No Address, No Identity

How verification systems treat stable housing as proof of personhood — and what that costs everyone else.

The system does not lose people experiencing homelessness. It identifies them as unverifiable and stops processing them.

That is a meaningful distinction. Loss implies accident. Rejection implies design. And the design of identity verification in the United States — embedded in federal benefits systems, state Medicaid programs, county social services, and private institutions alike — was never built to accommodate the people most likely to need it.

The logic is clean: if you have a permanent address, you are real, traceable, and accountable. A lease. A utility bill. A bank statement. A government-issued ID tied to that address. These are not just documents. They function as evidence of belonging within the administrative state. Without them, you do not disappear from the system. The system simply decides you do not qualify for it.

What follows is a structural analysis of how that logic operates, what it costs in human health and public safety terms, and why the people who built these systems were almost certainly not thinking about who gets left out.

The Address as a Stability Filter

Verification systems did not set out to exclude unhoused people. They set out to prevent fraud, manage eligibility, and produce audit-defensible records. An address accomplishes all three. It is standardized, documentable, and legally defensible. When a program auditor reviews a beneficiary file, a confirmed address signals that someone ran a check.

The problem is that a stability filter and an identity filter are not the same thing. One screens for fraud. The other screens for housing. When they are built into the same gate, the result is that permanent housing becomes a prerequisite for accessing the services designed to help people achieve it.

Nationally, an estimated 650,000 people experience homelessness on any given night, according to federal counts. Millions more cycle through housing instability across a year. For these individuals, documentation is not just difficult to obtain. It is difficult to maintain. Identification is lost or stolen. Documents are destroyed during encampment clearances. Constant movement degrades records. There is no secure storage. The system that requires continuity of documentation is dealing with a population defined by its absence.

Michigan reflects this pattern. The state’s unhoused population is concentrated in Wayne, Kent, and Washtenaw counties, with significant rural homelessness undercount in counties like Barry and Montcalm that lack permanent shelter infrastructure. The point-in-time count methodology used to track homelessness is itself constrained by the same assumption: that people can be found in a fixed location.

The Catch-22 Has a Name. Nobody Fixed It.

The documentation loop is not a secret. It has been named, studied, and reported on for decades. The shorthand is the “you need ID to get ID” problem. The more precise description is a self-referential eligibility barrier: most documents required to obtain identification themselves require identification or proof of address to obtain.

To get housing assistance in Michigan, an applicant typically needs a state ID. To get a Michigan state ID from the Secretary of State, an applicant needs a certified birth certificate and proof of Michigan residency. A certified birth certificate requires payment and, for many people, knowledge of which state to contact and a mailing address to receive it. Proof of residency requires a utility bill, lease, or bank statement — none of which exist if you do not have housing.

Shelters can sometimes serve as the residency anchor. Michigan allows a shelter address on a state ID application, but this depends entirely on the shelter having a formal agreement with the Secretary of State’s office, and on the individual having stable enough access to that shelter to use it as a mailing address. Both conditions are more variable than the policy documents suggest.

Far from accidental, the loop emerged from systems that were separately designed by separate agencies, each with its own eligibility logic, fraud prevention rationale, and compliance requirements. What they engineered was administrative tidiness; exclusion was the downstream effect.

Reentry: The State Creates the Problem and Then Punishes It

The documentation gap is just as problematic for people leaving Michigan correctional facilities. MDOC releases approximately 10,000 people per year. Upon release, individuals receive a parole address — typically a family member’s home, a transitional housing placement, or a parole-approved residence. What they do not automatically receive is a valid state ID.

Michigan law allows MDOC to assist with ID procurement before release, and some facilities have implemented programs to do so.

Implementation, however, is inconsistent.

A 2019 report from the Michigan Joint Task Force on Jail and Pretrial Incarceration identified documentation gaps as a barrier to successful reentry, but the systemic solution — ensuring every person leaves custody with valid identification — has not been standardized across the system.

The consequences are predictable and well-documented. Without a valid ID, returning citizens cannot open a bank account, which means they cannot receive direct deposit from an employer, which means they are limited to cash-pay work, which means they are excluded from a significant portion of the legitimate labor market. They cannot apply for most rental housing, because landlords require ID verification. They cannot access Medicaid, Food Assistance, or unemployment benefits without documentation that proves both identity and residency.

Michigan has approximately 32,000 people incarcerated at any given time, and spends roughly $2.1 billion annually on corrections. The state simultaneously structures a reentry process that withholds the one document most likely to reduce recidivism risk. This is not an oversight. It is a gap that has been studied, named, and left open.

The parole system then monitors compliance. Employment is typically a condition. Stable housing is typically a condition. These are reasonable conditions for public safety purposes. They are also conditions that depend on having identification. When a returning citizen cannot secure ID and therefore cannot secure employment and therefore violates a parole condition — that violation is treated as a supervision failure. The structural cause does not appear in the record.

The Court System Routes Notices to Addresses That No Longer Exist

Courts send notices by mail. That is not a metaphor for an outdated system. It is the operational reality of how criminal and civil proceedings notify parties of hearings, deadlines, plea offers, sentencing dates, protective order modifications, and civil judgments. The notice goes to the address on file. If the address on file is wrong, the notice is undelivered. The court has no mechanism to know it was not received. The clock keeps running.

For people without stable housing, the address on file is almost never current. An unhoused individual may have provided a shelter address when initially processed, a family member’s address as a courtesy, or an address that was accurate at the time of arrest but has since changed. None of those scenarios come with automatic notification to the court when the address becomes invalid. The individual is responsible for updating it. The practical barriers to doing so — accessing a courthouse, navigating a clerk’s office without ID, finding transportation — are significant and often insurmountable.

The criminal consequence is a bench warrant. Under Michigan Court Rules, a defendant who fails to appear at a scheduled hearing may be held in contempt and a bench warrant issued for their arrest. The warrant is not contingent on intent. It does not require a finding that the individual knew about the hearing and chose not to appear. The failure to appear is sufficient. A person who never received the notice because they moved between encampments, or because the shelter they listed closed, is legally indistinguishable from a person who ignored a summons.

Bond is often revoked at the same moment. A person who was released pretrial — who had not been convicted of anything — is now subject to arrest and detention because a piece of mail did not reach them. The warrant remains active and visible to law enforcement. A routine traffic stop, a wellness check, any contact with an officer running plates or names, surfaces the warrant and results in arrest. The original charge, whatever it was, is now compounded by a failure to appear. That compounds bail exposure, sentencing risk, and the complexity of the underlying case.

The overlap with reentry is direct. A returning citizen released from MDOC custody is assigned a parole officer and given reporting requirements. The parole address is the address of record for court purposes as well. If that address changes — because the family member can no longer accommodate them, because transitional housing ended, because a rental placement fell through — and the change is not formally updated with both parole and the court, notices continue routing to an address that is no longer valid. A supervision condition that requires court appearance becomes impossible to satisfy when the appearance date was never communicated.

The civil side compounds the problem in a different direction. Eviction proceedings in Michigan require service on the defendant. If the defendant has already lost stable housing, the address used for service may be the very property from which they are being evicted. A person who left because conditions were unsafe, or because they received an informal notice to vacate, may not appear at an eviction hearing because they did not receive the formal summons. The court enters a default judgment. That judgment appears on their credit and rental history. Every future housing application now carries the weight of a judgment they did not know existed, from a hearing they did not know was scheduled.

Child custody matters follow the same pattern. Protective order modifications, custody hearing dates, and parenting time enforcement proceedings all route to addresses on file. A parent experiencing housing instability who misses a custody hearing because notice went to a former address does not appear in the record as a person with an address problem. They appear as a parent who did not show up. Family court proceedings have their own contempt and default mechanisms. The documentation failure produces a legal record that reads as a conduct failure.

Debt collection judgments are the most invisible version of this problem. A creditor files suit. Notice routes to an old address. The defendant does not appear. A default judgment is entered. The judgment accrues interest. The creditor may then seek garnishment of wages or bank accounts. The person subject to that garnishment may not learn of the judgment until money disappears from a paycheck or account — if they have either. The court record shows a valid proceeding. It does not show that service was attempted at an address that had not been current for months.

What all of these scenarios share is the same structural feature: the court system is designed to treat address-based notice as legally sufficient service. Due process requirements in Michigan, as in most states, are satisfied when notice is sent to the last known address by first-class mail. The system is not required to confirm delivery. It is not required to verify that the address is current. It is required to send the notice. When it sends the notice, its obligation is met.

That framing makes legal sense within a system designed for people with stable addresses. It produces compounding injustice for a population that does not have them. The bench warrant, the default judgment, the revoked bond, the contempt finding — these are not neutral outcomes. They are legal records that follow people. They appear in background checks, housing applications, employment screenings, and future court proceedings. Each one makes the next one more likely.

The system is not recording this as an address problem. It is recording it as a compliance problem. The address is not in the record. The warrant is.

What Happens to Your Health Without an Address

Healthcare access follows the same pattern. Michigan’s Medicaid program, administered through the Michigan Department of Health and Human Services, requires applicants to provide proof of identity and Michigan residency. For unhoused individuals, the identity requirement is often the harder barrier to clear. Without a qualifying document, the Medicaid application cannot be processed.

Hospitals are required under federal law — specifically the Emergency Medical Treatment and Labor Act — to provide emergency stabilization regardless of identification status. EMTALA does not require ID. It does not, however, extend to non-emergency care, follow-up treatment, prescription coverage, or primary care. The unidentified patient gets stabilized and discharged. The underlying condition goes untreated. The prescription does not get filled. The specialist referral goes nowhere because the patient has no address for the appointment confirmation and no insurance card to present at check-in.

Pharmacies require ID for controlled substance prescriptions as a matter of standard practice. Without it, even a patient who has managed to see a physician and receive a prescription may be unable to fill it. This is not theoretical. It is a documented pattern in emergency medicine research: unhoused patients present in greater severity because primary care entry points have already turned them away.

The cost of that failure routes directly to the most expensive point in the healthcare system. Emergency department visits for conditions that should have been managed in primary care represent billions in annual healthcare spending nationally. Michigan’s uncompensated care burden — concentrated in safety-net hospitals in Detroit, Grand Rapids, and Lansing — reflects this pattern. The system saves the administrative cost of processing an unverified application and then pays for it at a multiple in the ED.

The Counterargument Is Real. It Does Not Hold.

The case for address-based verification is not frivolous. Benefit fraud is a real problem. Eligibility systems exist because public resources are limited and targeting them requires some mechanism for establishing who qualifies. An address is a reasonable proxy for stability, and stability correlates with legitimate need in ways that matter for program integrity.

The Government Accountability Office has documented the challenges federal agencies face in preventing improper payments, and address verification is one of the tools those agencies use. State Medicaid programs face federal matching fund requirements that include eligibility verification as a condition. The administrative incentive to use standardized, auditable documentation is real and legally backed.

The problem with this argument is not that it is wrong about fraud prevention. The problem is that it treats fraud prevention as the only relevant variable. A verification system that blocks 53% of eligible food assistance applicants because they lack a qualifying address is not a well-calibrated fraud prevention tool. It is a blunt instrument that incidentally prevents fraud by preventing access entirely.

Alternatives exist. Federal and state agencies have, in limited contexts, implemented caseworker attestation, shelter address acceptance, and fee waiver programs for ID replacement. The National Alliance to End Homelessness and Building Changes have documented functional identity models that work. They are not widespread because implementing them requires administrative will, updated eligibility rules, and inter-agency coordination. None of those things happen automatically. They happen when someone with decision-making authority decides the current system is unacceptable.

Administrative Invisibility Is Not a Side Effect

When identity cannot be verified, individuals become invisible within administrative systems in a specific, compounding way. They cannot be accurately counted in point-in-time surveys because those surveys require a location. They cannot be tracked across service systems because each system maintains its own records and those records cannot be reconciled without a common identifier. They cannot trigger automated eligibility pathways because those pathways require input data that does not exist.

The result is not just that individuals are denied services. It is that the systems responsible for serving them have no reliable record of their existence. A health system cannot track an unhoused patient’s care continuity if that patient has no ID. A benefits system cannot detect a service gap if the person experiencing it never successfully entered the system. The data that would justify expanding services or reforming eligibility rules is structurally unavailable because the people it would describe were never counted in the first place.

This is not a side effect of poor system design. It is a predictable outcome of systems that were designed to process verifiable data and were never updated to handle the populations that cannot provide it.

The unhoused individual, the person leaving prison, the domestic violence survivor who left without documents, the disaster evacuee whose records were destroyed — these populations share a common characteristic. They are operating outside the conditions that verification systems were built to process. The system does not adapt to them. It rejects their applications and generates no record of the rejection.

What Legislators Were Thinking About (It Was Not This)

The verification requirements that produce these outcomes were not written by people imagining the scenario of an unhoused individual trying to access food assistance. They were written by people thinking about fraud rings, duplicate enrollments, interstate eligibility conflicts, and federal audit exposure. Those are legitimate concerns. They are also the concerns of people operating inside a well-functioning bureaucracy, managing risks that are visible from inside that bureaucracy.

The risk that is not visible from inside the bureaucracy is the person who never successfully submits an application. That person generates no record. There is no denial letter, no appeal, no grievance, no data point. The system experiences their exclusion as silence — not as a signal that something is wrong.

Michigan’s legislature has had multiple opportunities to address documentation barriers in reentry and homelessness contexts. The Joint Task Force on Jail and Pretrial Incarceration produced a 2019 report with specific recommendations on ID procurement for people leaving custody. Those recommendations have not been implemented uniformly. The lame-duck session of 2024 ended without advancing the pretrial reform package that would have touched some of these structural issues. Reintroduction in the current session remains incomplete.

The pattern is consistent: the problem is identified, the fix is specified, the political moment to act passes, and the documentation loop continues generating the same outcomes for the same populations. The system is not unaware. It is not prioritizing.

The Practical Stakes

This is a systems design problem with measurable public costs. The health cost routes to emergency departments and Medicaid uncompensated care pools. The incarceration cost routes to corrections budgets — Michigan’s is $2.1 billion annually. The recidivism cost routes back to the criminal legal system, which then routes back to corrections. The administrative cost of processing appeals, managing shelter systems, and tracking chronic service users who cycle through without ever receiving stable access is distributed across agencies and fiscal years in ways that prevent clear accounting.

The population affected extends beyond people currently experiencing homelessness. Survivors of domestic violence who left without documents face the same loop. Disaster evacuees, particularly those displaced without advance notice, face it. Elderly individuals in cognitive decline whose families have not maintained their documentation face it. These are not marginal populations. They are people whose circumstances temporarily placed them outside the conditions that verification systems were built to process.

The question for Michigan is whether the state’s eligibility and verification infrastructure will be updated to handle that reality, or whether the current system will continue producing the same exclusions while generating no record of them.

An address proves where you live. It does not prove you exist. The systems that treat those two things as equivalent have decided, without stating it plainly, that existence within institutional frameworks is a privilege of stable housing.

That is a policy choice. It can be made differently.