The Justice Department has chosen four jurisdictions to serve as demonstrations of a new federal public-safety strategy. The department says its Model Cities Initiative will direct nearly $300 million toward Jacksonville, Florida; East Baton Rouge Parish, Louisiana; Jackson, Mississippi; and St. Louis, Missouri.
The promise is substantial: stronger law enforcement, better forensic and prosecutorial capacity, new technology, mental-health and substance-use services, and efforts to reduce repeat offending. The department describes the initiative as a comprehensive model that other communities may eventually follow. The Justice Department’s October 30 announcement presents the selections as part of a broader effort to reduce violent crime and restore public order.
That is the political argument for the program. The administrative question is more basic: What, precisely, has been awarded, and what evidence will show that it worked?
The public announcement identifies the four jurisdictions and the agencies involved. It does not identify each jurisdiction’s allocation, the final approved budgets, the amount formally obligated, the timetable for disbursement or the specific performance targets each recipient must meet.
That distinction is not a technicality. The initiative’s own solicitation says that DOJ will work with the selected jurisdictions to approve detailed implementation plans and final budgets, with funds released as milestones are completed. The awards are structured as cooperative agreements, not ordinary grants, and the expected award period is 36 months. The FY 2026 Model Cities solicitation also requires applicants to describe timelines, responsible entities and the way DOJ and the public will measure success.
In other words, Friday’s announcement begins the accountability process. It does not complete it.
A broad program is not necessarily an easy one to evaluate
The initiative is broad by design. Eligible activities include hiring sworn officers and civilian personnel, expanding forensic and prosecutorial operations, purchasing body cameras, license-plate readers, drones and artificial-intelligence systems, building real-time crime centers, supporting victim services, improving reentry programs and funding mental-health and substance-use treatment.
That range reflects an increasingly common understanding that public safety is not produced by police staffing alone. A prosecutor’s office, a forensic laboratory, a behavioral-health provider and a community-supervision agency may all affect what happens after a violent incident.
But breadth creates an evaluation problem. If one jurisdiction spends heavily on officers, another on forensic technology and a third on reentry services, a national claim about the “Model Cities” approach may conceal several different programs operating under one label.
The National League of Cities warned before the selections that the cooperative-agreement structure would give DOJ a more active role in planning, implementation and oversight. It also raised the practical issue of sustainability: personnel, specialized units, technology platforms and service programs may continue to require money after federal support ends.
That is the point at which many public initiatives become less impressive than their launch announcements. A jurisdiction can purchase a system in year one and struggle to maintain it in year four. It can hire federally supported staff without having the local revenue to retain them. It can adopt a data platform without having enough analysts, privacy controls or technical support to make the system useful.
Falling crime does not identify the cause
The department has also placed the initiative inside a favorable national crime narrative. DOJ says violent crime is falling and cites earlier federal involvement in Memphis and Washington, D.C., as evidence that concentrated intervention can produce large reductions.
The claim may prove correct in some places. It still requires careful testing.
Crime statistics are not receipts that identify which policy purchased a result. A city’s homicide rate may fall because of policing, prosecution, demographic change, reduced firearm availability, changes in drug markets, improved emergency medicine or ordinary statistical variation. Several forces can operate at once. A decline during a federally supported period does not, by itself, establish that federal money caused the decline.
The national data also require some care. The Bureau of Justice Statistics’ preliminary 2025 National Crime Victimization Survey found that the rate of violent victimization excluding simple assault fell 25 percent from 2024 to 2025. BJS cautions that these are preliminary estimates from a redesigned survey and are not comparable to older NCVS measures. The survey includes crimes reported and not reported to police, which makes it different from police-recorded crime data. BJS’s report explains those limitations.
That difference is precisely why the four jurisdictions need more than a single favorable number. They need baselines that distinguish homicide, robbery, aggravated assault, vehicle theft, victimization, emergency calls, arrests, prosecutions, convictions, recidivism and public perceptions of safety. They need to explain how changes in reporting practices will be handled. They need comparison communities or other credible methods for estimating what might have happened without the intervention.
Otherwise, almost any improvement can be claimed as success, and almost any setback can be attributed to circumstances outside the program’s control.
The federal role makes the accounting more important
The cooperative-agreement structure may be useful. DOJ can provide technical assistance, require better planning and withhold funds when milestones are not met. Federal involvement can also prevent local agencies from spending heavily on equipment or programs that they cannot operate responsibly.
But greater federal involvement creates greater federal responsibility. When Washington helps design a local public-safety strategy, reviews its budget and releases money according to milestones, it is no longer merely writing a check. It is participating in the design of the intervention.
That raises a question the department should answer before the first progress report: Who is responsible if the strategy fails?
The answer cannot be buried in a cooperative agreement that the public never sees. The final documents should identify the program’s objectives, the data to be published, the agencies responsible for each milestone, the safeguards governing surveillance and information-sharing technology, and the consequences of missing targets.
The public should also be able to distinguish money announced from money obligated, money obligated from money disbursed, and money disbursed from services actually delivered. Those are different events. Government accounting becomes misleading when they are treated as one.
The first test begins before the crime statistics arrive
Residents of Jacksonville, East Baton Rouge Parish, Jackson and St. Louis may see changes in police staffing, technology, prosecution, behavioral-health services and reentry programs. People subject to enforcement may also encounter new surveillance systems, data-sharing arrangements and forensic procedures. Those systems require rules for access, error correction, retention and public oversight.
Local elected officials should disclose which costs will remain after the 36-month federal period. They should identify whether city councils, county governing bodies or other public institutions must approve final spending plans. They should explain what existing programs will be expanded, what will be replaced and what will be created from scratch.
DOJ should publish the final award notices, cooperative agreements, action plans, budgets, milestone schedules and evaluation templates. It should disclose whether independent evaluators will be involved and what happens when a jurisdiction misses a target. The four jurisdictions should publish the applications that won them selection, subject to legitimate privacy and law-enforcement restrictions.
A national model is not established by naming four cities. It is established by showing what was done, what it cost, what changed and what would probably have happened without it.
For now, the Model Cities Initiative is a serious public-safety experiment with an incomplete public record. The appropriate response is neither automatic suspicion nor automatic applause. It is an insistence that the accounting begin before the political narrative hardens.
