05 May 2025

Fighting Organized Crime

Hedgehog thinking is not the right approach to law enforcement led drug enforcement. 

Also, some solutions, like reducing easy access to guns (and making it harder to smuggle guns abroad), or legalizing drugs to reduce illicit organized crime supplied demand, are politically unpopular with conservatives and centrists, who may have median voter political power. 

International drug trafficking dominates the conversation on organized crime, but equally common and serious are urban systems of organized crime—criminal groups focused not on exports or transshipment, but on dominating local markets, neighborhoods, and politics. 
When policymakers do pay attention to this problem, they consistently make the same mistake—believing there are best practices. But systems of organized crime are not all alike. There is no one blueprint or general solution, and so following the latest fad is unlikely to deliver the desired result. Instead of trying to copy the quasi-mythological success of a Giuliani or Bukele, policymakers need to understand what kind of organized crime problem they have, what capabilities their city possesses, and what tools are appropriate to the circumstances. 
This paper looks in-depth at how organized crime is organized in three cities: Chicago, Medellín, and San Salvador. It also considers New York, Bogotá, and Port-au-Prince. 
I argue that the primary driver of their organization and incentives is their source of criminal revenues. A second major driver of behavior is the degree of criminal political organization, which broadly-speaking takes three forms: atomized individuals, fragmented groups, and competing confederations. These forms are not just the product of their revenues, but are also the result of decades of competition with the state and one another. 
What policy tools will work hinges on this diagnosis. I look at the evidence for a range of standard policies—from crackdowns to street outreach—and explain why we can expect them to have wildly different impacts depending on the context.
Christopher Blattman, "Bad medicine: Why different systems of organized crime demand different solutions" 2024 IMF & IDB Conference Crime in Latin America (November 9, 2024). The conclusions are as follows:
5 Conclusions

5.1 Terrible trade-offs

In tackling everyday crime, the tradeoffs are obviously difficult. Having more and better-trained security forces usually lowers crime. But this comes at a steep fiscal cost, collides with systems and cultures that are resistant to change, and buffets some continuities with heavy-handed tactics.

In cities with established gangs and organized crime, however, mayors and police chiefs wish their tradeoffs were so easy. At least in the short run, there are almost never options that reduce violence, drug use, criminal power, and political corruption at the same time.

Achieving some of these objectives often comes at the expense of others.

Take crackdowns and mano dura approaches. These policies can “work” in the sense that they disrupt illicit business, fragment powerful criminal factions, and loosen criminal control of neighborhoods and politicians. They use existing police capabilities, without asking forces to do something new or different. And the effects are immediate and easy to splash on the pages of newspapers. The public often supports these efforts, at least at first. All this help explain why these policies are often popular among police and politicians.

As we’ve seen, however, heavy-handed and unconditional tactics bring new challenges. In the worst case, they lead to a more fragmented landscape. Decapitating criminal organizations can destabilize a city’s fragile criminal systems, launching inter-factional wars of succession and dominance. And when state control over prisons is weak (as in El Salvador before 2019), competition and mass imprisonment has often prompted criminal confederations a chance to consolidate and strengthen.

Strengthening state capabilities is no panacea. Cities like Chicago managed to completely dismantle their meso-level structures. Perhaps El Salvador has accomplished the same. But Chalfin and McCrary (2017) in the U.S., at least, the effect on violence has been mixed and at times unstable. Lower control by confederations stopped large-scale inter-factional warfare. It also eliminated these rich and powerful structures from the social and political scene. But the remnants of these former empires—the fragmented, anarchic landscape of hundreds of small street gangs—has proven almost impossible to eradicate, and has been vulnerable to violence spirals that are hard to control.

A more savvy and focused approach—cracking down on the most violent gangs, and incentivizing the meso actors to maintain a city-wide peace—carries trade-offs too. Medellín has managed more than a decade of incredible peace, with homicides dropping to a historic low of 12 per 100,000 in recent years. But its criminal factions have leveraged this influence to enrich themselves and city’s gangs—fixing retail drug markets and earning monopolistic profits, growing the ranks of drug consumers and addicts, expanding into legal markets like consumer products and security-provision, and channeling some of these winnings into political influence and police corruption.

In previous work my collaborators and I have called this the “terrible tradeoff” between violence and criminal control. Ben Lessing has also called it a trilemma. Policymakers want low violence, weak criminal groups (with low control over the state), and low drug flows, but they can seldom achieve all three at the same time. The more a city tackles drug flows and weakens criminal group control, the fewer incentives criminal groups have to maintain peace, and the more potential for violent spirals.

While these trade-offs are real, there is some reason for hope. 
First, Chicago (like many American cities) experienced an improvement in all three during the great homicide decline of the late 1990s to roughly 2014. They are trying to recreate that success now by combining more and better policing with focused deterrence and funding a “civilian architecture” for community violence reduction—one of the largest and most targeted outreach and service efforts in history.

New York gradually whittled down its organized crime problem through years of creative prosecutorial and policing strategies. 
Medellín’s new mayor is also trying to attain all three goals by maintaining incentives for peace (through conditional repression) while simultaneously weakening the economic power of the gangs in a variety of ways.

5.2 Becoming problem-oriented, adaptive, and iterative

Some problems are relatively straightforward to solve. They have clear causes and measures for progress, the technology exists to solve it, it doesn’t require enormous local adaptation, they don’t require the coordination of many actors, or if they do, that logistical effort isn’t complicated by conflicting interests and information. If we stick with the medical analogy that started this article, a good example would be the delivery of flu or smallpox vaccines.

Other problems are “wicked” in nature. This is a technical term, coined in the 1960s by a management scholar to describe the really hard-to-solve social challenges. The causes are unclear, the metrics are hard to observe, the right tools don’t exist, and if they do they require customization and adaptation. Implementing them requires coordinating many actors, most of them with conflicting interests, where it’s hard to observe their effort or intentions. Wicked problems are not insoluble, however, and a longstanding literature on tackling wicked policy problems contains a few lessons that we can apply to organized crime.

Problem-driven and diagnostic

Instead of leaping to policy solutions—crackdowns, hotspots policing, focused deterrence, or cognitive behavioral programs—we have to understand the ailment before prescribing a treatment.

Typically, most cities have enough local experts and police intelligence to quickly assess the basic organizational structures and criminal situation. Going beyond the basics takes more investment, but organized crime is also a problem that will take decades to tackle, and so any investment in diagnostic and monitoring has a great deal of time to deliver returns.

Common diagnostic investments include:

Organizations tend to manage what is measured, and a first step is to establish metrics and measurement systems, especially to capture systematically unmeasured outcomes  (such as gang activities, extortion payments, and criminal and state legitimacy.

Foster a research community to do the descriptive research–qualitative and quantitative research—to establish the facts and regularities on the ground, and to build this intelligence gathering and processing capacity within and also outside of government.

• As circumstances often vary by neighborhood, build out the local capacity to understand and communicate with local criminal actors—not just within the state apparatus, but also through the “civilian architecture” of neighborhood-level outreach organizations that understand, build relationships with, and address the issues of the most violent groups and actors.

Establish incentives for the invention, proposal, and piloting of new policy solutions— within the bureaucracy, but also in the public sphere, by making information and data public. One example, building on the experience of U.S. cities, is Thomas Abt’s Community Violence Problem Analysis (CVPA).

In U.S. cities, violence tends to cluster around small networks of people and places. Shootings among these fragmented groups or atomistic individuals often occur in cascades of retaliatory violence. “Problem analyses” are designed to identify these high-risk networks through the systematic analysis of incidents, studying groups and social networks, and identifying the highest-risk places. The goal is to interrogate and challenge assumptions, while also focusing policy attention on the small number of highest-risk people, groups, places, and behaviors. Many of the same tools could be applied in situations of more hierarchical organized crime, especially the idea of intelligence gathering on groups, with the proviso that the most powerful groups may be committing the least violence.

Another example example is an ongoing collaboration in Medellín, Colombia between local and international universities, a research organization, the local government and police, as well as local media and experts. Independent, full-time ethnographers and journalists conduct continuous interviews with community members and criminal group members to understand criminal business lines, organization, hierarchies, peace arrangements, and causes of violence. We collaborate with the city government to conduct annual city-wide surveys of gang and state governance, fee payment, and legitimacy (published in the main newspaper). We also support the city government in building their information systems and analyzing data. We work with the national and local governments to integrate administrative data on neighborhoods, blocks, and individuals to track crime and other issues, and to build predictive models of crime and extortion for targeting. And we help the school system build mechanisms for monitoring dropout and recruitment risk and to target interventions to reduce risk.

This “urban lab” draws in many ways on the success of the University of Chicago Urban Labs—a nearly two-decade collaboration between the government, foundations, police, and non-profits designed to develop and share data, challenge assumptions, and inspire new approaches. The main question is why so few cities in the hemisphere have developed similar partnerships and capabilities.

Becoming iterative and experimental

The trouble with any first diagnosis is that it is probably mistaken. The first round of policies and solutions are probably wrong as well. And if they do work, criminal groups are smart and strategic and will adapt. 
Therefore, cities do not simply need the capacity to diagnose at intervals, they need to be able to engage in a process of trial and error, trying our multiple approaches and solutions at once, carefully trying to discern which are working or not, and using the patterns of observed success and failure to re-diagnose the problem.

The information systems discussed above–metrics, surveys, qualitative, and administrative data—are important here. Consider Medellín again. The annual surveys identified surprisingly high levels of gang governance and legitimacy in the city. Further investigation revealed that historical increases in police presence did not have the expected effects on criminal rule—often, the threat of police presence was the main factor driving gangs to provide order on the street, to protect drug profits. Increasing police presence pushed gangs to behave better, but did not counter gang rule. Retrospective analysis and a experimental pilot suggested that gang rule was a response to a strong and capable state, especially in sectors where drug rents are high.

This pointed the administration to focus on strategies that reduce rents while also maintaining gang incentives to avoid abusing and extorting residents.

Another example comes from studying gang entry among adolescents. While the city of Medellín had many promising youth programs, investigation revealed that the highest-risk youth were difficult to reach and engage, and often not deliberately targeted, meaning programs were not reaching the right recipients. Schools were also good at judging who had already joined a gang, but had few capabilities to identify the risk in advance. Large-scale surveys in schools identified the factors most predictive of gang entry, and these are now being integrated into routine administrative data collection. In partnership with non-profit organizations, Medellin’s schools are now testing and refining a series of counter-recruitment programs.

Another model for identifying and incentivizing experimentation is through competitive calls. The two major cognitive-behavioral programs in Liberia and Chicago that inspired so many community violence interventions around the world each emerged from competitive bids for new and innovative violence reduction programs—one from the UN Peacebuilding Fund and Liberian government, the other from the University of Chicago Crime Lab.

Fostering self-correcting mechanisms

One challenge with these recommendation is that few city administrations or police have incentives to diagnose and iterate. Policymakers who want to engage on careful trial and error constantly compete against political opponents who promise big, bold steps against organized crime. Worse still, crackdowns, kingpin strategies, and mano dura approaches are simple and clear, and often prove popular at the polls, at least temporarily.

As for schools, mayor’s offices, or police departments, trial and error asks these bureaucracies not to do what they are already equipped to do, but rather develop new capabilities that may or may not prove effective. It is one thing to ask authorities to be nimble and to adopt the best programs for their city’s circumstances, and another to ask them to innovate and iterate on their own. This points to the need for public–private partnerships between the state and universities, community organizations, and the media. Schools and police departments seldom have the analytical capabilities, the bandwidth, or the skills to engage in diagnostics or trial and error alone.

Ultimately, however, there needs to be accountability for results. This is a further advantage of the data produced through the diagnostics and monitoring described above. Bureaucracies have a tendency to mange what is measured. Most police commanders can tell you how crimes per sector have changed since last year, but not whether gangs have sold more drugs. Mayors and newspapers can quote week to week changes in homicide rates but not gang legitimacy. Better information systems, available to the public and newspapers, and transparent evaluation is one step towards accountability for improvement on these unmeasured margins.

Tackling organized crime is the ultimate wicked problem. Success will not come from simply importing solutions from abroad. Nor will it come from simply funneling more fiscal resources into policing or intelligence systems. It was be more difficult than that, requiring a change in mindset and approach.

No comments: