FCRA and fair housing in tenant screening: what consistent has to mean

This is general information from an operator, not legal advice. Screening rules vary by state and city and your counsel is the authority.

The practical point holds regardless of jurisdiction: the exposure is rarely a single wrong decision. It is two similar applicants receiving different answers, with no written rule anybody can point at.

Consistency is a mechanical property, not an intention

"We treat everyone the same" is a statement about intent. Consistency is a statement about process, and it has four testable parts.

Written criteria that exist before the applicant does. Income ratio, whether gross or net, what counts as income, credit and criminal standards, and what compensates for a shortfall.

The same documents requested of every applicant in the same category. Asking one applicant for two months of statements and another for six is a difference that has to be explained.

The same evaluation applied. This is where automation genuinely helps: the same checks run on every file in the same order, and nobody's file gets more scrutiny because of who they are.

Every deviation recorded, with a reason. An override is not a compliance problem. An unrecorded override is.

If you cannot produce all four for a given application, the difference between two outcomes has to be explained some other way, and that is the conversation nobody wants.

Where an automated flow helps, and where it creates new risk

Worth being even-handed, because screening software is often sold as a compliance answer and it is only partly one.

It helps by making the process identical. Every applicant gets the same checks, in the same order, and the record of what ran exists without anyone assembling it. That is a genuine improvement over human review that varies by workload and time of day.

It creates risk in two ways. A criterion that correlates with a protected class produces disparate impact whether a human or a model applies it, and applying it consistently makes it consistent rather than lawful. And a black-box score that cannot be explained is difficult to defend, which is why every ProofUp flag carries its reason in plain language rather than only a number.

Adverse action is a process with a specific trigger

When a decision is based in whole or in part on information from a consumer report, FCRA requires an adverse action notice. The mechanics matter and the details are jurisdictional, so confirm the specifics with counsel.

What is worth knowing operationally: your notice needs to be tied to the actual reason. A notice that says "did not meet criteria" when the file was declined for a specific verified income shortfall is weaker than one that names it.

Our own status here, stated plainly: ProofUp verifies income, identity, and document authenticity. Native credit and criminal checks are planned rather than shipped, and an FCRA and adverse-action compliance review is on the roadmap as work to be done rather than work completed. Today ProofUp runs alongside whatever consumer-report screening you already use, and the adverse action obligations attach to that provider's output. We would rather say that than let you assume we have taken it off your plate.

The affordable population is where consistency breaks first

This is the practical finding worth acting on.

An affordable applicant frequently has Social Security, a pension, child support, a housing voucher, and part-time wages, in combination. A screening flow built around wage income and pay stubs cannot process that, so those applicants get handled as manual exceptions.

Manual exceptions are, by definition, the files with the least consistent process and the thinnest documentation. Which means the population most likely to be protected is also the population most likely to receive an undocumented decision.

That is a compliance problem before it is an efficiency problem, and it is a strong argument for a flow that natively handles Social Security, W-2, child support, benefits, and asset documents rather than routing them to a human who will do their best.

Five things to be able to produce for any application

If an inquiry arrives, these are what get asked for. Being able to produce them in an afternoon rather than a week is most of the difference.

Your written criteria as they existed on the application date. Version them.

The documents requested and received, with dates.

Which checks ran and what each returned, in plain language.

The decision and the specific reason for it.

Any override, who made it, and why.

ProofUp keeps a full per-attempt verification history with the method and source on every record, and logs overrides as overrides. That is the audit position rather than a feature.

Read your own criteria out loud

The diagnostic that takes ten minutes: read your written screening criteria and ask whether a new leasing agent could apply them to a borderline file without asking anybody.

If they would have to ask, the criteria are incomplete and the gap is being filled by judgment that varies. That gap is where the exposure lives.

Could a new hire apply yours without asking?