It certainly isn’t Macron or [Interior Minister Gérald] Darmanin, that’s for sure. And the more you flood the streets with police, the more the branleurs will come out and clash with them—and with Black Blocs joining in—as that’s what they like to do. The military? Don’t even think about it. One thing is certain: If even one more “jeune”² gets killed by the police, all hell will break loose.

Claire: That’s why this needs to end. And to that end, you need a massive number of police or the military. If you put enough on the streets, the violence ends. It’s Riots 101. It’s not an extreme idea.

Arun: No, Claire. No, no, and no! There is no such course as Riots 101.

Claire: There is, absolutely, a massive body of empirical research on crowd violence.

Arun: Massive empirical evidence? Such as?

Claire: I’m writing about it right now.

Arun: Look forward to reading.

---

The U-shaped Curve

Scholars in Europe and the United States have done extensive work on the relationship between riots and repression. (The latter is the academic term for “putting a ton of cops on the street and arresting the rioters.”) The answer to the question, “Does repression work?” is yes.³

But it’s not a linear relationship. It’s not true that for each additional unit of police on the streets, you get one less unit of rioting. In most cases—there are always exceptions—there’s a U-shaped curve. Repression can lead to an increase in rioting—which I think is Arun’s concern—if it’s perceived to be heavy-handed and arbitrary, and if this causes more of the public to sympathize with and join the rioters.

This contradictory impact is called the “repression paradox,”⁴ and recent research suggests that the paradox is heightened by social media. During the Arab Spring protests, for example, and also during the Occupy movement, police (or state) repression had a strong paradoxical effect on protest diffusion; the mechanism of transmission was social media.⁵

So a little bit of violent policing—especially if videos of the cops behaving brutally go viral—will make things worse. But if you put enough calm police on the ground, the riots stop. Many of you will say, “Well yes, of course.” But as my exchange with Arun suggests, this isn’t intuitively obviously to everyone.

Why exactly do the riots stop if you put enough police on the street? Rational choice theorists have done a lot of work on this topic. Karl Dieter Opp and Wolfgang Roehl, for example, argue that rioters respond to incentives. If they perceive the likelihood of being arrested as high, they will conclude that the cost of rioting exceeds its benefits.⁶ A riot, remember, is made up of a series of individual crimes. But it’s also a tipping-point phenomena. The cost of participating—the risk of the cops kicking the snot out of you, being arrested, doing time, having a criminal record—decreases as the size of the riot increases. If I walk outside my apartment now and set the nearest car on fire, there’s a high likelihood that I’ll be arrested. But if thousands of other people are doing the same thing, the police will be distracted and overwhelmed, and there’s a good chance I’ll get away with it.

Riots end, therefore, when there’s enough law enforcement on the streets to suggest to individual rioters that there’s little chance they’ll get away with it. To that end, it is much more important that many arrests be made than it is for the punishment to be extremely severe. The credible threat of arrest is usually enough to break the spell of the mob and cause people to disperse.

But if it isn’t, you need to start arresting everyone involved. It actually works quite well to arrest them, hold them overnight, then let them go. It clears the streets and breaks the momentum of the riot. There’s no need to ruin a bunch of over-adrenal teenagers’ lives by sending them to prison for years. You just need for arrest to be a certainty if they don’t go home. There’s a massive amount of criminological research indicating that no matter the crime, the severity of the punishment is far less important than the certainty of it.

Trying to stop riots with an inadequate force—and particularly, an exhausted and politically radicalized force that has long since dehumanized the rioters, as the French police have—is a terrible idea. France would literally be better off putting no cops on the street at all. The way they’re approaching riot control is a recipe for a paradoxical effect. The cops in the video below need to be taken off the streets before they kill someone. They’re certainly not helping to keep the peace.

---

Riots proceed in stages. The initial explosion usually occurs among people who are genuinely outraged by the trigger event. These emotions burn out quickly, however, and after the first day or two, the first rioters are often replaced by replaced by political agitators and opportunists. The initial rioters also tend be teenagers and younger adults. They engage in the destruction of property. Later, older rioters—often substantially older—replace them, and they’re the ones who initiate the looting. If the riot isn’t stopped at this point, other members of the community will then join them, because they figure they’d be fools not to get their share of the loot.⁷

The opportunists will generally keep going as long as they’re allowed, and the longer it goes on, the more damage to life, limb, and property there will be. At this stage, if the state fails to control the riots, vigilantes step in. Vigilantes are neither trained nor equipped to put down riots without using deadly force. Bad things happen. And if the cops are exhausted, outnumbered, and frightened, this is when they’re apt to lose their minds and kill someone—or just say or do something outrageous, which then becomes a viral video that brings more people out on the streets.

The key to stopping a riot effectively, then, is a massive law enforcement presence, on the streets, at the first sign that a riot is gathering. Critically, it should not be a threatening or a hostile presence. The police should appear relaxed and friendly. Ideally, they will be wearing their normal uniforms, not the Robocop-Gladiator ensemble, and they will not be in a phalanx, truncheons at the ready, but talking to the people around them calmly. They should not be spoiling for a fight. They should just be there—everywhere you look. That is usually enough to stop a riot before it starts.

Below is an example from a far-left march against police violence earlier today in my neighborhood. This is not—to say the least—the calm, relaxed, and confident demeanor one hopes to see the police display in a tense and riot-prone crowd:

---

What if it’s already started? Restraining as many rioters as possible, using as many cops as possible, is the safest and fastest way to end it. You don’t have to keep them for long—and you shouldn’t, in the case of teenage delinquents who just got over-excited. You just need to restrain them until everyone settles down. (The professional agitators, however, like the Black Blocs, need to be put away for a long time.)

Police must also be trained to appreciate that there are differences in motivation among members of riotous crowds. Some participate directly; others merely observe; some lead, some follow; some are there on impulse, others have carefully planned to exploit the situation. Some act as counter-rioters and try to dissuade the rest of the crowd from violence. Police should not treat the crowd as a homogenous mass. They should prioritize the arrests of the leaders and make allies of the counter-rioters. Failing to do so, again, raises the risk of inflaming the situation. To this end, if your police unions are ravening about “hordes” and “vermin,” it’s a fair bet that they are not making fine distinctions among the crowd, not using these strategies, and thus not as effective as they ought to be. Here’s another example of the way police shouldn’t behave:

There are a lot of riots we never hear about because they never happened, and they never happened because experienced and professional police forces use these strategies, if there are enough of them to do it. During the 2004 Republican National Convention in New York, for example, the police arrested more than a thousand people and then let them go. There was no riot. It’s impossible to prove there would have been a riot otherwise, but the initial signs were there. Nor was there a riot when Trump was arraigned in Florida. Nor when he was arraigned in New York. Nor when he was found liable of sexual assault in New York.

I’ll concede one point: The riots here died down despite the lack of overwhelming force on the streets and despite the police’s lack of professionalism and calm. So “calm, overwhelming force” is not a necessary condition for ending riots. But it sure helps.⁸

---

Smelser’s Gateways

The nature, size, and conduct of the police force is only one part of the formula. In 1962, the sociologist Neil Smelser proposed that for a riot to emerge, six elements were crucial:

• Structural conduciveness. These are the things that make riots possible, such as spatial proximity.

• Structural strain. Social malaise, economic pain, or political tensions that create a general sense of unease or dissatisfaction.

• Generalized belief. A widely shared interpretation of the cause of the strain.

• Precipitating factors: The spark that lights the fire.

• Mobilization for action. The process of preparing to riot: securing resources, communicating and coordinating with others, taking the initial steps.

• Failure of social control: How authorities react—or don’t.

Subsequent research suggests he was right about most of this, with one important exception. Researchers have looked over and over for evidence of a reliable correlation between “structural strain” and rioting. They haven’t found it. Nor have they found a relationship between rioting and individual strain. There just doesn’t seem to be a relationship between being deprived or frustrated, personally or as a group, and participating in a riot.

Although that’s a surprising finding, probably it shouldn’t be. After all, young men are overwhelmingly more likely to riot than middle-aged women. But what kind of social strain would affect only young men? Perhaps the prospect of being drafted, I suppose, but that’s a very limited set of riots. Do middle-aged women not suffer from the afflictions said to cause young men to riot?

Generally, though, the other five elements must be present, and you can prevent riots by interrupting one of these gateways. There are many ways to do this.

---

Generalized beliefs

French opinions about the recent riots are polarized along lines quite similar to those in the US in the wake of the murder of George Floyd. The left in France sees the violence as a function of structural strain. The right sees it as a failure of social control.⁹ These ideas are often discussed as if they’re incompatible.¹⁰

The way we talk about riots tends to contribute to the generalized belief systems required for riots to start. There’s no way, in a democracy, to demand that everyone just shut up and stop talking about riots. But it wouldn’t hurt for France—or any other riot-prone country—to have a discussion with itself about the value of discussing riots and the events that precede them calmly. In the era of social media, the “someone or some force” that can “turn down the temperature” is not just a politician—it’s all of us.

The generalized beliefs of France’s rioters last week, for example, seem to have been something along the following lines:

• France hates people like us.

• The police are indifferent to our lives.

So if you want the rioting to end, don’t reinforce these beliefs. This is common sense, which leads me to suspect that the people here raving about hordes and laxisme don’t want the riots to end so much as they want the pleasure of ranting about laxisme.

There’s nothing to be done about what the public says at moments like this, save to point out that it often does more harm than good. But for God’s sake, government ministers and the police unions should know when to keep their traps shut, and as Arun notes, the ones France has right now just don’t.

I don’t mean to suggest that 14-year-olds in Nanterre study the newspaper columns over their morning coffee and conclude, “Why goodness, it seems to be true. The Interior Minister is a neocolonialist brute! The only remedy is to loot the Conforama. J’attaque!” But I do mean to suggest that you can hardly expect teenagers to master their impulses and control their passions when the leaders and the adults around them are hysterical.

Attempting to change the generalized beliefs of a riot-prone community is not a practical short-term strategy. It’s not easy to change a whole culture. But it’s useful to see where these beliefs come from, and if understanding this convinces a politician or a pundit to think twice about reinforcing these beliefs, more’s the better.

Something could be done, however, about massive social media campaigns, aimed at aggravating riots, sponsored by hostile powers, usually Russia. If you’re a French teenager of North African origin who uses social media, you’ll be relentlessly exposed to the message that the French hate people like you and think the police should shoot more of you. Of course, some of this is real. But a lot of it is fake. On Twitter last week it was nearly impossible to find real news about France: There was nothing but bots and trolls, all of them trying to make things worse.

Macron has been pilloried for suggesting that the state needs the power, in an emergency, to switch off social media. He was immediately accused of outrageous authoritarianism. But he shouldn’t have been. He’s right. Social media spreads riots. Hostile powers use this as a weapon against free countries. How to prevent this is a debate every democracy needs to have.

---

If police presence is an aspect of social control and commentary about the riots helps to shape generalized beliefs, there are still three more gates upon which pressure may usefully be exerted: structural conduciveness, precipitating factors, and mobilization for action.

The majority of riots begin at or near high-density housing or major traffic thoroughfares. They usually begin after five in the afternoon, or on weekends. They continue in 24 hour cycles, peaking just after midnight, then dropping to their lowest intensity between 4 and 8 am—rioters need to sleep, too. The intensity then increases slightly until noon, then steadily until it again hits a midnight peak.

A riot is most likely to begin when rioters learn of a precipitating incident while they’re physically together, for example, in a crowd outside the stadium after a football game. This is when the idea of rioting is first discussed, and where they decide what they plan to do, when, and where. Whether they carry out their plans will depend on their schedules: Do they have transport? Do they have other commitments?¹¹

We saw this recently in France. I recommend reading this entire interview with Le Monde reporter Luc Bronner, whose account of the rioters’ behavior confirms all of Smelser’s research. Rioters here didn’t, for example, just spontaneously start looting. They organized the raids in advance and came from all over the region to carry them out:

In Bondy (a northeast Paris suburb), I witnessed the organized looting of [home furnishing chain] Conforama. Information had been given in advance on social media, and when the time came, around midnight, a group of around a hundred people attacked the shopping center. The police, including from the RAID tactical unit, tried to block them. They were not entirely successful. I witnessed a procession of vehicles (with licence plates from all over the region) pulling up in front of the store to load up flat screens, Nespresso machines and boxes of all kinds.

He confirms, too, that they did it because they felt they could get away with it:

Did you observe a feeling of “omnipotence,” a conviction that they will be able to escape legal consequences, among the groups of young people you came across?

Yes, the feeling of omnipotence struck me. I’ve seen kids thrilled to stand up to the police, even elite police units. One mayor told me he’d seen graffiti reading: “We are the law.” So, through this violence, there is a form of territorial takeover by groups of a few dozen to a few hundred young people in each public housing development. The people I spoke to told me they were astonished to discover that adults had no control over these teenagers either.

Similarly, the rioters were all male:

Do you know if many young girls were also involved in the riots?

This is an essential question. From what I've been able to observe and from what officials and police officers have told me, it was almost exclusively young men. Many were minors, sometimes 14 or 15 years old, according to observers.

---

To prevent and quell riots, all of these processes should be interrupted.

Around the world—and not just recently in the United States, but everywhere, for as long as the phenomenon has been studied—the most common cause of a riot is a police shooting or a similar abuse by the police. To prevent riots, police must be trained not to kill people for no good reason. Not only is this morally obvious, it’s essential to public order. Everyone, everywhere, just hates being shot by the police. They hate it far more than they do being shot by a fellow civilian. This may be illogical, but it doesn’t matter: It’s a sociological fact.

There’s a certain kind of critic in America who will reliably point out—correctly—that the number of black men killed in the United States by the police pales in comparison to the number of black men killed by other black men. Therefore, they conclude, all the hullaballoo about Black Lives Mattering is phony. But they’re wrong. Wise people work with human nature, not against it, and it’s abundantly obvious that people have far less tolerance for being killed by a cop than by a criminal. Whether this is rational isn’t important. If you don’t like riots, train your cops well.

It’s in the nature of policing, however, that there will be accidents, no matter how well the cops have been trained. Sometimes there will be justified shootings, too, that don’t look justified on the video that goes viral—perhaps it doesn’t show what happened right before, or what was taking place off camera. When this happens—and when the video begins to circulate widely on social media—a riot is now predictable. If you want to prevent it, you need to spring into action and cover with cops the areas where people are likely to gather to plan a riot, like high-density housing in neighborhoods where people of the same ethnicity as the victim live. The cops should not be aggressive or invasive. But they should be there, as far as the eye can see. Domestic intelligence services should closely follow developments on social media to see if people are organizing a riot. The cops should beat them to every place they decide to go. The decision should never be made to let the rioters “burn off some steam,” which is a common but deadly mistake. It’s far easier to stop a riot from starting than to stop one in full swing.

If despite this riots seem to be gathering, don’t wait for days to do what the French ultimately did last week. Declare curfews immediately. And yes, shut down social media. It’s important to explain that this is the plan well before there’s a riot and also to explain why it’s the plan.

No one, including the rioters, really wants a riot to happen. A thousand-odd teenage boys here now face prison and lives blighted by serious criminal records. For their sakes as much as their victims’, they should have been calmly but firmly discouraged from being idiots. Teenage boys riot because they get caught up in the excitement, all of their friends are doing it, they see videos on TikTok that show them how and make it look cool, and they think there will be no punishment. Teenage boys are idiots. Their brains aren’t developed enough to assess risk. They should have seen a sea of cops the moment they got it in their heads to riot. That they didn’t represents a failure of the state.

There should be a lot less handwringing here about immigration and whether France has become “ensavaged” and a lot more handwringing about the moment of state failure—when the police were not where they should have been, and where, when they were been present, they behaved like this:

---

No matter how legitimate the grievance, the state has an absolute obligation to prevent and end riots. In the case of the recent riots in France, the grievance was real. No self-respecting country permits the police to pistol whip and then kill an unarmed teenager for a traffic offense. When the police then then behave in the way they are in the videos above, it rubs salt in the wounds.

But no self-respecting country allows teenage boys to run around looting, vandalizing, setting fires, and tyrannizing their neighborhoods, either.

Riots can be as destructive as any natural disaster. They’re extremely dangerous. Every year, thousands of people are killed in riots around the world. By one early estimate, last week’s riots in France did a billion dollars worth of damage, which comes right out of the pockets of French taxpayers. If you allow people to run riot because the state committed an injustice, you’ve added a second grievous injustice to the first. A state that permits this to happen—and to happen frequently—is failing in its most fundamental duty.

The political consequences of riots are well-studied. They cause electorates to swing right. In a country like France, where a dangerous far-right force is standing in the wings, salivating, it’s not just incompetent to fail to control riots. It’s deeply dangerous, politically. The malevolent effects of riots like those France has recently experienced go far beyond the economic. They poison relations between immigrant communities and the rest of France. That’s unfair, because these communities are doing nothing the rest of France doesn’t do, but a fact nonetheless. Uncontrolled riots are less caused by ethnic hostility than the cause of it. Immigrants to France and their children will suffer by association for years because of it.

It doesn’t matter why people are rioting or who is responsible. The state needs to stop it, and to do that, it must have a police force big enough that if needed, it can be on the streets in overwhelming numbers. France’s police are exhausted, poorly trained, and terrified they’ll be overrun. There simply aren’t enough of them. You can see this in the videos above. They’ve come to hate the people they’re supposed to protect. They need a vacation. They need help. They’re acting like animals because they’ve been asked to do the impossible. This is unfair to them and to the public. That France does not have adequate force at its disposal—despite one riot after another—is a serious indictment of its government.

Yes, it would be costly to create a police force big enough reliably to prevent mobs from whipping themselves into a frenzy before law enforcement arrives. But it would be a whole lot less costly—in money, lives, and social harmony—than the riots themselves.

1. Branleur: wanker, literally. (Twerp, knicker-sniffer, mutt, chode, spanky, mouth-breather … )

2. “Youth.” (As in English, it’s the euphemism of choice for those wankers, twerps, knicker-sniffers, mutts, chodes, spankies, and mouth-breathers. )

3. See: The Differential Policing of American Activism, 1960 to 1990; Repression and Mobilization: Insights from Political Science and Sociology; The Dark Side of Purpose: Individual and Collective Violence in Riots; Dynamics of Repression and Mobilization; Breakdown Theories of Collective Action, Public Order Policing Approaches to Minimize Crowd Confrontation During Disputes and Protests in Australia.

4. See: How social media matter: Repression and the diffusion of the Occupy Wall Street movement, Ibid.

5. Ibid.

6. See Repression, Micromobilization, and Political Protest. How Riots Spread Between Cities: Introducing the Police Pathway.

7. Riots: Behavioral Aspects.

8. France dispatched 45,000 police to quell the riots, but only after they were already well underway. By the time they did this, riots had already broken out in more than 500 cities, towns, and villages—spread by social media. Of that 45,000, only 5,000 were in Paris. That’s not nearly enough. Had they put them on the streets at the outset, it might have been sufficient to prevent them, but it’s far too few for a riot in full swing. It should be impossible for the rioters to overwhelm the cops—visibly, unmistakably impossible. You must not have the police wind up in hand-to-hand combat with the rioters, with the two sides seemingly equally matched.

People are advised, if confronting an aggressive dog, to be calm, stand tall, and stand still. Don’t try to run, because a running animal looks to a dog like prey, and he will chase you. He can’t help it: It’s a canine instinct. Likewise, if you’re confronting a disorganized group of aggressive young men, don’t organize yourself in a way that suggests a tribe. Males of that age are subject to similar ancestral instincts, and if you allow them to think they can form their own tribe and overpower you, they won’t be able to resist trying. It’s a human instinct. If they end up confronting one another in a configuration that evokes tribal warfare, both police and rioters are apt to lose their minds and do things they regret.

9. For example:

The situation is simply explained by a single word: “laxity!”

For years, we have heard the jeremiads of good souls and other salon dwellers who strive to find a thousand excuses for the delinquents of the suburbs, accusing society of being the culprit. At the same time, the media practices omertà about the multiple attacks on police and firefighters with mortars and even firearms; the number of police officers and gendarmes injured, killed by so many assassins at the wheel they can no longer be counted.

Each time we hear soothing remarks from the government highlighting the calm of the police who suffered the attacks. Delinquents and other scum quickly understand that they risk nothing, and above all that they are the kings of the street to impose their laws and their drug traffic.

This laxity is the real cause of a situation that is worsening, day by day, in the absence of repression—controlled, but adequate to these challenges. The policy of social workers isn’t the solution. Years of that policy have led us to today’s riots.

Let’s be clear: If the State does not repress these riots by force—if necessary by giving the order to shoot the rioters—it should not be surprising that our fellow citizens take matters into their own hands. There will be deaths for sure, and the good souls will be surprised, in lamentations!!!

Laxity generates chaos.

10. For example:

Nothing seems to be able to stop the spiral of violence and riots in the French suburbs. … France looks more and more like a powder keg, where it only takes a spark to ignite everything. A France where there is a real mistrust of everything that represents authority. A mistrust that draws its source from a feeling of exclusion and injustice suffered by part of the population, the part that lives in working-class neighborhoods. Neighborhoods that are the glaring expression of the failure of several decades of public policies of the various governments that have succeeded one another in France. The multiplicity of initiatives deployed over the past forty years has, in fact, only increased social inequalities and between territories, creating a two-tier France. Those who have everything or almost everything. And those who have nothing or almost nothing: Those who manage, who fight to survive, even if it means tipping into illegality, the one who is a victim of unemployment, the one who is marginalized and abandoned ...

11. See, e.g., Riots: Behavioral Aspects: Precipitating Incidents And Underlying Conditions.