Research: Documenting Reasonable Accommodation Requests in Rental Operations

A reasonable accommodation request is one of the few rental interactions where the quality of the paper trail can matter as much as the decision itself. This research note describes a documentation model that separates what was received, what was considered, what was decided, and what was actually put in place. It is written for rental owners and managers who need a repeatable way to review those records.

The model does not decide whether a particular request must be granted. That call depends on the facts, the applicable law, and the judgment of a qualified decision maker. The purpose here is narrower: to make the evidence visible so a reviewer can see whether a process happened at all.

Methodology and scope

On September 18, 2026, we reviewed ten public sources from federal housing, disability rights, and record-keeping authorities. We used the Fair Housing Act text at Cornell Law, the HUD fair housing program page, the HUD Fair Housing Act overview, the HUD assistance animals page, the ADA service animal guidance, the ADA service animal topic page, the DOJ Fair Housing Act overview, the eCFR Part 100 fair housing regulations, the HUD fair housing office page, and NIST information quality standards.

We then translated the general principles into a four-state administrative model and applied it to hypothetical records: a phoned request with no written follow-up, a written request with an incomplete medical note, a request approved but never scheduled, and a request denied without a recorded reason. No resident, applicant, medical file, or tenancy record was reviewed.

What triggers the record

A request does not need a specific form or a particular phrase. A resident or applicant may ask for a change in a rule, policy, practice, or service because of a disability. The corresponding obligation is often described as the duty to provide a reasonable accommodation, unless doing so would impose an undue burden or fundamentally alter the nature of the housing provider's operations. Those determinations are legal and factual, and they belong to a qualified decision maker.

For documentation purposes, the important point is that the interaction itself is an event worth recording. A request that is handled verbally and never written down leaves the next reviewer with nothing to inspect. The record begins when the request is received, not when it is approved.

The four states

The first state is received. A usable record identifies the date, the channel, the person who received the request, and the substance of what was asked in the requester's own terms. Paraphrasing early is risky because later reviewers may not be able to tell what was actually requested.

The second state is evaluated. Here the record shows who reviewed the request, what information was available, and whether the provider asked for clarification or supporting documentation. Because disability-related information is sensitive, the record should show that the review happened without unnecessarily collecting medical detail. NIST privacy guidance supports limiting collection to what is needed for the stated purpose.

The third state is decided. A decision record names the decision, the decision maker, the date, and the reasoning at the level appropriate for the file. An approval and a denial both need a record; a denial that exists only in someone's memory is not reviewable.

The fourth state is implemented. This is the state most often missing. An approval that was never carried out, or was carried out on a different date than the resident was told, is an open operational item. A completion entry should identify what changed, when it changed, and how the resident was informed.

Accommodation compared with modification

Two related concepts are frequently merged in practice. A reasonable accommodation is generally a change in rules, policies, practices, or services. A reasonable modification is generally a physical change to the unit or common area. The documentation should make clear which one was requested, because the evidence and the responsible party can differ. The eCFR Part 100 regulations and the DOJ overview provide the governing framework.

Requests involving animals

Requests for an assistance animal are a common source of confused records because teams sometimes fold them into a pet policy workflow. A pet approval and an assistance animal accommodation are not the same record. The HUD assistance animals page and the ADA service animal guidance help distinguish service animals from assistance animals in housing, but the file should still show that the request was evaluated under the correct process. Pet fees and pet rules generally should not be applied to approved assistance animals, and the record should show how that distinction was handled.

Analysis: where records fail

Three failure patterns recur in hypothetical review. The first is the invisible verbal request. The interaction happened, but nothing durable exists, so no one can later confirm timeliness or substance.

The second is the orphaned approval. The decision was made and communicated, but implementation was never confirmed. The resident believes the accommodation is active; the operations team has no task.

The third is the over-collected file. The provider asks for more medical detail than the review requires, creating a privacy exposure that is larger than the decision itself needed. HUD fair housing guidance treats confidentiality as part of the process, not an afterthought.

A fourth, quieter pattern is the decision with no reasoning. Even a short note that names the factors considered gives a later reviewer a basis for consistency review.

Record fields that make review possible

A practical record does not need to be long, but it needs a small set of stable fields. Request received date, requester, channel, and requested change. Reviewer identity and review date. Information considered and any clarification requested. Decision, decision date, and decision maker. Implementation date, method, and resident notification. Confidentiality handling notes. These fields allow a later reviewer to reconstruct the sequence without re-reading a long narrative, and they make sampling tractable across many files.

Teams should decide where the record lives before a request arrives. If accommodation requests can be opened in a maintenance queue, a leasing file, and a resident portal, the portfolio may hold three partial records and no complete one. A single named queue, with a clear owner, is usually the difference between a reviewable process and a reconstruction exercise.

Inference boundaries

A complete-looking record does not prove that the legal analysis was correct. A missing field does not prove that discrimination occurred; it proves that the evidence is incomplete. Consistency across files can suggest that a process is being followed, but it cannot establish that any individual outcome was right. Those questions require legal review and, in a dispute, an adjudicator.

Limitations

This model is administrative. It does not interpret the Fair Housing Act for a specific jurisdiction, does not decide whether a request is reasonable, and does not address state or local requirements that may be stricter. It was not tested against production files and should not be used to delay an urgent safety response. A qualified owner or attorney must make the substantive calls.

Practical measurement

A portfolio can sample accommodation records and score each one against the four states, tracking missing fields rather than pass or fail. Reporting a completion rate for approved accommodations, and a median time from receipt to decision, gives operators something concrete to review. The rule version, the sampling window, and the exclusions should be fixed before the numbers are published. Teams that want a second set of eyes on the pattern can request a portfolio review, and the wider method library is available in the research archive. Readers comparing this model with a related review of assistance animal policy records will see that both depend less on the form than on a consistent, timely record of the process.