Documenting chronic homelessness

Chronic homelessness is the status that moves someone up a prioritization list, and it is the one HUD asks the hardest questions about afterward. The definition itself is not complicated. Proving it is, and the proving is where organizations lose findings.

Most guidance on this covers the definition and stops. The part that decides whether you pass a monitoring visit is further down: an evidence hierarchy, and three percentage caps that apply across your whole caseload rather than to any one file.

The definition, briefly

There are two routes to the same status. Both require a disability, and both are assessed on the head of household.

Route one Twelve consecutive months Homeless continuously for twelve months, plus a documented disability. The simpler route to describe and usually the harder one to evidence, because it requires an unbroken chain.

Route two Four occasions over three years At least four separate occasions within three years, totalling at least twelve months, plus a documented disability. Institutional stays of under ninety days may fall inside the period.

Route two carries a requirement people miss: four occasions means a minimum of three breaks, and each break has to be documented. The occasions are defined by the gaps between them, so the gaps are what you are evidencing.

The evidence hierarchy

HUD ranks the kinds of proof. You are expected to try for the highest available and to show you tried before dropping down.

FIRST Third-party documentation

HMIS or comparable database records. Individual records of shelter or safe haven stays. Street outreach contact logs. Written observation by an outreach or intake worker describing the living situation. Institutional discharge paperwork or a referral carrying dates. Written observation from a community member who physically saw the situation, paired with a referral from a housing or service provider.

SECOND Intake worker observation

Your own staff, recording what they saw. This is more powerful than most organizations treat it, and it is free. It is also the thing that quietly disappears when a note gets written three weeks later from memory.

LAST Self-certification

Written and signed by the person, not notarized. It has to be accompanied by intake worker documentation of the living situation and by a record of your attempts to obtain better evidence. And it is capped, which is the next section.

The three caps, which is the part nobody tracks

These are the numbers that make this a portfolio problem rather than a file-by-file one. They apply across the households you serve in an operating year, which means no individual case tells you whether you are compliant.

Measure these across your operating year, not per client
100%	of households may self-certify up to three of the twelve months Every household you serve can have three months of the required twelve covered by self-certification. This one is permissive and it is the allowance most organizations underuse.
75%	must have third-party documentation for nine of twelve months At least three quarters of households served in an operating year need real third-party evidence covering nine of the twelve months. This is the one that gets checked.
25%	is the ceiling on full-period self-certification No more than a quarter of households served may rely on self-certification for the entire period, and that allowance exists for people who were unsheltered and out of contact with any provider for a long stretch. It is for the hardest cases, not for the ones where nobody chased the paperwork.

One more, and it runs the other way. For the four-occasions route, 100 percent of the breaks between occasions may be documented by self-report. The gaps do not need third-party evidence. That is a genuine and underused piece of relief, because reconstructing where somebody was during a two-week break three years ago is close to impossible.

Why this shape matters

You can be fully compliant on every individual file and still fail the portfolio test, because the 75 percent and 25 percent thresholds are about the mix. An organization that serves mostly long-term unsheltered people can drift over the 25 percent ceiling without a single caseworker doing anything wrong. Nobody notices until someone counts.

Documenting the disability

The condition has to be expected to be long-continuing or of indefinite duration, and to substantially impede the person's ability to live independently. Acceptable evidence includes verification from a state-licensed professional, Social Security Administration documentation, or receipt of disability benefits such as SSDI or veteran compensation.

There is also a path that buys you time. Intake staff may record an observation of disability, provided supporting evidence follows no later than forty-five days from the application for assistance. That is a real window and it means an undocumented disability is not a reason to delay someone's application. It is also a forty-five day clock that somebody has to be watching, and it is exactly the sort of clock that gets lost between two staff members.

What to do about it

Count your own mix, once a quarterWhat share of your chronically homeless households are relying on self-certification for the full period? If you do not know, that is the finding waiting to happen. It is one number and it takes an afternoon.

Write the observation the day you make itWorker observation is second in the hierarchy and it costs nothing, but only if it exists. An observation written from memory weeks later is weaker evidence and everyone knows it, including the monitor.

Record your attempts, not just your resultsSelf-certification requires evidence that you tried for something better. "Called the shelter twice, no records retained before 2024" is documentation. An empty field is not.

Put the forty-five day clock somewhere it will be seenA staff observation of disability is a promise to produce evidence within forty-five days. If that lives only in one caseworker’s head, it survives exactly as long as that caseworker does.

Use the three free months deliberatelyEvery household may self-certify three of the twelve months. Spend that allowance on the hardest-to-evidence stretch rather than letting it go unused while somebody chases a shelter that closed.

Where the status shows up

Prioritization. Chronic status is one of the strongest factors on most CoC prioritization lists, which is the whole reason the documentation standard is this strict.
Your APR, which has a dedicated data quality table for chronic homelessness alongside the questions counting chronically homeless people and households.
The PIT count, which reports chronic status among its demographic breakdowns.
Monitoring. This is a documentation standard with named percentages, which makes it unusually easy for a monitor to test and unusually hard to argue about afterward.

About COMPASS

We make case management software for organizations doing this work. Worker observations get written inside the session rather than in a separate document, notes are versioned so an edit never quietly replaces what was originally recorded, and documents attach to the client record.

Counting your own portfolio against the three caps is not something COMPASS does for you today. That is the only pitch in this guide, and the rest is true whatever system you run.

Has a monitor tested your mix against these caps?

We would like to hear how that went and what they asked for. It is the sort of detail nobody writes down and everyone wants.

hello@illumipath.io

Sources

Your CoC may apply stricter local standards than the federal floor described here, and your HUD field office is the authority on how these requirements are monitored where you are. This page reflects HUD's published guidance as of August 2026 and we update it in place.