Redlining used to require a map and a marker. Someone had to physically draw the line around the neighborhoods where certain people wouldn't be lent to or rented to, and the discrimination was crude, visible, and — once the country decided to care — illegal. So it didn't disappear. It got a redesign.
Today a landlord doesn't have to say "we don't rent to people like you." They don't have to say anything at all. They can let a screening algorithm, a credit-score cutoff, a criminal-record filter, or a rigid "income standard" deliver the same result, wrapped in the language of objective, neutral, businesslike policy. The rejection comes back as a number, a flag, a "you didn't meet our criteria" — and the human decision to exclude, which would be obviously illegal if spoken aloud, hides inside a system that looks like math. Same line, same people kept on the wrong side of it, better branding.
That's the argument of this article, and it's not a metaphor stretched for effect — it's a description of how exclusion actually operates in the rental market now. Screening tools that present themselves as neutral can produce sharply un-neutral outcomes, systematically disadvantaging the same groups the old redlining targeted. And because the mechanism looks technical and impersonal, tenants tend to accept the verdict as objective fact rather than recognizing it as a decision that can be biased, wrong, and in many cases unlawful. So let's strip the branding off and look at the machine underneath — how "neutral" rules discriminate, why automated denials are so dangerous, why "it's company policy" is not the end of the conversation, and what evidence you can gather if a screen shut you out.
The central trick of modern screening is that a rule can be perfectly neutral on its face and profoundly discriminatory in its effect. It never mentions race, or disability, or family status. It just quietly excludes them at higher rates.
This has a name in fair housing law: disparate impact. The idea is that a policy which is neutral on its surface — applied identically to everyone — can still be unlawful if it disproportionately excludes a protected group and isn't justified by a genuine, necessary business reason. Intent isn't required. A landlord doesn't have to want to exclude a protected group; if the policy does it in practice, that can be enough to make the policy a legal problem. This matters enormously, because it means "but I apply the same rule to everyone" is not the defense landlords think it is. Applying a discriminatory filter evenly is still discrimination if the filter lands on protected groups harder.
Look at how specific screens do this. Criminal-record screens are the starkest example, and it's not a matter of opinion — housing regulators have taken the explicit position that blanket denials based on criminal history have a disparate impact on people of color, because of well-documented racial disparities at every stage of the criminal legal system. A policy of "no one with a record" never says a word about race, yet it functions, in the regulators' own words, as a proxy for it. Credit-score cutoffs work similarly: credit scores reflect accumulated wealth, banking access, and financial history that break sharply along racial and other protected lines, so a rigid score threshold can screen out protected groups at elevated rates while looking like pure fiscal prudence. And income standards — the classic "you must earn forty times the monthly rent" — can operate as illegal source-of-income discrimination when applied to voucher holders, because the whole point of a voucher is that the tenant's income is supposed to be low; the subsidy covers the gap. Demanding that a voucher holder personally earn forty times the rent is demanding something the voucher program's own eligibility rules make impossible, which excludes voucher holders as a class under the cover of a neutral-sounding income rule.
In each case the structure is identical: a rule that mentions no protected characteristic, produces outcomes that track protected characteristics, and hides the exclusion inside the appearance of objective standards. That's not the opposite of redlining. It's redlining with a spreadsheet.
Layer on top of this the fact that a great deal of screening is now automated — run through tenant-screening companies and scoring systems — and a second problem appears, one that has nothing to do with disparate impact and everything to do with the machinery simply being wrong.
Automated screening systems are frequently, demonstrably inaccurate, and the errors fall hardest on the people least able to absorb them. The tenant-screening industry is largely unregulated, and companies often assemble reports from messy, non-official data sources — which produces mistakes at scale. Investigative reporting has found renter background checks to be wildly unreliable; in one case, a lawsuit alleged that a single screening company generated eleven thousand inaccurate reports over a handful of years. These aren't rare glitches. They're a structural feature of an industry built on cheap, fast, poorly-verified data.
Consider the specific failure modes. Mistaken identity — you get matched to someone else's record because you share a name or a birthdate, and suddenly you're carrying a stranger's eviction or conviction. Outdated eviction records — a case that was dismissed, sealed, settled, or resolved in your favor still shows up as a red flag, because the database never updated, so a case you won gets read as a case that disqualifies you. Stale or wrong criminal data — charges that were dropped, records that should be sealed, convictions attributed to the wrong person. And the whole thing is often delivered as a score or a flag with no explanation, an unchallengeable-seeming verdict that the landlord treats as gospel and you're never invited to contest.
That last part is the deepest danger: the unaccountability of it. When a human rejects you, there's at least a person who made a decision. When an algorithm rejects you, the landlord can shrug and point at the system, the system points at its data, the data came from somewhere no one will name, and you're left denied by a process with no visible author and no obvious way in. The automation doesn't just make errors — it launders them, converting a mistake or a bias into an authoritative-looking number that everyone treats as neutral truth. But it isn't truth. It's often just bad data wearing the costume of objectivity, and you have more right to challenge it than the system wants you to believe.
Here's the sentence these systems hide behind, the one meant to shut down all further discussion: it's just our policy. The rule is the rule, it's applied to everyone, nothing personal, end of story. Tenants hear it and deflate, because it sounds both final and fair. It's neither.
A policy is not automatically lawful just because it's a policy, and "we apply it to everyone" does not save a rule that produces discriminatory outcomes. That's the entire point of disparate impact: a uniformly-applied policy that disproportionately excludes a protected group can be illegal as a policy, regardless of how evenhandedly it's enforced. So when a landlord or a management company says "it's company policy," the right response isn't resignation — it's a question. What exactly is the policy, and does it create a discriminatory outcome? The policy being consistent is not the same as the policy being legal, and a blanket criminal-record ban or an income rule that excludes all voucher holders doesn't become lawful by being written down and applied to everybody.
There's a further reason "it's policy" shouldn't end the conversation: policies have to obey the law, and the law in this area has been moving. Jurisdictions are increasingly regulating exactly these screens. New York City, for instance, enacted the Fair Chance for Housing Act, effective at the start of 2025, which sharply restricts how landlords can use criminal history: they generally can't even run a criminal background check until after making a conditional offer based on the applicant's other qualifications, they can only look back a limited number of years, and they can't advertise criminal-history exclusions at all. A "policy" of screening everyone for any criminal record, up front, and rejecting on that basis isn't just arguably discriminatory in effect — in a place like that, it may be flatly illegal on its face. The company policy doesn't override the law. The law overrides the company policy.
So don't accept "it's our standard criteria" as a closed door. It's a claim about a policy, and policies can be discriminatory, outdated, or unlawful. The fact that a rule is systematized makes it look more legitimate; it doesn't make it more legal.
If a screening process shut you out and something about it feels wrong — the criteria seem designed to exclude, the denial rests on data that isn't accurate, the rule lands on a protected characteristic — you can gather evidence, and the evidence is often surprisingly gettable. Here's what to collect.
Screenshots of everything. The listing, the advertised criteria, the application portal, any stated requirements — captured with dates. Advertised screening rules are especially valuable, because a criterion that's discriminatory or unlawful on its face (an income multiple that excludes voucher holders, a flat criminal-history exclusion where those are restricted) is evidence the moment you preserve it. Systems change their public language quickly once challenged; your screenshot freezes what they actually said.
A written request for the screening criteria. Ask, in writing, what criteria were used and what standard you allegedly failed. The request itself is useful whether or not they answer: a clear, specific written statement of the criteria can expose a discriminatory rule, and a refusal to provide it is itself telling. Put it in writing so both the question and the response — or the silence — are on the record.
The rejection notice and the underlying report. Keep any denial you receive, and if the denial was based on a screening report, you generally have the right to obtain that report and to dispute errors in it. This is critical for the bad-data cases: request the report, find the mistaken identity or the outdated eviction or the sealed record, and formally dispute it. The correction can both reverse your denial and document that the system was wrong.
The advertisements. Save the ads for the unit and for the landlord's other units, because ads that state discriminatory preferences or exclusions — "no vouchers," criminal-history carve-outs where prohibited — are direct evidence, and often plainly unlawful in themselves.
Comparator applicants. This is the quiet powerhouse of a disparate-treatment or impact case: information about who did get approved, or how differently-situated applicants were treated. If applicants without the protected characteristic sailed through the same screen that stopped you, or if the "standard" seems to bend for some and not others, that comparison is powerful. Testing and comparator evidence is exactly how fair housing organizations build these cases, and anything you can document about differential treatment feeds it.
Assembled, these turn a screening denial from an anonymous, authoritative-seeming verdict into something examinable — a decision with visible criteria, traceable data, and a paper trail that a fair housing organization, a human rights agency, or a lawyer can evaluate for exactly the discrimination the branding was designed to hide.
Step back and see the whole design. The genius of screening-as-redlining is the branding — the way it takes a decision that would be nakedly illegal if spoken and dresses it in the vocabulary of objectivity: scores, criteria, standards, policies, systems. The branding does two jobs at once. It hides the discrimination from regulators, by giving every exclusion a neutral-sounding rationale. And it hides the discrimination from you, by convincing you that a number is a fact, a policy is a law, and an algorithm is a neutral judge — so that you accept the verdict and never think to challenge it. The whole apparatus runs on tenants mistaking the appearance of objectivity for the real thing.
So refuse the mistake. A screening rule is not neutral just because it's expressed as a number or a policy; it can carry every bit of the old discrimination inside a cleaner interface. A neutral-sounding criterion that disproportionately excludes protected groups can be unlawful no matter how evenly it's applied. An automated denial can be built on bad data you have the right to see and dispute. And "it's company policy" is the beginning of a question, not the end of a conversation. The exclusion may look like math, but math doesn't discriminate — the people who design the criteria and choose the cutoffs do, and their choices are subject to the law.
If a screen shut you out and it doesn't sit right, don't accept the number as a neutral verdict on your worth as a tenant. Screenshot the criteria, request the report, dispute the errors, save the ads, and find out whether the "objective standard" that rejected you was actually a line drawn around people like you — just with better branding. Find out where you stand.
