The Gap Is the Exploit: How Fragmented Communication Across Michigan Courts and Law Enforcement Costs Taxpayers and Enables Harm

Michigan has 83 counties. Each one operates its own court system, its own law enforcement records, and its own administrative universe. This structure is not an accident of geography. It is a deliberate feature of how Michigan distributes governmental authority, and for most purposes it functions well enough that no one with budget authority has been compelled to change it.

For a specific category of actor, it functions perfectly.

The person who has been sanctioned for filing abusive litigation in Kent County and drives forty minutes to file the same pattern in Ottawa County is not exploiting a loophole. They are exploiting a structural design choice that makes pattern recognition across jurisdictions essentially impossible without manual intervention by someone who already knows what to look for. The courts do not share a flag. Law enforcement does not share a flag. Nobody is looking at the cross-county picture, because the infrastructure to generate that picture does not exist.

What Fragmented Communication Actually Enables

There are three categories of actors who benefit from Michigan’s communication silos. The first is the vexatious litigant: someone who uses civil court filings as a harassment instrument, filing protective orders, restraining order petitions, or civil complaints not to win legal relief but to force a target to respond, appear, and pay attorneys. Michigan has no statewide mechanism to flag a person who has been sanctioned for abusive filings in one county when they walk into a different court and start again. The second filing looks identical to the first one. The court has no institutional memory of the pattern.

The second category is the fraud actor who operates across law enforcement jurisdictions. A person running a check-kiting scheme, a fraudulent contractor operation, or a benefits fraud pattern in multiple counties generates criminal history records in each county’s system in isolation. An investigator in Ingham County may have no way of knowing that the subject they are building a case against already has documented fraud activity in three adjacent counties, unless they manually query each county separately and know to look. That kind of manual cross-referencing requires time, requires someone to have a reason to do it, and requires knowing which counties to check. The pattern completes before anyone has assembled it.

The third category is the most expensive and the least discussed: the institutional revolving-door case. Someone cycling through the court system, through incarceration, through pretrial detention, through civil proceedings, across multiple jurisdictions, accumulating system costs at every stop, while no single institution maintains a full picture of the resource consumption pattern. Michigan’s approximately 60 percent pretrial jail population, the roughly 16,000 people in pretrial detention on any given day, is not a monolith. A meaningful fraction of that population has a cross-county pattern that would be legible in a unified system. The system is paying to process the same actors repeatedly, in isolation, without the data to know it is doing so.

The Vexatious Litigant Problem Is Not a Fringe Issue

Courts across the country have struggled with the vexatious litigant problem for years. A small number of serial filers consume a disproportionate share of court resources. Federal courts have developed mechanisms to address this, including pre-filing injunctions and vexatious litigant lists that require judicial approval before a flagged person can file a new action. Some states maintain statewide lists. Michigan does not have a robust statewide mechanism for this.

What Michigan has is a county-by-county approach, meaning a judge in one county can sanction a filer for abusive litigation and order them to seek leave of court before filing again. That order is enforceable in that court. It is not visible to the next court. The filer moves counties, and the pattern starts fresh.

The people who know about this gap are the people who have been targeted by serial filers. They learn, usually after significant financial and personal cost, that the court in their county has no idea about the other counties. They are told to document everything. They are told to bring the information themselves, because the system will not retrieve it. They become their own institutional memory because the institution has none.

Why Courts and Law Enforcement Are Not Talking to Each Other

The communication gap between courts and law enforcement is not simply a matter of different agencies failing to pick up the phone. It is a structural problem with several layers.

Courts and law enforcement operate under different record-keeping mandates, different confidentiality frameworks, and different technology systems. A law enforcement agency running LEIN, the Law Enforcement Information Network, is operating in a different data environment than the circuit court three blocks away. LEIN contains criminal history, but it does not contain civil court activity. A person who has been using civil courts for harassment, fraud facilitation, or financial manipulation may have no criminal record at all. The law enforcement picture and the civil court picture exist in parallel and are never synthesized.

Add to this the county-level siloing: a county sheriff’s department, a municipal police department, and a county circuit court may each have partial information about the same actor, housed in separate systems, with no mechanism for triangulation unless a specific investigator or judge manually pursues it. That kind of manual connection-building happens in high-profile cases. It does not happen systematically. Most patterns are not high-profile until after the harm is complete.

MiFile was intended to modernize Michigan’s court records infrastructure and improve access. Its implementation has been complicated, contested, and widely criticized by attorneys and court staff for a range of technical and operational failures. Whatever MiFile eventually becomes, it is a court records modernization project, not a cross-system integration project. It does not connect courts to law enforcement. It does not flag cross-county patterns. It is not the unified infrastructure Michigan needs.

What a Unified System Would Actually Cost to Build Versus What Fragmentation Already Costs to Maintain

The argument against a unified flagging and communication infrastructure is almost always the same: it costs money, it raises privacy concerns, and county autonomy is a legitimate governmental value. These objections are worth taking seriously, because none of them are wrong on their face. They are wrong in the cost-benefit framing.

Michigan spends $2.1 billion a year on corrections. It spends another $478 million on jail and pretrial operations. It collects approximately $418 million annually in criminal fines and fees from defendants. The system is running a massive, distributed, poorly tracked expenditure on individuals who cycle through it repeatedly, across jurisdictions, without generating a unified record that allows anyone to understand the full resource cost of any individual actor.

A unified cross-county flagging system would not require building new legal authority. It would require a policy decision to share existing records across existing systems, a data governance framework for that sharing, and a dedicated technical integration project. States have done this. The technology is not experimental. The infrastructure to run shared databases and flag cross-county patterns exists in commercial off-the-shelf products that governments already procure for other purposes.

The privacy objection is real but manageable. A shared flagging system is not a public surveillance database. It is a court and law enforcement internal tool. Access controls, audit logs, and use-limitation policies are standard features of government data systems. The privacy framework required to run a vexatious litigant registry or a cross-county fraud flag is less complex than the frameworks already governing LEIN. This is not novel territory.

The county autonomy objection is the most politically durable, because it has constituency support. County sheriffs, county clerks, and county commissioners have structural incentives to resist mandates from Lansing. But county autonomy is not a costless value when the cost of that autonomy is borne disproportionately by the people being targeted by actors who exploit the jurisdictional seams, and by taxpayers funding the system’s repeated first-encounters with the same bad actors.

From Shared Records to Predictive Analytics: The Path Is Shorter Than It Sounds

Predictive analytics in government contexts often lands as a term that implies expensive, speculative, and possibly dystopian technology. In the context of a unified court and law enforcement flagging system, it means something considerably more modest: the ability to look at a pattern of behavior across jurisdictions and time, and make a better-informed resource allocation decision.

The starting point is not a machine learning model. The starting point is a shared database that makes cross-county patterns visible to judges, prosecutors, and law enforcement investigators. Once that baseline exists, the analytical layer builds naturally from the data. Which case types are generating the most resource consumption? Which actors appear repeatedly across courts and agencies? Which court procedures are being exploited most frequently, and in which jurisdictions? These are not sophisticated questions. They are basic operational questions that any institution managing a $2.5 billion annual system should be able to answer. Michigan cannot answer them because the data is fragmented across 83 county silos.

The next layer is resource allocation. A court system that can see its own pattern data can direct judicial capacity, prosecutorial attention, and clerk resources toward the case types and actors generating disproportionate costs. A county prosecutor who knows that a person before them has fraud filings in four adjacent counties is making a materially different charging decision than a prosecutor operating blind. A judge who can see that the filer before them has been sanctioned for abusive litigation in three other Michigan courts is making a materially different ruling on a motion to dismiss than a judge seeing only the local file.

What comes after that, genuine predictive tools that model likely system costs, likely recidivism patterns, likely fraud vectors based on accumulated cross-county data, is a legitimate goal for a system that has committed to data integration. It is not where Michigan starts. It is where Michigan gets to, if it makes the foundational decision to treat fragmented data as the institutional liability it is.

The Counterargument: What Could Go Wrong with Centralization

A unified flagging infrastructure carries risks that critics raise with legitimate concern. Cross-system data sharing has a documented history of generating false positives, encoding biases present in local enforcement patterns, and creating records that follow people unjustly across jurisdictions. A shared vexatious litigant list, poorly designed, could flag people who filed meritorious claims that a hostile local judge dismissed abusively. A cross-county fraud flag system, without robust challenge mechanisms, could generate a record that prevents someone from accessing courts for legitimate purposes based on a pattern that was never properly adjudicated.

These are not hypothetical concerns. They are documented failure modes of centralized justice data systems, and Michigan’s existing systems have produced their own version of these failures in narrower contexts. LEIN errors have resulted in wrongful arrests. Court record errors have followed people into employment and housing decisions. Adding a cross-county flagging layer without robust error correction, challenge rights, and audit mechanisms would amplify those failures at scale.

The answer to this is not to forgo the infrastructure. The answer is to build the governance framework before building the technical system. That means: a defined and limited set of flag categories with specific evidentiary thresholds, a challenge mechanism that allows a flagged person to contest a record with a neutral reviewer, an audit log that tracks every access and every flag action, automatic expiration or review timelines on flags that are not refreshed by new activity, and legislative oversight with defined reporting requirements. None of this is unprecedented. Most of it exists in analog form in existing court procedures. The design task is translating those protections into the data governance framework.

The system Michigan has now is not neutral on these concerns. The current system also produces unjust outcomes: the vexatious litigant who keeps filing because no court has institutional memory of their prior conduct, the fraud actor whose pattern is never assembled because no agency has the cross-county picture, the incarcerated person whose documented wrongful conviction record in one county never reaches the agency making a parole decision in another. Fragmentation does not protect against institutional failure. It just distributes it across more jurisdictions where it is harder to see and harder to challenge.

What Reform Requires

Michigan has a $2.5 billion annual system that treats every actor as a first encounter. That is not fiscal responsibility. It is fiscal negligence dressed up as county autonomy. The gap between what the system knows and what the system could know is not a technical problem awaiting a solution. It is a policy choice awaiting a different decision.

The actors who depend on that gap staying open know exactly what they are doing. The question is whether Michigan’s institutions are going to keep pretending the gap does not exist.