Historic Declines, Contested Borders

There are six violent deaths in William Shakespeare’s “Romeo and Juliet,” a tale of love and tragedy set in 14th century Verona — then, as now, a prosperous city in northern Italy. With 28 named characters, the play’s death rate is a sobering 21%, a reflection of the brutal reality of the place and time in which the story unfolds. Renaissance Italy is often remembered for its flourishing trade, rising prosperity and the creation of some of the most celebrated works in the Western canon. But it was also, by modern standards, an extraordinarily violent society.

As Manuel Eisner documents in his seminal work on homicide across European history, the murder rate in 14th century Florence — just 145 miles south of Verona — may have reached 150 per 100,000 residents. That level of violence is unmatched today in any American city and would be seen only in a handful of the most distressed communities anywhere in the United States. By contrast, the U.S. homicide rate in 2024 was roughly five per 100,000. In New York City — the nation’s flagship metropolis — it was 4.7, fully 97% lower than Florence’s medieval peak.

While the comparison is selective, it is perhaps notable that a 97% decline in homicides compares favorably to the effectiveness of another remarkable lifesaving intervention: vaccination. Indeed, two doses of the measles, mumps and rubella (MMR) vaccine are estimated to be about 97% effective at preventing infection. A century ago, measles killed roughly 8,000 Americans annually. In 2024, it killed none. Through vaccination, disease had effectively been eliminated as a cause of death in the United States — until the antivaccination movement, encouraged by new leadership in Washington, gave the measles virus a new lease on life last year, when it caused three casualties. Of course, we cannot quite make the same assertion about homicide, which still claims roughly 17,000 American lives each year. And yet viewed through a long historical lens, the conclusion is unavoidable: To a remarkable degree, modern societies have inoculated themselves against lethal violence. Were the United States to experience the homicide rate of 14th century Florence, more than half a million Americans would be killed by other people each year.

The long-run decline in homicide is nothing short of astonishing. What makes 2025 exceptional, however, is that the short-run trends are astonishing as well. According to a recent Council on Criminal Justice report, large U.S. cities saw homicides fall by roughly 25% in 2025, following an already remarkable 15% decline in 2024. Homicide rates are now about 25% lower than in 2019, before the sharp pandemic-era spike. In Philadelphia and Baltimore, homicide is as low as it was in the mid-1960s. In San Francisco, one must look back to 1942 — when many young men were overseas fighting World War II — to find a comparably low rate.

When measles deaths plummeted in the 1920s, the explanation was clear: Vaccination worked. The causes of the post-COVID-19 homicide decline, by contrast, remain elusive. There is no shortage of theories, including a generalized return to normalcy after the pandemic, but many familiar explanations do not fit the facts. Poverty has barely changed. The unemployment rate is essentially the same as it was in 2022. Police staffing is flat or declining. The prison population has stabilized after a decade-long fall from its 2008 peak. Nor can this decline plausibly be attributed to the long-lag effects of legalized abortion or lead abatement, factors sometimes credited with contributing to the crime drop of the 1990s.

Understanding what went right in 2024 and 2025 is critical. Without a clear account of the forces behind this decline, policymakers risk squandering a rare opportunity to learn how to sustain and deepen gains in public safety. The contributors to this volume take up that challenge. They identify the most consequential public safety developments of the year, examine how New York City compares with other major U.S. cities, analyze what police have focused on in 2025, consider what advocates and officials should prioritize in 2026 and — critically — grapple with the central question of why 2025 proved to be such an extraordinary year for public safety.

Contributors also engage the other defining public safety story of 2025: a dramatic and highly visible escalation of interior immigration enforcement.

What happened to crime in 2025?

Alex Piquero and Adam Gelb set the stage by identifying what they see as the five most important crime and justice developments of the year. They describe mounting attacks on federal crime data and a growing distrust of official statistics as well as deep and abrupt cuts to federal justice grants, disrupting prevention efforts and meaningful progress in rebooting courts, policing and corrections after pandemic shutdowns. They also highlight a sharp disconnect between public fear of crime and mounting evidence that crime is falling. Their essay underscores the historic nationwide drop in homicide — reaching lows in some cities not seen since the 1960s — while emphasizing how much remains unknown about its causes and how essential rigorous, independent research will be for future policy.

New York City led the nation during the great crime decline of the 1990s and achieved an extraordinary 50% reduction in homicide between 2010 and 2019. While the city had a banner year in 2025 with respect to homicides and shootings — which were down by 25% and 20% respectively — Vital City’s End of Year Crime Report suggests that recent public safety trends in the city are actually something of a mixed bag. Indeed, with respect to murders, the city’s 20% decline mirrors national trends and is broadly similar to peer cities like Los Angeles, Chicago and Philadelphia. What’s more, assaults, robberies, burglaries and especially motor vehicle thefts remain notably higher than they were in 2019. Crimes on the subway also remain elevated, which, given the universality of subway travel among city residents, forms the basis for widespread discontent even as murders and shootings are in free fall. This is despite an increase in street stops and the rate at which crimes are cleared by arrest. Together, the data suggest that although the police have successfully engaged the power few who drive the problem of lethal violence, they have not solved the problem of illegal behaviors that have a broader base of participation.

Building upon the end-of-year crime report, Marcos Soler emphasizes that while the city has since reached historic lows in gun violence and homicide — largely mirroring state and national trends — it has failed to reduce overall crime. Total reported crime remains roughly 45% higher than its 2017-2019 low and declines now lag behind those in other large U.S. cities. Soler attributes this divergence to policy and management failures under the Adams administration, pointing to sharp increases in assaults and larcenies that now dominate citywide crime. He contends that simply increasing arrests has been ineffective because diminished certainty of prosecution has eroded deterrence. To restore New York to its prepandemic crime levels, Soler calls for a renewed focused deterrence strategy, place-based interventions in high-crime areas such as the Bronx, people-centered strategies for high-recidivism adults aged 35-44 and stronger supervision and accountability for individuals on pretrial release.

In a provocative essay that gets right to the crux of the matter, John Roman asks why 2025 was, in fact, such a banner year. He argues that the nationwide crime decline was not driven by changes in policing or incarceration — the traditional levers of crime control — but by large-scale federal investments in prevention. Pandemic recovery funds, Roman suggests, stabilized and expanded local social services by funding teachers, counselors, clinicians and prevention workers, thereby reducing risk and creating opportunity for the young people most vulnerable to violence. What makes this argument especially compelling is its scope: Because the decline was broad and consistent across cities and evident across crime types, Roman contends that the cause must be a national one. Alternative explanations — changes in police staffing, gun availability, drug markets, or long-term demographic trends — fall short. Based on his hypothesis, Roman warns that enforcement-only strategies have diminishing returns and argues that sustaining crime reductions will require a philosophical shift toward prevention as a central public safety strategy.

What should the police be doing?

In a Vital City interview, NYPD Chief of Department Michael LiPetri offers a rare window into the Department’s operational strategy, explaining how New York drove shootings to a historic low in 2025 — fewer than 700 citywide — through highly targeted, data-driven policing. The approach emphasizes hyperlocal “violence reduction zones,” intensive foot patrols during peak hours, coordinated narcotics and gang enforcement and sophisticated analytics to identify high-risk individuals, groups and locations. Close collaboration with prosecutors and sustained gun arrests reinforced deterrence and produced broad declines in shootings and homicides, including notable gains in the Bronx.

At the same time, LiPetri acknowledges the limits of these successes. As Soler observes, assaults — particularly domestic violence, assaults on public servants and unprovoked attacks often linked to mental illness — remain stubbornly high. The Department is responding by restructuring domestic violence investigations to be more victim centered, increasing warrant enforcement, focusing on transit and school zones, and reassessing strategies for persistent problem populations and places. The interview captures a central tension of the moment: There is extraordinary progress in reducing lethal violence alongside more limited progress in addressing the everyday challenges in everyday safety.

To this end, what should the Police Department be thinking about in 2026? This is a key question at any time but is an especially important question during a period in which there is a new mayoral administration that has a philosophy on crime control that is notably different from that of the prior administration. Where can the new administration and the NYPD, staffed by senior personnel who have served previous administrations, find consensus? And which issues are most critical when it comes to maintaining public safety?

Anthony Braga and Christopher Winship argue that gang databases — including the NYPD’s Criminal Group Database, which Mayor Mamdani staunchly opposed as a candidate — remain essential tools for reducing gun violence and should be reformed rather than abolished. Critics contend that such databases are discriminatory and overly broad. Braga and Winship counter that gun violence in New York is highly concentrated among a small number of individuals involved in gangs and crews and disproportionately harms Black and Hispanic communities — the same populations overrepresented in the data. Research shows that focused, data-driven policing targeting high-risk groups can substantially reduce gun violence without displacing crime. Eliminating these databases, they warn, could lead to more indiscriminate policing by forcing officers to rely on subjective judgment, which is more prone to bias. Reform — clear criteria, transparency, oversight and due process — offers a better path forward.

Alex Chohlas-Wood argues that artificial intelligence provides an extraordinarily powerful opportunity for the NYPD to deploy its limited resources more effectively. Drawing on his experience developing Patternizr — an early AI tool used by NYPD detectives to identify patterns based on crimes’ modus operandi, Chohlas-Wood argues that the Department has fallen behind as AI has advanced rapidly elsewhere. Carefully tested systems, he suggests, could improve officer training through analysis of reams of body-worn camera footage, automate freedom of information requests and reduce administrative inefficiencies. Chohlas-Wood urges proactive but cautious adoption, emphasizing that practice must be grounded in transparency, rigorous evaluation and collaboration with outside experts to enhance legitimacy while maximizing learning.

Should public safety be reconceived?

Gloria Gong closes the public safety section by questioning whether conventional crime metrics ultimately fail to capture what matters most — how people actually experience safety in public spaces. While CompStat revolutionized crime tracking, Gong argues that it fails to measure what residents care about most: whether they feel safe. She proposes a new metric — a citywide “Felt Safety Index” — built from real-time, anonymized behavioral data such as foot traffic, park usage and patterns of public presence. The goal is a next-generation CompStat tool that treats safety not merely as the absence of crime, but as the presence of confident, vibrant public life. A decade ago, such an idea might have seemed utopian. Today, given the explosion and democratization of data, it feels firmly within reach.

Immigration enforcement and public safety

The other defining public safety story of 2025 is a dramatic escalation in the scale and visibility of interior immigration enforcement by federal authorities — culminating in the killing of two U.S. citizens in Minneapolis by Immigration and Customs Enforcement and Border Patrol agents.

There is much that can be written about this topic, but, for starters, it is fair to wonder whether undocumented immigrants do, in fact, make the United States less safe. While the scholarly literature is clear that immigrants, as a whole, do not increase violent crime rates and are less likely than U.S. natives to end up in a state prison pursuant to committing a serious crime, studying the effect of undocumented immigrants specifically is a little harder to do due to the difficulty of obtaining accurate data. Still, several conclusions are well supported. Most research finds that undocumented immigrants are not an important driver of violent crime and are not disproportionately represented in state prisons. Moreover, ICE’s own data show that, in 2025, only about 13% of detained individuals had prior convictions for violent or property crimes — suggesting that enforcement targeted a population at lower risk of serious offending than those typically arrested by local police. This finding aligns with earlier research on the Secure Communities program, which expanded federal-local cooperation on immigration enforcement during the Bush and Obama administrations. While investments in local policing reliably improve public safety, Secure Communities — and the increase in removals it produced — did not generate measurable crime reductions.

Beyond empirical debates lies a deeper conflict over the appropriate role of federal immigration enforcement in local communities: how such enforcement should be carried out, how constitutional norms should be respected and how federal law enforcement officers should be trained, managed and held accountable. These questions intersect with enduring disputes over federalism, the balance of power between states and the federal government and the meaning of sanctuary policies.

Erwin Chemerinsky addresses the constitutional limits of federal authority, arguing that President Donald Trump’s threats to withhold federal funds from so-called sanctuary cities are unlawful. Supreme Court precedent, he explains, bars the federal government from coercing states and localities into enforcing federal law, and only Congress and not the president may attach conditions to federal spending. Courts have previously struck down similar efforts. Chemerinsky also clarifies that sanctuary cities do not shield undocumented immigrants from enforcement — they simply decline to assist federal authorities, often for practical reasons such as encouraging crime reporting and protecting public health.

With respect to the tactics employed, Brandon del Pozo sharply criticizes recent ICE actions in Minneapolis — particularly the killing of Alex Pretti — as emblematic of reckless and illegitimate force that corrodes public trust. He contrasts ICE’s conduct with the norms of professional municipal policing, emphasizing restraint, legitimacy and accountability. Del Pozo argues that low training standards, quota-driven enforcement, erratic use of force and resistance to oversight have damaged ICE’s credibility and weakened community safety. He suggests that this moment offers municipal police departments an opportunity to model legitimate policing by deepening reforms, engaging communities transparently and pairing enforcement with preventive strategies.

The volume concludes with an interview with Jeh Johnson, former U.S. Secretary of Homeland Security, who reflects on the Trump administration’s aggressive interior enforcement strategy. Johnson criticizes the deployment of large, militarized federal immigration forces into local communities, warning that it fuels fear, undermines safety and erodes trust between residents and law enforcement. Echoing del Pozo, he notes that a fixation on arrest quotas departs from traditional enforcement priorities and has intensified resistance in sanctuary cities such as New York. Johnson calls for better training, stronger emphasis on de-escalation and a more collaborative relationship between federal and local authorities, suggesting that ICE itself may require a fundamental rethinking of its mission and public role.

When it comes to serious violence, 2025 was a banner year for public safety in cities across the United States. The homicide rate in large U.S. cities is as low as it was during the early 1960s and, for a majority of Americans, the country has never been safer in their lifetimes. Yet it’s not clear that the public has noticed. Do people actually feel safer? Do Americans even believe the crime data in a world in which truth sometimes takes a back seat to ideology?

In such a world, the challenge for politicians and policymakers is not only to achieve greater safety but to ensure that Americans feel safer. Given that crime is concentrated to a considerable degree in a small number of places, this is a challenge. Most Americans live in communities in which crime is uncommon. Their beliefs about crime then center on what they believe is happening in places in which they probably spend little time. A recent New York Times article captured this phenomenon well, observing that tourists who flocked to San Francisco for the recent Super Bowl were surprised to find that the city was pleasant and not postapocalyptic as they’d heard about in social media postings. Indeed, San Francisco experienced only 30 murders during the past 12 months, a rate of homicide among the lowest in the United States and only slightly higher than European capitals such as Paris and Brussels (and, in fact, lower than Berlin). But will anyone believe it?