The moment is familiar to anyone who screens applicants. The year-to-date math on a pay stub does not reconcile. The deposits on a statement do not match the pay the stub claims. The PDF says it was produced by design software. Something is wrong, and the file is sitting open on your screen with a decision attached to it.

The first move is not the decision. It is separating what you observed from what you have concluded.

That distinction carries the whole process. An observation is a fact about a file: the numbers do not reconcile, the metadata says what it says. A conclusion is an inference about a person. Observations are defensible and specific. Conclusions made quickly, on one signal, are how good-faith applicants get harmed and how screeners end up unable to explain a decision later.

1. Slow down and re-verify the finding

Before anything else, confirm the finding is real. A surprising share of apparent discrepancies are artifacts of the review rather than the document.

If the finding survives re-verification, write it down in the narrowest factual terms you can. “Year-to-date gross on the July 15 stub is inconsistent with the stated hourly rate and the pay periods elapsed” is a usable record. “Fake stub” is not.

2. Consider the ordinary explanations first

Most documents that fail a first review are not fraudulent. They are wrong, old, incomplete, or confusing, and the applicant frequently does not know it. Working through the boring explanations first is not naivety; it is what keeps the process accurate.

There is also a category worth naming plainly. Some applicants use a paid service to “reformat” documents they believe are accurate, without understanding that the output is a fabricated document regardless of whether the underlying numbers are true. The document is still a problem. The intent may be very different from what the file suggests.

3. Ask one specific question

The most effective next step is usually a narrow, factual question about the specific discrepancy, asked without characterizing it. It costs one message, it resolves a large share of cases, and it protects you if the explanation turns out to be legitimate.

A workable question names the document, names the field, and asks for the source rather than an explanation. Asking for a document you can verify independently is more useful than asking someone to account for a number. Requesting the original download from the institution’s portal, or the adjacent month, or authorization to confirm with the employer directly, moves the question somewhere it can actually be answered.

How the applicant responds is informative, but it is not proof. People become defensive when questioned about money, and an anxious response is not evidence of anything. A refusal to provide a source that would be easy to provide is more meaningful than a tone.

4. Verify at the source before you conclude

If the question does not resolve it, the answer is almost always available outside the document. File review narrows the question; source verification answers it.

Never ask an applicant for account credentials or for live access to a financial account. It is unsafe for them, it creates liability for you, and it is not necessary to establish the facts.

5. Document the file properly

Whatever you decide, the record is what makes the decision explainable months later. This is the step most often skipped and most often regretted.

Consistency matters here beyond good practice. A screening standard applied to some applicants and not others is difficult to defend, and the defense usually depends entirely on whether the file shows the same process was run every time.

6. Handle the decision carefully

This is where the ground shifts from method to obligation, and where general guidance stops being useful.

Decisions that turn on information you obtained about an applicant frequently carry notice and process requirements. In the United States, tenant screening and lending decisions commonly implicate the Fair Credit Reporting Act, which has specific rules about adverse action notices and about what a consumer is entitled to see and dispute, particularly when a third-party screening report was involved. Housing decisions also sit under fair housing law at federal, state, and local levels, and several jurisdictions have added their own tenant screening statutes with their own notice and record-keeping requirements. Which of these applies to you depends on what you are deciding, what information you used, whether a consumer reporting agency was involved, and where you operate.

Because that answer is genuinely specific to your situation, treat the following as the shape of the problem rather than as instructions:

Ocolta does not provide legal advice, and this page is not a substitute for it. The purpose of the review process above is to make sure that whatever you are required to do, you are doing it on the basis of facts you can actually support.

7. On reporting it

People frequently ask whether a fabricated income document should be reported somewhere. There is no single answer, and the honest version is that it depends on the jurisdiction, the amount at stake, whether a transaction completed, and whether the matter involves a regulated institution with its own reporting duties.

What is reasonable to say generally: if the application was withdrawn or denied and nothing completed, most screeners document it and move on. If money changed hands, or if the document appears to have come from an organized service rather than an individual, it becomes a question for counsel and possibly for the institution whose documents were imitated. Do not make a public accusation, do not share the applicant’s documents outside your process, and do not add them to any informal shared list, which creates its own legal exposure.

The thing worth remembering

A document that fails review tells you the document failed review. It does not tell you who produced it, or why, or whether the underlying facts are true. Holding that line, in the file and in the conversation, is what separates a process that holds up from one that produces a defensible-sounding conclusion and a real problem.

Related guides