Medicine can reduce crime

GLP-1 and Crime Reductions: Reducing crime through better health

There’s a lot of debate about whether Medicaid should pay for access to GLP-1s. Most of the debate takes the shape of an informal cost-benefit analysis comparing the gains from better health to the large costs of providing coverage. One big potential benefit of widespread access to GLP-1s, lower crime rates, is almost never included among the potential benefits. It should be, and it might tip the scales in favor of broader coverage if it were.

The mechanisms through which GLP-1 medications modulate reward-related brain circuits and improve health, including reduced substance use, lowered impulsivity, and increased stability, echo the effects of widely adopted psychiatric medications in the 1990s, a period when population-wide health interventions reduced crime. GLP-1 use is growing, but it remains too expensive for many, even in pill form. Given the positive externalities that appear to result from broadening availability, it seems cost-beneficial for Medicaid to cover their use. Let me make that case.

Medicine and Justice

Many medicines help reduce risk factors for criminality, but few are integral to our crime-reduction paradigm. You can find bits and pieces of the medicine-as-crime-policy idea here and there. But, as far as I know, it does not exist anywhere as a foundational principle of an anti-crime strategy. This is despite strong evidence from the 1990s that widespread access to behavioral health medicine helped to spur the Great American Crime Decline, by removing an enormous risk condition in millions of Americans and carving a path toward opportunity. This was a lesson we barely acknowledged at the time, then promptly forgot.

In the interim between then and now, many other medications emerged with crime-fighting benefits. Vivitrol, for instance, blocks receptors in the brain that cause euphoria when you drink or use opioids. This medication is extended-release Naltrexone, so someone receiving Vivitrol at the end of a course of drug or alcohol treatment will have a 30-day window to begin their recovery, knowing that even if they use, they won’t get high. The clinical results are very encouraging.

While there has been a long, slow embrace of medically assisted treatment (MAT), the tied fight between abstinence-only and MATs continues. The field seems scarred by problems with methadone and suboxone, which created their own addiction problems, and has been struggling in the adoption of new medicines. But they work, and I suspect they will eventually be a routine part of substance treatment protocols.

Now, the next class of medications, GLP-1s, has emerged and appears to offer substantial benefits with few side effects. Two of those benefits are closely related to criminality. One is that GLP-1s appear to reduce users’ appetites for substances, including alcohol and opioids. Reducing the consumption of both would have enormous crime-reducing consequences. The other is that GLP-1s broadly improve users’ health and, in turn, their life-course outcomes. We don’t know much about the relationship between people’s general health and their likelihood of committing crimes, but we do know that people who commit crimes are much less healthy than the average person.

Medicines and the Great American Crime Decline

There are hundreds of ways to reduce crime without involving the criminal justice system. Many of these explanations are front and center in debates about why crime declined substantially in the 1990s all across the US, and essentially, all at once. Some of these explanations—like the removal of lead from drinking water, paint, and gasoline (but particularly gasoline!)—have been widely accepted. One that has seen little attention is the widespread adoption of SSRIs and antipsychotic drugs over that decade. The correlational evidence that SSRIs and antipsychotics is strong, but the empirical evidence is weaker. Some studies show reductions in aggressive behavior and lower violent crime rates, but the results are not as strong as the correlational data suggest (though data limitations are a big impediment to his research). I note that Franklin Zimring identifies this relationship between SSRI availability and violence reductions as one of the few explanations for the crime decline that holds up in international comparisons.

Source: Marcotte, Dave E., and Sara Markowitz. “A cure for crime? Psycho‐pharmaceuticals and crime trends.” Journal of Policy Analysis and Management 30, no. 1 (2011): 29-56.

In the 1990s, psychiatry saw a major shift with the widespread use of selective serotonin reuptake inhibitors (SSRIs) for depression and anxiety, and second-generation antipsychotics for psychotic disorders. SSRIs work by increasing the availability of serotonin in the brain, which helps regulate mood, anxiety, and impulse control. Compared to older antidepressants, SSRIs are safer and have fewer side effects, and as a result, were prescribed broadly for the first time. At the same time, new types of antipsychotics became available. These reduced severe psychiatric symptoms like hallucinations and delusions, also with fewer side effects than earlier antipsychotics. Together, these medications expanded treatment access and improved day-to-day functioning for many people with serious mental illness. By 1998, almost 12% of all drug spending was for these new SSRIs.

SSRIs can reduce crime mainly by affecting emotional regulation and impulse control. By reducing severe depression, chronic anxiety, irritability, and impulsive aggression, SSRIs can reduce violence in several ways. SSRIs can directly reduce aggression, reduce provocative behaviors, and prevent reactive offending. They may also lower the likelihood of behaviors linked to violence, particularly substance misuse. Finally, SSRIs support people through classic protective mechanisms, including helping them remain engaged in work, school, and relationships.

Atypical antipsychotics may reduce crime risk in a different but complementary way. For individuals with schizophrenia or bipolar disorder, these medications reduce paranoia, hallucinations, and disorganized thinking that can increase the risk of confrontations with others or police. Improved symptom control can mean fewer psychiatric crises, fewer involuntary hospitalizations, and greater stability in housing and supervision, all of which are associated with lower justice system contact. Atypical antipsychotics prescription also grew rapidly during the 90s.

By the late 1990s and early 2000s, SSRIs were being used by tens of millions of people worldwide, becoming some of the most commonly prescribed medications in primary care. Antipsychotics were prescribed to far fewer people—roughly 1–2% of the population—but their use rose steadily after 1993 as newer drugs replaced older ones and expanded into bipolar disorder and other conditions. These prescribing trends matter because even modest behavioral or stability effects can have large population-level consequences when applied at scale. A simple way to see this shift is to look at prescription trends over time, which show a sharp rise in SSRI use beginning in the early 1990s and a later but steady increase in atypical antipsychotics through the late 1990s and 2000s.

GLP-1’s and Crime Reduction

GLP-1 receptor agonists may reduce crime indirectly by lowering alcohol and opioid use and dampening reward-driven impulsivity—mechanisms linked to violent incidents and disorder—though direct evidence on crime outcomes remains limited.

The strongest evidence for GLP-1s and crime risks is around reductions in alcohol use following regular semaglutide use. A randomized clinical trial in adults with alcohol use disorder found that semaglutide reduced alcohol outcomes versus placebo, especially on cravings but also consumption. Less alcohol use then leads to fewer alcohol-involved assaults, DV incidents, disorder, and DUI. Recent reviews/meta-analyses also summarize an overall pattern: GLP-1 RAs are associated with fewer alcohol-related events across studies, though results vary by outcome and design.

A second potential crime-reducing mechanism derives from reduced opioid use, leading to fewer overdoses and fewer drug-market/withdrawal-driven crimes. A large observational study found that semaglutide was associated with lower opioid overdose risk among patients with comorbid type 2 diabetes and opioid use disorder. There are also ongoing clinical trials testing semaglutide for opioid use disorder outcomes.

Additionally, GLP-1 signaling appears to interact with reward circuitry, including dopamine pathways, to reduce craving, cue reactivity, and compulsive drug-seeking. Because substance use and withdrawal are tightly linked to impulsive behavior, violence, and economically motivated crime, even modest reductions in craving and relapse risk can lower exposure to the situations and pressures that often precede criminal activity.

Finally, it appears that GLP-1s help people stabilize in ways associated with better quality of life and better long-term outcomes, including reduced criminality. If GLP-1s reduce heavy drinking, relapse, and overdose risk, that can translate into fewer acute crises, fewer destabilizing events (job loss, eviction, and relationship problems), and less exposure to criminogenic settings.

In Closing

Crime policy has long rested on the assumption that public safety is produced primarily by police, courts, and prisons. In the 1990s, population-wide access to behavioral health medications quietly reduced key risk factors for crime, even as public debate focused elsewhere. That experience offers a useful lens for thinking about GLP-1s today, not as a law-enforcement tool, but as a health intervention that operates upstream of violence and disorder.

None of this requires claiming that GLP-1s are a proven crime-control strategy. It requires only recognizing that medications capable of reducing heavy substance use, impulsivity, and instability act on the same conditions that often precede criminal behavior. When such interventions reach millions of people, even modest individual-level effects can accumulate into meaningful population-level change.