Chris Blattman Profile picture
May 11 33 tweets 12 min read
Medellin, Colombia has 100s of powerful gangs, but still went from one of the most violent cities in the world to homicide rates far lower than that of St Louis, Baltimore or Detroit. What happened?

The story here: great success, but at a terrible & hidden price.

A 🧵
Americans think of shootings as an individual problem: greed, passions, feuds, and hot reactive thinking drive killers.

That’s true to an extent.…
But this overlooks something important: that, in an awful lot of cities, crime is organized. That makes some killings calculated and strategic, not simply crimes of passion or reflex.

A gang leader I met in Chicago once told me how he became a killer.
Someone robbed his group of their drugs. He looked weak. He knew he had to show he was resolved. Construct a reputation to protect him and his group. So he killed and killed until no one robed him anymore. That wasn't passion. It was cold, cruel, and strategic.
Here's another thing: criminal groups, being strategic, prefer peace. No one sells drugs in a gang war. So groups have incentives to make peace.
Take Medellin, where every low and middle income neighborhood has a street gang known as a "combo". There are a lot of them:
One leader told us about a brawl that broke out over a pool game in prison. Men got shot. The violence and revenge killings started to spread to the street. Every gang in the city lined up between one side or the other to prepare for war. But that billiards war never happened.
The gangs didn't want a war. And the big crime bosses definitely didn't want one. They'd lose the most. From prison they orchestrated sanctions, mediation, & commitment to peaceful deals. They walked street gangs back from war. They call their peacemaking confederation La Oficina
La Oficina is a little like the UN Security Council. Dominated by a few powerful bosses. Selective in their enforcement of the peace. Only partly effective. Highly unequal. But they do help keep the peace a little. They call it El Pacto del Fusil—the pact of the machine gun.
I talk a lot about this in my book, Why We Fight. Like gangs, nations, political factions and ethnic groups don't want to go to war. When they do fight, it's for similar reasons. And the solutions resemble one another too.

But that's not this thread.
This thread is about the terrible trade-off that governments make for peace. Because peace—meaning an absence of prolonged battling—isn't always just or pleasant. It sometimes means unpalatable deals.
Take a look at this figure of homicides in Medellin every month of 2019. They started to spike as the factions began to skirmish and start a war. La Oficina was fracturing and needed help finding a deal.

What happened at the red line?
That's the week the govt coincidentally shifted each criminal leader from their prison. Scattered about, they suddenly found themselves in the same holding cell for a few days. Coincidentally, a trusted criminal mediator was arrested & placed in the same cell. Killings plummeted.
Informally and formally, states all over the world try to help gangs keep the peace.

President Bukele in El Salvador (and his predecessors) are famous for organizing secret ceasefires and pacts.
Mexico’s single party state once did the same with cartels. Why wage a costly fight when you can compromise & share spoils?
But democracy makes this hard. It’s unpopular (plus it’s harder to make deals with rotating politicians.)
In Brazil, the state inadvertently helps criminals keep the peace by ceding prison control to gangs. This can help criminal heads strike truces, work out differences, and even fuse their organizations. These (imperfect) peacemakers are we know as the PCC. and the Red Command.
US cities do this too. They call it focused deterrence. City authorities meet with gang leaders & pledge swift, certain, and severe punishments for future homicides.
Comparing cities that do this to those that don’t suggests the threats have some effect.…
Violence interrupters also mediate between rivals. Usually they leap between feuding individuals, but some also try to mediate between leaders and arrange pasts. Comparing neighborhoods suggests this too works reasonably well.…
So where's the terrible tradeoff? Sounds great so far!

The trouble is that helping gangs make peace can make them more profitable and stronger. Especially the criminal syndicates who provide this valuable conflict management service to the smaller gangs and groups.
People don’t buy as many drugs or pay extortion in the middle of a gang war. So peace is an opportunity to sell more dope, extract more protection money, and gain legitimacy from local people.
Also, the same syndicates that help keep peace can also coordinate to form cartels—in the economic sense of the term—coordinating everyone to sell as monopolistic prices. Certainly this is what’s happened in Medellín.
So mayors face a tough choice.

Voters hate homicides. Why disturb any Pacto del Fusil through crackdowns on the gangs? They’d just find their cities engulfed in turf wars so bloody and destructive that they undermine their election prospects next term.
Yet leaving organized criminals be makes them powerful.
Here, for example, is a heat map for Medellin, showing who (on average) is more likely to help you resolve a dispute with your neighbor, address a theft, or protect your street: the state (blue) or the local gang (red).
Here’s how gang (combo) loyalty and legitimacy correlates with this criminal governance. In many neighborhoods, the gang has about as much trust and respectability as the state.
This is a high price to pay for lower violence. It might be the right price to pay. Every city has to make their own choice.

But few mayors or voters are even aware of it.
Most police & mayors have information on one thing: crime and homicides. They get daily reports, down to the level of street corners. Mayors can hold their staff o& police accountable for improvemenrt. If they don’t, newspapers report them every week. Voters pay attention.
But they have no way of knowing whether criminal groups are growing stronger or weaker. Few collect data on gang strength or drug profits. Fewer track when gangs provide local security, or ask citizens whether they find the gangs as legitimate as the state.
I want to get to how we can collect these data, but first here's a polity note with more detail, put out with support from @poverty_action and @FCDOREDGCSD and the SOC research programme at U Birmingham with @hamarquette…
This is work with @SantiagoTobon, @BigBigBLessing, @gusduncan & @JuanPabloMesaM.

Need help? We and @poverty_action are working with more and more cities to do this. Call us.

Here's where we give tips on how to do this in other cities. CAREFULLY.…
Now for the traveling salesmen pitch:

For more on how to counter violent gangs, ethnic groups, political factions, nations, and Putins, do check out my new book.

@FT, @WSJBooks, @latimes, @amazonbooks and others all recommend it as a best new book
Check out the other working papers in this series…
And for more posts like this, follow me on Twitter or sign up to receive blog posts by email.

(Yes, I still kick it old school in the blogosphere.)

• • •

Missing some Tweet in this thread? You can try to force a refresh

Keep Current with Chris Blattman

Chris Blattman Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!


Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @cblatts

May 8
If you have ANY interest in conflict & development, check out the student-run Root of Conflict podcast @PearsonInst and @HarrisPolicy. Every amazing policymaker or scholar that passes through our doors gets interviewed. Here are a few examples.…
Nina Jankowicz @wiczipedia on disinformation…
@Afrahnasser from Human Rights Watch on Yemen…
Read 6 tweets
May 8
Happy mother's day! And the natural question for the occasion: Would there be less war in the world if moms were in charge?

The research says 'probably yes', but perhaps not for the reasons you think.

A 🧵, starting with the Peloponnesian War and a play called Lysistrata. Image
Women in Athens couldn’t vote, but the conflict was killing their kin. So Lysistrata called on Greek women to deny their husbands sex until Athens & Sparta stop fighting.

By the end of the play the men are... frustrated. Peace ensues.

(That's the most NSFW this gets folks) Image
Now, the play isn’t an enlightened treatise on gender. It was a bawdy comedy, written by a man, acted by men, in a society dominated by men, playing up stereotypes for hilarity.

Still, Aristophanes captures a common view: if women were in charge, we wouldn’t fight so much.
Read 21 tweets
Apr 17
There are some furious reactions to this Noam Chomsky interview. There are problems with Chomsky's line of argument, and the reactions worry be too. A thread.
@Noahpinion points out: whether to settle and on what terms is mainly a Ukrainian decision. There's an arrogance in Chomsky's "we" to which too many armchair analysts fall prey. True. I'm prone to it myself.
@Noahpinion and others also point out that there are Russian politicians and talking heads signaling that Russia is unlikely to settle on any terms less than the extermination of Ukrainian identity.

Read 25 tweets
Dec 1, 2021
Important study. A lot of people are focused on comparing these results to cash and other interventions to say this combo is more successful. Maybe, but that’s not a the only difference. In my mind, the big difference is in the kind of people receiving the assistance. Ultra-poor.
The main thing we have learned from all these studies is “read your economics 101.” An intervention relieves poverty in a sustained long term way only if it relieves a constraint that doesn’t otherwise get relieved. Programs have sustained impacts when he control group stagnates.
In many cash transfer studies, effects dissipate over time not because the initial impact wasn’t large, but because the control group caught up. They were poor but not especially so. The constraints they faced were not insurmountable. And so with time the control group converged.
Read 7 tweets
Nov 9, 2021
I find the UAustin ridicule cringeworthy. The fact that I’m at a leading policy school at one of the more “freethinking” places and I can’t think of a single conservative on the faculty (and a handful—at most—in our other professional schools) ought to be deeply concerning.
Young conservatives are reading our tweets and are being dissuaded from academia. Young progressives are reading us and confirming their view that it is perfectly fine to dismiss and ridicule ideas they don’t share (even if they’ve never thought about them deeply).
Whenever I hear students in my school discuss conservative viewpoints, I am struck by how naive, astonishing, false, and elitist they are. This should be worrying most of all to those who care about progressive causes, for these kids do not understand their opponents.
Read 10 tweets
Aug 28, 2021
A behavioral scientist declaring behavioral economics is dead and will not be a respected field in 10y. Some interesting points, but (as a non behavioral economist) it seems like a super narrow conception of the field…
Mostly it hinges on things like loss aversion and priming failing to replicate. Fine. But if we think of behavioral econ as the expanding our understanding of how we make choices as individuals and groups, then the field seems far from dead.
Economists begin with models of people maximizing expected utility over mostly material goods. You can think of behavioral science as introducing:
1. Other preferences
2. Failures to get beliefs right about states of the world
3. Failure to maximize
Read 16 tweets

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!


0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy


3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us on Twitter!