Published by HUD May 2024. U.S. DEPARTMENT OF HOUSING AND URBAN DEVELOPMENT
Office of Fair Housing and Equal Opportunity
Guidance on Application of the Fair Housing Act to the Screening of Applicants for Rental Housing
I. Introduction
This guidance from HUD’s Office of Fair Housing and Equal Opportunity explains how the Fair Housing Act protects certain rights of applicants for rental housing.1 It discusses how housing providers and companies that offer tenant screening services can screen applicants for rental housing in a nondiscriminatory way and recommends best practices for complying with the Fair Housing Act. This guidance may also help applicants understand their rights and recognize when they might have been denied housing unlawfully.
Housing providers have a legitimate interest in selecting tenants who will pay their rent and otherwise comply with lawful requirements of their lease. However, some tenant screening practices do not in fact serve these goals. Tenant screening based on imprecise or overbroad criteria may unjustifiably exclude people from housing opportunities in discriminatory ways.
These issues have been magnified in recent years by the increasing reliance by housing providers on tenant screening companies to drive tenant selection decisions. An increasing number of tenant screening companies claim that they use advanced technologies, such as machine learning and other forms of artificial intelligence (“AI”). These technologies can increase these companies’ capacity to access and analyze information about applicants that has not been widely used for rental decisions until recently but may have little bearing on whether someone will comply with their lease. These technologies can also lead to a less transparent process by obscuring the precise reasons for a denial from the housing provider and applicant.
The Fair Housing Act applies to housing decisions regardless of what technology is used.
Both housing providers and tenant screening companies have a responsibility to avoid using these technologies in a discriminatory manner.
This guidance first provides background on tenant screening companies. It then
1 Additional civil rights laws apply to recipients of Federal financial assistance, such as Title VI of the Civil Rights Act of 1964, 42 U.S.C. §§ 2000d– 2000d-7, and Section 504 of the Rehabilitation Act of 1973, 29 U.S.C. § 794(a). Although those laws are not directly addressed by this guidance, similar considerations apply under those authorities as under the Fair Housing Act. In addition, some tenant screening practices are governed by other Federal laws, such as the Fair Credit Reporting Act, 15 U.S.C. §§ 1681–1681x, the obligations under which are not covered by this guidance but are addressed in guidance from other agencies. See, e.g., Credit reporting requirements (FCRA) Resources to help industry understand, implement, and comply with the Fair Credit Reporting Act and Regulation V, CONSUMER FIN. PROT. BUREAU, https://www.consumerfinance.gov/compliance/compliance-resources/other-applicable-requirements/fair-credit-reporting-act/ (last visited April 26, 2024); What Tenant Background Screening Companies Need to Know About the Fair Credit Reporting Act, FED. TRADE COMM’N (Oct.
2016), https://www.ftc.gov/business-guidance/resources/what-tenant-background-screening-companies-need-know-about-fair-credit- reporting-act; Using Consumer Reports: What Landlords Need to Know, FED. TRADE COMM’N (July 2023), https://www.ftc.gov/business- guidance/resources/using-consumer-reports-what-landlords-need-know.
explains how the Fair Housing Act applies to both housing providers and tenant screening companies, describes common fair housing issues, and suggests how to avoid discriminatory screenings. This guidance covers screening practices with varying levels of human involvement and automation, including machine learning and other forms of AI.
II. Background
- Tenant Screening Companies
Hundreds of companies offer tenant screening services in an industry that earns about one billion dollars per year.2 Tenant screening companies obtain and analyze information about applicants from many sources and then generate a report about the person being screened, often including a recommendation for whether to accept the applicant. Housing providers then use these reports to determine which applicants to accept, deny, or accept with an additional fee or some other condition. Some housing providers also use these reports to determine whether to decline to renew the leases of any existing tenants (e.g., when a new owner or management company takes over a property).3
Tenant screening companies often market their services as a necessary, efficient, and effective tool to screen out problem tenants.4 In their marketing materials, tenant screening companies often advise housing providers to screen for credit, eviction, and criminal records, along with other items in applicants’ backgrounds that the companies claim cannot be assessed effectively without access to the software and data sources the screening company can provide.5 Some tenant screening companies also advertise that automated screenings are less likely to be discriminatory and that their services can help housing providers comply with the law.6
B. Tenant Screening Reports
Tenant screening reports vary in features, content, and price. Many companies offer basic screening reports that include information from credit reports and court records.7 These court records tend to include eviction records regardless of the case disposition, as well as a broad swath of criminal records (e.g., administrative citations, bench warrants, and traffic tickets, along with misdemeanors and felonies). Some companies also offer screening reports with additional details, such as medical debt, foreclosures, student loans, and even the number
2 CONSUMER FIN. PROT. BUREAU, TENANT BACKGROUND CHECKS MARKET 10 (2022) [hereinafter, CFPB TENANT BACKGROUND CHECKS MARKET],
3 See, e.g., Charge of Discrimination ¶¶ 14–16, Sec’y of Dep’t of Hous. & Urb. Dev. ex rel. Loveless v. Wesley Apt. Homes, LLC, No. 17-AF-0046- FH-004 (H.U.D. O.H.A. filed Jan. 18, 2017).
4 CFPB TENANT BACKGROUND CHECKS MARKET at 13.
5 Id. at 11.
6 See, e.g., Conn. Fair Hous. Ctr. v. CoreLogic Rental Prop. Sols., LLC, 478 F. Supp. 3d 259, 275 (D. Conn. 2020).
7 Tech, Bias, and Housing Initiative: Tenant Screening, TECHEQUITY COLLABORATIVE (Feb. 23, 2022), https://techequitycollaborative.org/2022/02/23/tech-bias-and-housing-initiative-tenant-screening/.
of phone numbers an applicant has had.8 Some reports include a recommendation to accept or deny the applicant or provide a numerical score or grade. Some reports detail the records found, while others simply state if the applicant “passed” or “failed” in various areas.9 Sample tenant screening reports are appended to the end of this guidance.
Generally, automated software is used to generate these screening reports, and companies tend not to disclose how this software works, including the extent to which it uses advanced technologies, such as AI. Companies typically input a prospective tenant’s self- disclosed personal information into a computer program that then matches this information to records in numerous data sources. Companies utilize public databases of records and may also purchase additional datasets from court systems or data aggregators.10 Some of these datasets are fixed at the time they are purchased, meaning that screening reports will not reflect updates made afterwards to the underlying records.11
The level of involvement by housing providers in setting the criteria and standards used for screening reports varies. The specific criteria that are part of the screening process (e.g., credit score) and the specific standards set for a positive recommendation (e.g., over 700) may be set by the housing provider, the tenant screening company, or both, depending on the company and services offered. Tenant screening companies decide whether to have default settings and control the range of options they make available to housing providers.
Tenant screening companies have increasingly begun advertising that they rely on advanced technologies, including machine learning and other forms of AI.12 However, companies tend to disclose few details about which specific functions these technologies perform and what safeguards are in place. Some systems might rely on technology simply to gather and synthesize data from a variety of sources.13 For example, these systems could comb through thousands of court records and report that a prospective tenant has three misdemeanors, leaving it to an employee of the screening company or the housing provider to set the standard for whether that is a reason to reject the applicant. Other systems might go a step further, by also relying on technology to determine which criteria and standards are
8 Erin Smith & Heather Vogell, How Your Shadow Credit Score Could Decide Whether You Get an Apartment, PROPUBLICA (Mar. 29, 2022, 6:00 AM), https://www.propublica.org/article/how-your-shadow-credit-score-could-decide-whether-you-get-an-apartment.
9 Lauren Kirchner & Matthew Goldstein, Access Denied: Faulty Automated Background Checks Freeze Out Renters, MARKUP (May 28, 2020, 5:00 AM), https://themarkup.org/locked-out/2020/05/28/access-denied-faulty-automated-background-checks-freeze-out-renters; see also Tenant Screening. Why It’s Important and How We Do It., WEICHERT, REALTORS – LILAC PROPS., https://www.weichertlilac.com/tenant-screening-why-its- important-and-how-we-do-it/ (last visited Apr. 22, 2024).
10 CFPB TENANT BACKGROUND CHECKS MARKET at 19.
11 Id. (citing NAT’L CONSUMER L. CTR., BROKEN RECORDS REDUX: HOW ERRORS BY CRIMINAL BACKGROUND CHECK COMPANIES CONTINUE TO HARM CONSUMERS
SEEKING JOBS AND HOUSING 17 (2019), https://www.nclc.org/images/pdf/criminal-justice/report-broken-records-redux.pdf). Under the Fair Credit Reporting Act, tenant screening companies must have procedures in place to include any existing disposition information for court filings. See Advisory Opinion, Fair Credit Reporting; Background Screening, 89 Fed. Reg. 4171, 4172 (Jan. 23, 2024).
12 CFPB TENANT BACKGROUND CHECKS MARKET at 11–13.
13 Valerie Schneider, Locked Out by Big Data: How Big Data, Algorithms and Machine Learning May Undermine Housing Justice, 52 COLUM. HUM. RTS. L. REV. 251, 267 (2020) [hereinafter, Locked Out by Big Data].
relevant.14 For example, a model could be given information about a group of people, some of whom were evicted and some of whom were not, and be directed to “learn” which variables correlate with whether an individual was evicted. The model would then use those variables as criteria for screening prospective applicants going forward.15
III. Legal Responsibilities Under the Fair Housing Act
The Fair Housing Act (“Act”) prohibits discrimination in the sale, rental, and financing of dwellings and in other housing-related services because of race, color, religion, sex (including sexual orientation and gender identity), national origin, disability or familial status (children under eighteen being present, seeking legal custody of minor children, or pregnancy).16 The Act prohibits intentionally discriminatory practices, as well as those with an unjustified discriminatory effect.17
“By drafting the Act in the passive voice, Congress ‘bann[ed] an outcome while not saying who the actor is.’”18 Thus, any entity that “play[s] a substantial role” in a housing decision can be liable under the Fair Housing Act, even if that entity or person is not the ultimate or sole decisionmaker.19 This includes those with input into what variables are considered or what standards should be used in making a housing decision.20
A. Liability of Housing Providers
Under principles of direct liability, housing providers are responsible for ensuring their rental decisions comply with the Fair Housing Act even if they have largely outsourced the task of screening applicants to a tenant screening company.21 The screening company may be responsible too, but housing providers retain authority over screening practices and decisions at their properties.
In addition, under principles of vicarious liability, a principal is liable for the actions of
14 CHI CHI WU, ARIEL NELSON, APRIL KUEHNHOFF & CAROLINE COHN, NAT’L CONSUMER L. CTR., DIGITAL DENIALS: HOW ABUSE, BIAS, AND LACK OF TRANSPARENCY IN
TENANT SCREENING HARM RENTERS 8 (2023) [hereinafter, NAT’L CONSUMER L. CTR. DIGITAL DENIALS], https://www.nclc.org/wp-
content/uploads/2023/09/202309_Report_Digital-Denials.pdf.
15 See, e.g., Locked Out by Big Data, at 275–278.
16 42 U.S.C. §§ 3601–19.
17 See 24 C.F.R. § 100.500; Tex. Dep’t of Hous. & Cmty. Affairs v. Inclusive Cmtys. Project, Inc., 576 U.S. 519, 540 (2015).
18 Conn. Fair Hous. Ctr. v. Corelogic Rental Prop. Sols., LLC, 369 F. Supp. 3d 362, 370 (D. Conn. 2019) (quoting NAACP v. Am. Family Mut. Ins. Co., 978 F.2d 287, 298 (7th Cir. 1992), cert. denied, 508 U.S. 907 (1993)).
19 Commonwealth v. Priority Home Sols., LLC, CL20000688-00, at 6 (Va. Cir. Ct. Dec. 6, 2021); Sabal Palm Condos. of Pine Ridge Ass’n, Inc. v. Fischer, 6 F. Supp. 3d 1272, 1293 (S.D. Fla. 2014).
20 Louis v. SafeRent Sols., LLC, Civ. No. 22-CV-10800-AK, 2023 WL 4766192, at *8–*10 (D. Mass. July 26, 2023); Conn. Fair Hous. Ctr., 369 F. Supp. 3d at 374–76; see also Conn. Fair Hous. Ctr. v. Corelogic Rental Prop. Sols., LLC, No. 3:18-cv-705-VLB, 2023 WL 4669482, at *19 (D. Conn. July 20, 2023), appeal docketed No. 23-1166 (2d Cir. Aug. 14, 2023) (appealed on other grounds).
21 See 24 C.F.R. 100.7(a).
their agent done within the scope of the agent’s authority.22 A principal can be liable if the agency relationship is actual or if it is apparent.23 “A person is vicariously liable for a discriminatory housing practice by the person’s agent or employee, regardless of whether the person knew or should have known of the conduct that resulted in a discriminatory housing practice.”24 If a tenant screening company is an actual or apparent agent of a housing provider, any discriminatory action taken by the company can also be attributed to the housing provider.
B. Liability of Tenant Screening Companies
Tenant screening companies can be held liable for discriminatory decisions caused by their practices.25 The Fair Housing Act provides exceptions from liability for a notably small selection of entities, and otherwise “focuses on prohibited acts,” not specific actors.26 The Act broadly permits entities to “be held liable when they have personally committed or contributed to a Fair Housing Act violation,”27 in other words if their “conduct . . . results in a discriminatory housing practice.”28 This view of the Act’s coverage is consistent with decades of judicial precedent interpreting the Act to apply to a wide range of persons and entities that provide services in connection with housing but do not directly provide housing themselves or make final housing decisions.29 In fact, two courts have adopted this argument to find that in certain circumstances tenant screening companies can be liable under the Act.30
Thus, a tenant screening company can be held liable under the Act for facilitating a discriminatory decision, even if the housing provider makes the ultimate decision as a formal
22 Meyer v. Holley, 537 U.S. 280, 282, 285–86 (2003); City of Chicago v. Matchmaker Real Estate Sales Ctr., Inc., 982 F.2d 1086, 1096–97 (7th Cir. 1992). A principal is also vicariously liable for the actions of their agents taken outside the scope of their relationship when the agent is aided in the commission of such acts by the existence of the agency relationship. See Quid Pro Quo and Hostile Environment Harassment and Liability for Discriminatory Housing Practices Under the Fair Housing Act, 81 Fed. Reg. 63,054, 63,072 (Sept. 14, 2016) (collecting cases).
23 See, e.g., Pinchback v Armistead Homes Corp., 689 F. Supp. 541, 550-51 (D. Md. 1988) judgment aff’d in part, vacated in part on other grounds, 907 F.2d 1447 (4th Cir. 1990); United States v. Balistrieri, 981 F.2d 916, 930 (7th Cir. 1992).
24 24 C.F.R § 100.7(b); see also Meyer, 537 U.S. at 282, 285-86; Matchmaker Real Estate Sales Ctr., 982 F.2d at 1096.
25 See FED. TRADE COMM’N, CONSUMER FIN. PROT. BUREAU, DEP’T OF HOUS. & URB. DEV. & DEP’T OF JUSTICE, TENANT BACKGROUND CHECK AND YOUR RIGHTS 4
(2024), https://www.hud.gov/sites/dfiles/FHEO/documents/HUD_Tenant_Background_Checks_and_Your_Rights.pdf (“The Fair Housing Act makes it illegal for tenant background check companies . . . to discriminate against you.”); DEP’T OF HOUS. & URB. DEV. OFF. OF FAIR HOUS. & EQUAL OPPORTUNITY, IMPLEMENTATION OF THE OFFICE OF GENERAL COUNSEL’S GUIDANCE ON APPLICATION OF FAIR HOUSING ACT STANDARDS TO THE USE OF CRIMINAL RECORDS BY PROVIDERS OF HOUSING AND REAL ESTATE-RELATED TRANSACTIONS 3 (2022) [hereinafter, HUD 2022 Criminal Records Guidance],
https://www.hud.gov/sites/dfiles/FHEO/documents/Implementation%20of%20OGC%20Guidance%20on%20Application%20of%20FHA%20Stan dards%20to%20the%20Use%20of%20Criminal%20Records%20-%20June%2010%202022.pdf (noting applicability of Fair Housing Act guidance to third-party screening companies).
26 Meyer, 537 U.S. at 285.
27 Sabal Palm Condos. of Pine Island Ridge Ass’n, Inc. v. Fischer, 6 F. Supp. 3d 1272, 1293 (S.D. Fla. 2014).
28 24 C.F.R. § 100.7(a)(1)(i).
29 See, e.g., Swanson v. Citibank, N.A., 614 F.3d 400, 406 (7th Cir. 2010) (applying the Act to lenders); Fair Hous. Council of San Fernando Valley
- Roommates.com, LLC, 521 F.3d 1157, 1166–1175 (9th Cir. 2008) (en banc) (housing search website); NAACP v. Am. Fam. Mut. Ins. Co., 978 F.2d 287, 290 (7th Cir. 1992) (insurers); United States v. City of Black Jack, 508 F.2d 1179, 1184–85 (8th Cir. 1974) (state and municipal bodies); Balt. Neighborhoods, Inc. v. Rommel Builders, Inc., 3 F. Supp. 2d 661, 665 (D. Md. 1998) (developers); United States v. Am. Inst. of Real Estate Appraisers, 442 F. Supp. 1072, 1079 (N.D. Ill. 1977) (appraisers); Rivera v. Mora Dev. Corp., 624 F. Supp. 3d 80, 88–92 (D.P.R. 2022) (condominium associations).
30 Louis v. SafeRent Sols., LLC, Civ. No. 22-CV-10800-AK, 2023 WL 4766192, at *8–*10 (D. Mass. July 26, 2023); Conn. Fair Hous. Ctr. v. Corelogic Rental Prop. Sols., LLC, 369 F. Supp. 3d 362 (D. Conn. 2019). But see Conn. Fair Hous. Ctr. v. Corelogic Rental Prop. Sols., LLC, No. 3:18-cv-705- VLB, 2023 WL 4669482, at *20 (D. Conn. July 20, 2023), appeal docketed No. 23-1166 (2d Cir. Aug. 14, 2023) (finding that the plaintiff had not proven the tenant screening company made housing unavailable in this instance).
matter.31 The screening report does not need to direct the housing provider explicitly to accept or deny the applicant to have such influence; as a practical matter, it often is sufficient that the report provides a score, a recommendation, or any other sort of explicit or implicit opinion as to the relevance of the information gathered to the disposition of the application. That is particularly true where the screening company contributed to the decision of what criteria to include, what standards to set for those criteria, or how to weigh them.32 Tenant screening companies also bear responsibility for policies and practices related to the accuracy or completeness of their records that cause discriminatory results.33 In general, disclaimers of liability cannot shield a party from potential liability under the Act.34
IV. Proving Discrimination
- Discriminatory Intent
Intentionally discriminatory tenant screening policies and practices violate the Fair Housing Act. Direct evidence of discriminatory intent (e.g., statements) may exist, or intent can be proven by circumstantial evidence using one of several well-established frameworks.35
A housing provider or a tenant screening company can violate the Fair Housing Act by using a protected characteristic – or a proxy for a protected characteristic – as a screening criterion.36 This is true even if the decision for how to screen applicants is made in whole or in part by an automated system, including a system using machine learning or another form of AI.
Intentional discrimination also includes screening applicants in a particular way in order to exclude members of a protected class or discourage them from applying. For example, if a housing provider adopts an overbroad criminal records screening policy as a way of preventing “too many” Black people from living at the property, that constitutes intentional discrimination even if some Black applicants are admitted and some White applicants are denied.37
Applying screening criteria inconsistently across protected classes is another form of intentional discrimination. For example, intentional discrimination includes making exceptions to a purported substantive requirement for White applicants but not similarly situated Black
31 See SafeRent, 2023 WL 4766192, at *8–*10; Vargas v. Facebook, Inc., No. 21-16499, 2023 WL 6784359, at *1–*3 (9th Cir. Oct. 13, 2023);
Roommates.com, LLC, 666 F.3d at 1219.
32 Conn. Fair Hous. Ctr., 2023 WL 4669482, at *2–*8, *19–*20.
33 See HUD 2022 Criminal Records Guidance at 3; see also 15 U.S.C. § 1681e.
34 See Smith v. Hillsdale Village, Civ. No. 17-0883(KM), 2018 WL 588923, at *8 (D.N.J. Jan. 26, 2018).
35 See, e.g., Village of Arlington Heights v. Metro. Hous. Dev. Corp., 429 U.S. 252, 264–68 (1977); McDonnell Douglas Corp. v. Green, 411 U.S. 792, 802–05 (1973); see also, Budnick v. Town of Carefree, 518 F.3d 1109, 1114 (9th Cir. 2008) (discriminatory intent may be proven by “simply produc[ing] direct or circumstantial evidence demonstrating that a discriminatory reason more likely than not motivated the challenged decision.”).
36 A proxy is a variable that is highly correlated with a protected class, as for example ZIP Code can be highly correlated with race.
37 See Pac. Shores Props., LLC v. City of Newport Beach, 730 F.3d 1142, 1161 (9th Cir. 2013) (“[W]illingness to discriminate against both minorities and White citizens . . . d[oes] not cleanse [a party] of any discriminatory intent that it may have harbored.”).
applicants.38 Providing more help to applicants with negative records from some groups than to those from others can also be intentional discrimination (e.g., only advising White applicants to pay down their debt and reapply).39
B. Unjustified Discriminatory Effect
Even when there is no intent to discriminate, a policy or practice violates the Fair Housing Act if it has a discriminatory effect and (i) the policy or practice is not necessary to achieve a substantial, legitimate, non-discriminatory interest or (ii) the interest can be served by a less discriminatory alternative.40 Courts have recognized discriminatory effects claims related to tenant screening practices.41 Discriminatory effects liability is assessed using a three- step burden-shifting framework and requires a fact-specific analysis.42
In the first step of the analysis, a plaintiff (or HUD in an administrative enforcement action43) has the burden to prove that a screening policy or practice has a discriminatory effect. In other words, they must show that (i) it causes or predictably will cause a disparate impact based on one or more of the characteristics protected by the Fair Housing Act, or (ii) it creates, perpetuates, increases, or reinforces segregation.44 National, state, or local statistics can be used to show that the policy or practice causes or predictably will cause a discriminatory effect.45 Certain screening practices will likely have a disparate impact based on race, national origin, sex, or disability because of systemic inequalities in areas such as credit, housing, and criminal justice (as discussed in Section V.C below). Screening practices perpetuate segregation if they cause Black or Brown applicants to be excluded or displaced from an area that is, or is becoming, predominantly White.
If a plaintiff or HUD proves the tenant screening policy or practice has a discriminatory effect, the burden shifts to the defendant (or respondent in an administrative action) to prove
38 See, e.g., Watson v. Pathway Fin., 702 F. Supp. 186, 188–89 (N.D. Ill. 1988).
39 See, e.g., United States v. Lorantffy Care Ctr., 999 F. Supp. 1037, 1041, 1045 (N.D. Ohio 1998); Hobson v. George Humphreys, Inc., 563 F. Supp. 344 (W.D. Tenn. 1982).
40 See 24 C.F.R. § 100.500; see also Tex. Dep’t of Hous. & Cmty. Aff. v. Inclusive Cmtys. Project, Inc., 576 U.S. 519, 541 (2015) (citing HUD’s 2013 rule). HUD’s 2023 Discriminatory Effects Rule formally restores, with minor unrelated modifications, HUD’s 2013 Discriminatory Effects Rule. An intervening version of the Rule published in 2020 was enjoined before it took effect, so the framework established by the 2013 Rule has been in effect without interruption since it was issued. See Mass. Fair Hous. Ctr. v. Dep’t of Hous. & Urb. Dev., 496 F. Supp. 3d 600 (D. Mass.
2020).
41 See, e.g., Louis v. SafeRent Sol.s., LLC, Civ. No. 22-CV-10800-AK, 2023 WL 4766192, at *8–*10 (D. Mass. July 26, 2023); La. Fair Hous. Action Ctr. v. Azalea Garden Props., LLC, Civ. No. 22-74, 2022 WL 1262642, at *6 (E.D. La. Apr. 28, 2022), rev’d on other grounds, 82 F.4th 345 (5th Cir. 2023); Jones v. City of Faribault, Civ. No. 18-1643 (JRT/HB), 2021 WL 1192466, at *19 (D. Minn. Feb. 18, 2021); Conn. Fair Hous. Ctr. v. CoreLogic Rental Prop. Sols., LLC, 478 F. Supp. 3d 259, 291–93 (D. Conn. 2020).
42 24 C.F.R. § 100.500. While HUD’s discriminatory effects rule is largely consistent with that of the Federal Courts of Appeals, some circuits may analyze discriminatory effects claims under a different framework.
43 Anyone who has experienced discrimination can file a complaint with HUD at no cost. HUD’s administrative process for these complaints is explained on HUD’s website. See also 24 C.F.R. pts. 103, 180.
44 24 C.F.R. § 100.500(a), (c)(1). The seven characteristics protected by the Fair Housing Act are race, color, national origin, religion, sex
(including sexual orientation and gender identity) disability, and familial status.
45 See, e.g., DEP’T OF HOUS. & URB. DEV. OFF. OF GEN. COUNSEL, GUIDANCE ON APPLICATION OF FAIR HOUSING ACT STANDARDS TO THE USE OF CRIMINAL RECORDS BY PROVIDERS OF HOUSING AND REAL ESTATE-RELATED TRANSACTIONS 3 (2016) [hereinafter, HUD 2016 Criminal Records Guidance],
https://www.hud.gov/sites/documents/HUD_OGCGUIDAPPFHASTANDCR.PDF.
that the policy or practice is necessary to achieve a substantial, legitimate, nondiscriminatory interest.46 This interest may not be hypothetical or speculative.47 Only screening criteria and standards – including complex models – that are well-tailored to predict future behavior relevant to tenancy can be considered necessary to achieving a legitimate interest.
If a defendant or respondent successfully meets this burden, the burden shifts back to the plaintiff or HUD to prove that such interest could be served by another practice that has a less discriminatory effect.48 If a screening policy is overbroad (i.e., it screens out unproblematic applicants), a more targeted screening policy could be a less discriminatory alternative.
V. Requirements and Recommendations for Fair Housing Act Compliance
This section discusses best practices that can help housing providers and tenant screening companies comply with the Fair Housing Act. Some practices discussed in this section may be required by other Federal laws, such as the Fair Credit Reporting Act,49 or by state or local laws. Screening practices prohibited by any applicable law will not be considered necessary to achieve a substantial, legitimate, nondiscriminatory interest under the Fair Housing Act.50
A. The Relationship Between Housing Providers and Screening Companies
- The Role of Housing Providers
Housing providers should remember that they are responsible for avoiding discriminatory housing decisions, even when they use a tenant screening company to assist in the process. They should develop policies and practices to ensure that all denials reflect their own sound judgment.
Housing providers should adopt screening policies that are clear, detailed, and publicly available, and only use tenant screening services that will help them implement these policies.51 Customizing the criteria, standards, and weights being used, rather than purchasing an “off the shelf” product, can help ensure screenings conform to stated policies.
When a housing provider receives a screening report with a denial recommendation, the housing provider should make an independent determination whether, under their screening policies, the information in the report is in fact disqualifying. If not, the housing provider should accept an applicant notwithstanding the denial recommendation and consider
46 See 24 C.F.R. § 100.500(c)(2); see also Inclusive Cmtys. Project, 576 U.S. at 541.
47 24 C.F.R. § 100.500(c)(2).
48 24 C.F.R. § 100.500(c)(3); accord Inclusive Cmtys. Project, 576 U.S. at 527.
49 15 U.S.C. §§ 1681–1681x; see generally Advisory Opinion, Fair Credit Reporting; Background Screening, 89 Fed. Reg. 4171 (Jan. 23, 2024).
50 See Langolis v. Abington Hous. Auth., 234 F. Supp. 2d 33, 69 (D. Mass. 2002) (“If the justification is in tension with other laws, it is difficult to see how it can qualify as ‘legitimate’ in order to justify a measure with a racially disparate impact.”).
51 See HUD 2022 Criminal Records Guidance at 9–10.
contacting the screening company to adjust the grounds for denial recommendations going forward.
Regardless of whether the housing provider used a screening company or found the negative information themselves, applicants should be given the same opportunity to dispute the accuracy or relevance of any negative information.52 Housing providers that use a screening company ultimately bear any responsibility they otherwise would have for ensuring applicants have a full opportunity to contest the accuracy or relevance of any potential ground for denial.53 (This process is discussed further in Section V.B.5 below.)
In selecting a tenant screening company, housing providers should inquire into the ways in which the company ensures its screenings are accurate and nondiscriminatory.54 Housing providers should select screening services that (a) offer customizability; (b) frequently update their data; (c) monitor for unjustified discriminatory effects; (d) report clear and specific reasons for denials; (e) allow individuals to correct inaccuracies; (f) publicly disclose key details about their screening systems; and (g) comply with all applicable Federal, state, and local laws.
- The Role of Tenant Screening Companies
As technology continues to evolve, all parties subject to the Fair Housing Act have an obligation to ensure that tenant screening models comply with the law. The more information a tenant screening company discloses, the easier it will be for housing providers to assess whether using a particular tenant screening service meets their needs and promotes compliance with their fair housing responsibilities. This information could include what, if any, advanced technologies the screening company uses, what functions these technologies perform, and what safeguards are in place; the sources from which the screening company draws records and how they ensure those records are complete and accurate; and how decisions are made about criteria, standards, and weights.
As a best practice, tenant screening companies should conduct civil rights monitoring – akin to that done routinely by mortgage lenders and others that rely on complex models – to ensure their models comply with the Act. This includes checking inputs for protected characteristics and their close proxies, as well as ensuring they are not being replicated by constellations of factors.55 Even if specific protected characteristics (and their more obvious proxies) are removed from a model’s inputs, machine learning and other forms of AI — without corrective adjustments — may compensate for their removal by altering the weights of the
52 Id. at 7.
53 The dispute process provided by some screening companies might be more limited (i.e., designed to meet the minimum requirements of consumer protection laws) in which case the housing provider should engage with the applicant and consider all mitigating information themselves.
54 See HUD 2022 Criminal Records Guidance at 9–10.
55 In complex models, a very large number of factors, each of which may be weighted differently, can function as a proxy for a protected characteristic. See Alice Xiang, Reconciling Legal and Technical Approaches to Algorithmic Bias, 88 TENN. L. REV. 649, 667 (2021).
remaining factors, thereby recreating the same discriminatory outcomes.56 Monitoring should also look at whether any datasets used are more inaccurate or incomplete for certain groups than others. It also includes evaluating the model’s outputs to ensure the model is not having an unjustified discriminatory effect. Even complex models should be interpretable enough that adjustments can be made to fix any fair housing problems that are found.
As a general matter, tenant screening companies should serve to help implement, rather than effectively set, a housing provider’s screening policies. To that end, screening companies should offer customizability, for example as to criteria, standards, and weights. Screening companies should not offer options likely to pose fair housing concerns (e.g., unlimited or lifetime lookback periods for criminal records). Screening companies should either not provide default standards (i.e., they instead should prompt housing providers to choose standards without any pre-selected) or they should set defaults to the options that are least exclusionary.
Denial recommendations or low “grades” should not be provided to housing providers in a conclusory fashion. Rather, screening reports should include all available relevant details about the basis for the determination (e.g., dates, locations, case numbers, and dispositions). Screening reports should also list all sources from which data about the applicant was drawn and state the standard that would have been needed for a positive result. These details should be in plain language, so a lay person can easily understand the reasons for the denial.57
Under the Fair Credit Reporting Act, tenant screening companies must provide a mechanism for individuals to dispute inaccuracies and have them corrected.58 To promote compliance with the Fair Housing Act, this process should also enable applicants to dispute whether a record should be included in their file as a potential grounds for denial of tenancy even if the record is accurate (e.g., an eviction related to domestic violence). The screening company should send corrected reports, scores, and recommendations to the housing provider and applicant, and carry forward corrections to all future screenings the screening company does of that individual. This process should be relatively quick, so that housing providers do not have to choose between holding a unit open indefinitely and unjustifiably denying a particular applicant.
56 Id.
57 Highly complex models – such as those that use machine learning or other forms of AI – can make the precise reasons for a denial difficult to discern and disclose. Additional considerations for such models are discussed in Section V.B.6 below. See generally CONSUMER FIN. PROT. BUREAU, CIRCULAR 2023-03, ADVERSE ACTION NOTIFICATION REQUIREMENTS AND THE PROPER USE OF THE CFPB’S SAMPLE FORMS PROVIDED IN REGULATION B 3
(2023), https://files.consumerfinance.gov/f/documents/cfpb_adverse_action_notice_circular_2023-09.pdf.
58 The Fair Credit Reporting Act (“FCRA”) requires tenant screening companies to investigate claims by screened applicants that records are inaccurate or incomplete. 15 U.S.C. § 1681i(a)(1)(A). The recommendations herein that go beyond what FCRA requires can help ensure compliance with the Fair Housing Act by preventing further erroneous denials that may disproportionately harm certain groups.
B. Guiding Principles for Non-Discriminatory Screenings
The principles in this section can be used by housing providers – those that conduct screenings themselves and those that outsource the task – as well as by tenant screening companies to help comply with the Fair Housing Act.
- Choose Relevant Screening Criteria
Screening applicants only for information relevant to the likelihood that they will comply with their tenancy obligations can help ensure that screenings are fair and nondiscriminatory.
If a screening policy disproportionately excludes applicants of a certain race or other protected class, conducting the screening in a more precise way may be a less discriminatory alternative.
Past actions unrelated to tenancy and past incidents unlikely to recur (e.g., eviction due to job loss) should not be the basis for denials. Screening criteria should be waived if they are not relevant for an applicant’s individual circumstances, even if the criteria might be relevant in general (e.g., minimum income for an applicant whose rent will be paid by someone else).
Some records are more relevant than others, and the most relevant records should be prioritized (e.g., recent records versus older ones), including when programming complex models. Housing providers should avoid screening existing tenants of a property who have lived there without incident because their behavior at that property is more relevant than a records-based screening.
Records without a negative outcome are not relevant. For example, the record of an eviction proceeding has no relevance if the tenant prevailed. If a court record does not provide enough information to determine who prevailed, it should be disregarded unless additional information about the disposition is obtained.
Whether done manually or by an automated system, all types of relevant information should be considered. For example, less common sources of income (e.g., SSDI) or other financial resources (e.g., Housing Choice Vouchers, formerly known as Section 8 vouchers) should be considered when assessing an applicant’s ability to afford the rent.
- Use Only Accurate Records
Ensuring records are accurate can help avoid discriminatory screenings. Certain types of inaccuracies are more likely to occur for members of some demographic groups than others.
Even if the rates of inaccuracies within a dataset are the same across demographic groups, groups that are overrepresented in that dataset will be disproportionately affected by any problems the dataset has. Screenings that have a discriminatory effect may not be considered legally justified if they are based on records rife with inaccuracies or if a more accurate option is
available.
Datasets used for tenant screenings are often incomplete, missing key personal identifiers, or updated infrequently.59 Automated systems might miscategorize records with missing or unclear information if they are not programed to account for those scenarios. For example, a court record might not clearly state whether the record is for a felony versus a misdemeanor or a criminal violation versus a civil violation.
Often records of people bearing the same or similar names are erroneously attributed to the wrong person, a problem that is more common for last names more prevalent among Hispanic, Asian, or Black individuals.60 This can happen when records are used that list only partial information (e.g., an initial instead of a first name) or that lack corroborating details (e.g., a name without a date of birth). This can also happen when a housing provider or screener puts less specific information into a search than is available to them (e.g., inputting a state instead of a full address). Records should match multiple pieces of identifying information, and “wildcard” or “name-only” matching should not be used.61
- Follow the Applicable Screening Policy
Records outside the scope of a housing provider’s stated screening policy should not be considered.62 For example, misdemeanors or civil violations should not be considered under a policy of screening for felony convictions. This can happen, for example, when a housing provider does not check whether their screening company is using the standards in the housing provider’s policy or when a screening company does not properly limit its screening to conform to the housing provider’s policy.63 It can also happen if records are inaccurately categorized, as described above.
Housing providers should not ask applicants about any criminal, credit, or housing history that falls outside the scope of their screening policies.64 Housing providers who use automated screenings should consider not asking applicants any questions about their history (i.e., not even within the scope of their policies) because such questions can confuse or discourage applicants while not giving the housing provider any information beyond that which
59 CHI CHI WU & ARIEL NELSON, NAT’L CONSUMER LAW CTR., MISSION CREEP: A PRIMER ON USE OF CREDIT REPORTS & SCORES FOR NON-CREDIT PURPOSES 7
(2022), https://www.nclc.org/wp-content/uploads/2022/08/Mission_Creep_rpt.pdf; see also Ariel Nelson, NAT’L CONSUMER LAW CTR., BROKEN RECORDS REDUX: HOW ERRORS BY CRIMINAL BACKGROUND CHECK COMPANIES CONTINUE TO HARM CONSUMERS SEEKING JOBS & HOUSING 15-23 (2019),
https://www.nclc.org/wp-content/uploads/2022/09/report-broken-records-redux.pdf.
60 Advisory Opinion, Fair Credit Reporting: Name-Only Matching Procedures, 86 Fed. Reg. 62,468, 62,469–70, (Nov. 10, 2021).
61 See CFPB TENANT BACKGROUND CHECKS MARKET at 26–28; Advisory Opinion, Fair Credit Reporting: Name-Only Matching Procedures, 86 Fed. Reg. at 62,469–70.
62 See DEP’T OF HOUS. & URB. DEV. OFF. OF FAIR HOUS. & EQUAL OPPORTUNITY, GUIDANCE ON COMPLIANCE WITH TITLE VI OF THE CIVIL RIGHTS ACT IN MARKETING AND APPLICATION PROCESSING AT SUBSIDIZED MULTIFAMILY PROPERTIES 7 (2022) [hereinafter, FHEO TITLE VI MARKETING GUIDANCE],
63 See id.
64 See id.
they will learn from the automated screening.
- Be Transparent with Applicants
Transparency helps ensure consistent, nondiscriminatory results by making sure everyone has clear, complete information.65 Applicants need to know how they will be screened to decide whether to pay an application fee and take the time to fill out an application. This challenge is particularly acute for low-income applicants who cannot afford to pay multiple application fees, as well as for applicants with inflexible work or caretaking schedules since some housing providers only accept applications in-person during business hours.66 Moreover, systematically giving applicants complete, detailed information ensures that some applicants are not more likely to pass the screening than others simply because they are better informed about how to navigate the process. Transparency is important throughout the application process so that applicants know beforehand how they will be screened and afterwards why they were denied.
Tenant screening policies should be in writing, made public, and readily available to potential applicants. Prior to applying, potential applicants should be given a copy of the screening policies or told where they can find them (e.g., the link to a website). Screening policies should contain enough detail for an applicant to tell whether they are likely to qualify (e.g., what records will be considered, what incidents will be disqualifying, and how far back the screening will look). Applicants should also be given information about how evidence of mitigating circumstances can be submitted and will be treated, how to request a reasonable accommodation for a disability,67 and how to contest an inaccurate, incomplete, or irrelevant record. Publicizing this information can increase the number of qualified applicants by attracting those who might have assumed different standards apply. Transparency up front can also reduce the number of unqualified applicants, saving housing providers and applicants time and expense.
Denial letters should contain as much detail as possible as to all reasons for the denial, including the specific standard(s) that the applicant did not meet and how they fell short. For example, rather than saying “you were denied because of your credit score,” the letter should say “you were denied because we require a credit score of XXX and you have a credit score of XXX according to XXX service.” If an applicant fails multiple screening criteria, all of those criteria should be included in the denial letter. All records relied on should be attached, including any screening reports. Denial notifications should also instruct applicants how to
65See id. at 5–6.
66 Accepting applications at a broad range of times and through a variety of means can help ensure equal access to housing opportunities because overly restrictive application procedures can be discriminatory. FHEO TITLE VI MARKETING GUIDANCE at 5.
67 The Fair Housing Act makes it unlawful for any person to refuse to make reasonable accommodations (i.e., a change, exception, or adjustment) to a rule, policy, practice, or service, when such accommodations may be necessary to afford individuals with disabilities equal opportunity to use and enjoy a dwelling, including public spaces and common use areas. 24 CFR § 100.204(a).
submit an appeal if a record is inaccurate, incomplete, or irrelevant; mitigating circumstance exists; or a reasonable accommodation for a disability is needed.
- Allow Applicants to Challenge Negative Information
Applicants should be allowed the opportunity to challenge any potentially disqualifying information. The applicant should be sent any information reviewed (e.g., their screening report) along with the precise standards at issue. This information should be provided in a way that makes it easy for the applicant to understand why they were denied.
The applicant should then be given the opportunity to dispute the accuracy or completeness of any negative information, for example by demonstrating that a record belongs to another person with a similar name or omits a court decision in their favor. The applicant should also get the chance to show – even if a negative record is accurate – that they will comply with their tenancy obligations regardless. Applicants may do this by disputing the relevance of a standard to their particular circumstance (e.g., minimum income requirements for those with Housing Choice Vouchers). Applicants may also demonstrate that any negative behavior is unlikely to recur by providing evidence of mitigating circumstances (e.g., participation in a rehabilitation or financial literacy program, a change in education or employment status, or a positive reference from a social service provider or another person who knows the applicant).68
- Design and Test Complex Models for Fair Housing Compliance
The Fair Housing Act applies broadly, including to housing decisions made using machine learning and other forms of AI. Screeners that use complex models should choose types that are more interpretable, especially when doing so does not meaningfully reduce accuracy or fairness.69 Housing providers and tenant screening companies may find it difficult to be sure they are complying with the Act when the reasoning behind automated decisions is not transparent. If a highly complex model has a discriminatory effect, the model’s lack of transparency could make it hard to prove that a legally sufficient justification exists for the criteria used for a denial decision. Evaluating the data used by a model, creating test data, and rigorously testing to validate the model can reveal whether it has disparate outcomes and what less discriminatory alternatives might exist.70
Complex models, such as those that use machine learning or other forms of AI, should
68 Individualized assessments must be conducted in a nondiscriminatory way, including by treating similar mitigating circumstances the same way for all similarly situated applicants. See HUD 2022 Criminal Records Guidance at 10.
69 See, e.g., CONSUMER FIN. PROT. BUREAU, CIRCULAR 2022-03, ADVERSE ACTION NOTIFICATION REQUIREMENTS IN CONNECTION WITH CREDIT DECISIONS BASED ON
COMPLEX ALGORITHMS, 3 (2022), https://files.consumerfinance.gov/f/documents/cfpb_2022-03_circular_2022-05.pdf.
70 See generally Emily Black, John Logan Koepke, Pauline T. Kim, Pauline, Solon Barocas & Mingwei Hsu, Less Discriminatory Algorithms, 113 GEO. L. J. (forthcoming 2024) (manuscript at 1) [hereinafter, Less Discriminatory Algorithms].
be programmed following best practices for nondiscriminatory model design and with attention to aspects especially likely to pose fair housing concerns. For example, in designing the purpose of a model (also known as question or problem formulation), adding levels of precision can help avoid unjustified discriminatory effects.71
Training the model on demographically representative data can help ensure that the model does not erroneously learn that certain protected classes should be screened out at higher rates.72 Demographic representation is important both in datasets intended to represent “good” tenants and those intended to represent “bad” tenants. Problems of over- and under-representation can arise when models are trained on real-world data because certain data for specific demographic groups can be limited.73 These problems can also arise because real-world datasets are influenced by historic and ongoing discrimination, so datasets need to be selected and processed in a way that ensures the model does not perpetuate those effects.74 Problems of representation can be addressed by using statistical techniques to adjust real-world data, carefully constructing synthetic data, or by using other standard validation tools.75 In addition, models should be trained on data that includes all relevant features (e.g., all types of income).76 Ongoing monitoring for these issues is important to ensure that changes over time (e.g., demographic shifts) do not cause a dataset to become unrepresentative or incomplete.
Improving the predictiveness of a model can help ensure a model is achieving its purpose in a nondiscriminatory way. Careful consideration should be given to the variables a model uses, how they are weighted, and how they might interrelate to ensure the model has comparable levels of predictiveness and error rates across demographic groups. Models should be validated as having accurately and equitably predicted behaviors in the past and reassessed to ensure they continue to do so.77 Models that are not, in fact, predictive of future tenant behavior do not further a housing provider’s interest in avoiding renting to unproblematic tenants.
C. Additional Considerations for Three Particular Types of Screenings
The following are particular considerations for housing providers and tenant screening companies when conducting three types of screenings particularly likely to pose fair housing concerns: credit history, eviction history, and criminal records. Because of the particular disparities discussed below, overbroad screenings in these areas, including screenings that fail
71 In some circumstances, programming models to assess data in a continuous, rather than binary, way may yield more precise and less discriminatory outcomes. Less Discriminatory Algorithms at 37, 41.
72 Less Discriminatory Algorithms at 37–39.
73 See Lerman, Jonas, Big Data and its Exclusions, 66 STAN. L. REV. ONLINE 55, 57–58 (2013).
74 See, e.g., Less Discriminatory Algorithms at 37–39.
75 See, e.g., id. at 37–39.
76 Id. at 39–40.
77 Id. at 41.
to account for individual circumstances, are especially likely to have an unjustified discriminatory effect based on race, national origin, sex, disability, or another protected characteristic.
- Credit History
Credit scores are calculations made based on a consumer’s credit report designed to assess the relative risk of consumers defaulting on a loan (i.e., not the risk that a tenant will fail to pay rent).78 To construct this score, national credit bureaus assemble information about a consumer’s credit history for their credit file, place information in the consumer’s credit report, and then use that information to calculate a numerical score.79 The vast majority of housing providers and tenant screening companies run credit checks on applicants.80 Some tenant screening companies go a step further, incorporating credit information into models that obscure how the information is used.81
Black and Brown persons are more likely to have inaccurate credit reports or have had experiences that resulted in low or no credit scores.82 Residents of majority-Black areas are more than twice as likely to dispute items in their credit report as errors than residents of majority-White areas.83 As of August 2021, Native American, Black, and Hispanic individuals had median FICO credit scores of 612, 627, and 667 respectively, while White individuals had a median credit score of 727.84 In fact, more than half of White households have FICO credit scores above 700 compared to 21% of Black households.85 Additionally, individuals with disabilities are more likely to report they have a “bad” or a “very bad” credit score than individuals without disabilities.86
Racial disparities also exist among those who are credit invisible (i.e., they have minimal
78 CFPB TENANT BACKGROUND CHECKS MARKET at 38–39; CONSUMER FIN. PROT. BUREAU, KEY DIMENSIONS AND PROCESSES IN THE U.S. CREDIT REPORTING SYSTEM: A REVIEW OF HOW THE NATION’S LARGEST CREDIT BUREAUS MANAGE CONSUMER DATA 10 (2007), [hereinafter, CFPB KEY DIMENSIONS AND PROCESSES],
https://files.consumerfinance.gov/f/201212_cfpb_credit-reporting-white-paper.pdf.
79 CFPB KEY DIMENSIONS AND PROCESSES at 7, 9, 10.
80 See TransUnion Independent Landlord Survey Insights, TRANSUNION (last updated Aug. 07, 2017), https://www.mysmartmove.com/blog/landlord-rental-market-survey-insights-infographic; CFPB TENANT BACKGROUND CHECKS MARKET at 10–11, 38.
81 NAT’L CONSUMER L. CTR. DIGITAL DENIALS at 10.
82 CONSUMER FIN. PROT. BUREAU, DISPUTES ON CONSUMER CREDIT REPORTS 8–9 (2021) [hereinafter, CFPB DISPUTES ON CONSUMER CREDIT REPORTS],
https://files.consumerfinance.gov/f/documents/cfpb_disputes-on-consumer-credit-reports_report_2021-11.pdf; CONSUMER FIN. PROT. BUREAU, DATA POINT: CREDIT INVISIBLES 16–18 (2015), [hereinafter, CFPB DATA POINT: CREDIT INVISIBLES],
https://files.consumerfinance.gov/f/201505_cfpb_data-point-credit-invisibles.pdf.
83 CFPB DISPUTES ON CONSUMER CREDIT REPORTS at 8–9.
84 Laura Swanson et al, Credit Health during the COVID-19 Pandemic, URB. INST. (Mar. 8, 2022), https://apps.urban.org/features/credit-health- during-pandemic/. FICO scores are a common credit scoring methodology developed by the Fair Isaac Corporation used by most lenders. The score is calculated based on data compiled by a consumer reporting agency in an individual’s credit report, and used to predict how likely someone is to pay back a loan on time. What is a FICO Score?, CONSUMER FIN. PROT. BUREAU (Sept. 4, 2020), https://www.consumerfinance.gov/ask-cfpb/what-is-a-fico-score-en-1883/.
85 JUNG HYUN CHOI ET AL., URB. INST., EXPLAINING THE BLACK-WHITE HOMEOWNERSHIP GAP 8 (last updated Nov. 2019),
86 NANETTE GOODMAN, BONNIE O’DAY & MICHAEL MORRIS, NAT’L DISABILITY INST. & FINRA INV. EDU. FOUND., FINANCIAL CAPABILITY OF ADULTS WITH DISABILITIES:
FINDINGS FROM THE NATIONAL FINANCIAL CAPABILITY STUDY 47 (2017), https://www.finrafoundation.org/files/financial-capability-adults-disabilities.
or no credit history). 16% of Hispanic individuals and 15% percent of Black individuals are credit invisible, compared to 9% of non-Hispanic White individuals.87 Racial disparities in credit scoring exist, in part, because Black and Brown persons lack access to equitable credit and homeownership opportunities,88 and credit scores generally do not capture timely rent, utility, or other bill payments.89 Credit scores also do not capture timely payments to payday lenders and other non-traditional lenders,90 which are disproportionately used by Black borrowers.91 Those who are credit invisible, though often treated as having poor credit, may not have any history that suggests they are credit risks, let alone rental risks; they simply have no record one way or the other. For example, an applicant who recently immigrated to the United States might have a credit record that lacks information regardless of their financial history in their country of origin.92
Survivors of domestic violence, who are disproportionately women, are also more likely to have had experiences resulting in no or low credit scores.93 A survivor might not have credit records because their partner prohibited them from opening their own accounts, or the survivor might have negative credit history from being forced to obtain credit for the perpetrator’s benefit or assume the perpetrator’s unpaid bills.
Because of these disparities, overbroad screenings for credit history may have an unjustified discriminatory effect based on race or other protected characteristics. HUD is unaware of any studies showing that credit reports and scores accurately predict a successful tenancy, and as mentioned above they were not designed for this purpose. Many households prioritize paying the rent over other debts during times of financial hardship, yet their choice to do so — which should indicate they will continue to prioritize paying rent — is generally not
87 CFPB DATA POINT: CREDIT INVISIBLES at 16–18, 33.
88 NAT’L CONSUMER LAW CTR., PAST IMPERFECT: HOW CREDIT SCORES “BAKE IN” AND PERPETUATE PAST DISCRIMINATION 2 (2024), https://www.nclc.org/wp-
content/uploads/2016/05/20240227_Issue-Brief_Past-Imperfect.pdf.
89 See CFPB DATA POINT: CREDIT INVISIBLES at 5; CFPB TENANT BACKGROUND CHECKS MARKET at 38–39; Jim Akin, What Kinds of Bills Affect Credit Score?, EXPERIAN (Nov. 29, 2023), https://www.experian.com/blogs/ask-experian/what-kinds-of-bills-affect-credit-scores/; Brianna McGurran, How Utility Bills Can Boost Your Credit Score, EXPERIAN (Aug. 16, 2023), https://www.experian.com/blogs/ask-experian/does-paying-utility-bills-help- your-credit-score/. Even when rental debt is reported to credit bureaus, it often contains errors. See generally NAT’L CONSUMER LAW CTR., UNFAIR DEBTS WITH NO WAY OUT CONSUMERS SHARE THEIR EXPERIENCES WITH DEBT COLLECTORS (2022), https://www.nclc.org/wp-
content/uploads/2022/10/UnfairDebts-Rpt.pdf.
90 I Heard That Taking Out a Payday Loan Can Help Rebuild My Credit or Improve My Credit Score. Is This True?, CONSUMER FIN. PROT. BUREAU (last reviewed Sep. 1, 2020), https://www.consumerfinance.gov/ask-cfpb/i-heard-that-taking-out-a-payday-loan-can-help-rebuild-my-credit-or- improve-my-credit-score-is-this-true-en-1611/; Ana Staples, What is a Car Title Loan and How Does It Work?, CNBC (May 8, 2023), https://www.cnbc.com/select/car-title-loans/#:~:text=Most%20of%20the%20time%2C%20title,report%20them%20to%20credit%20bureaus. 91Mehrsa Baradaran, Jim Crow Credit, 9 U.C. IRVINE L. REV. 887, 938–39 (2019).
92 See DEP’T OF HOUS. & URB. DEV. OFF. OF FAIR HOUS. & EQUAL OPPORTUNITY, FREQUENTLY ASKED QUESTIONS (FAQS) ON FAIR HOUSING ISSUES REGARDING EXCEPTIONS TO CREDIT CHECK POLICIES AND OCCUPANCY LIMITS, AFFIRMATIVE MARKETING, AND LANGUAGE ACCESS 1–2, [hereinafter, FHEO FAQ ON CREDIT CHECK
POLICIES] https://www.hud.gov/sites/dfiles/FHEO/documents/General%20FAQ%20-%20Housing%20Providers%20and%20Fair%20Housing.pdf (last visited Apr. 26, 2024).
93 See DEP’T OF HOUS. & URB. DEV. OFF. OF PUB. & INDIAN HOUS., PIH-2017-08, VIOLENCE AGAINST WOMEN REAUTHORIZATION ACT OF 2013 GUIDANCE 7
(2017), https://www.hud.gov/sites/documents/PIH-2017-08VAWRA2013.PDF; see also LEAH A. PLUNKETT & ERIKA A. SUSSMAN, NAT’L CONSUMER RIGHTS CTR., CONSUMER RIGHTS SCREENING TOOL FOR DOMESTIC VIOLENCE ADVOCATES AND LAWYERS 4 (2011), https://csaj.org/wp-
considered in their favor in the credit history analysis.94
Given these significant and recognized limitations of credit scores as a predictor of likelihood to pay rent and given the disparities noted above, overreliance on credit history poses a significant risk of having an unjustified discriminatory effect based on race or other protected characteristics.95 Note that no HUD program requires screening rental applicants for their credit score, and previous HUD guidance has noted that housing providers may forgo credit checks as long as they do not discriminate because of a protected characteristic.96
Furthermore, housing providers and tenant screening companies should avoid denials and denial recommendations based on an applicant’s credit score in circumstances when the applicant’s financial background has especially little relevance. Limiting the use of credit scores when more relevant financial information is available may be a less discriminatory alternative to using credit scores in all instances. For example, a government agency or other entity guaranteeing a significant portion of an applicant’s income should make it significantly more likely that the applicant’s rent will be paid on time notwithstanding any negative credit history, particularly if that history predates the applicant’s receipt of such assistance. Housing providers and tenant screening companies should keep in mind that under the Housing Choice Voucher program and other assisted housing programs, a public housing agency or other entity has already deemed the rent affordable based on the applicant’s income and will increase the amount of assistance if the applicant’s income decreases.97
Other circumstances in which an applicant’s credit history may not be relevant include when the applicant has a cosigner who satisfies the housing provider’s financial screening criteria. Negative credit history due to an event that is unlikely to recur, such as a family or medical emergency, also has little relevance. If the household member responsible for paying the rent passes the housing provider’s financial screening criteria, the credit qualifications of other household members have no relevance. Minimal or poor credit history that was due to domestic violence, dating violence, sexual assault, or stalking is not the fault of the survivor and does not bear upon the likelihood that the survivor will pay their rent on time in the future.98
These circumstances can be accounted for by manually disregarding the applicant’s credit score when it is not relevant or programming an automated screening model to do so
94 See CFPB TENANT BACKGROUND CHECKS MARKET at 39.
95 See Louis v. SafeRent Sols., LLC, Civ. No. 22-CV-10800-AK, 2023 WL 4766192, at *10–*14 (D. Mass. July 26, 2023); Lumpkin v. Farmers Grp. Inc., No. 05–2868 Ma/V, 2007 WL 6996584, at *4 (W.D. Tenn. Apr. 26, 2007). Note this guidance applies to credit scores currently in existence that have not been proven predictive of tenancy behavior. It would not apply to alternative credit scores designed and proven predictive for such purpose.
96 FHEO FAQ ON CREDIT CHECK POLICIES at 1.
97 For similar reasons, minimum income requirements are also not relevant to predicting whether a tenant who receives income-based housing assistance will pay their rent on time.
98 In addition to the Fair Housing Act requirements discussed herein, when screening tenants under certain housing programs the Violence Against Women Act (“VAWA”) prohibits taking into account poor credit history that is due to an underlying experience of VAWA violence/abuse. 24 C.F.R. § 5.2005(b).
(e.g., not screening for credit if the applicant uses a Housing Choice Voucher).99 Less discriminatory alternatives may also include giving applicants the chance to explain why their credit history is not relevant by presenting mitigating circumstances (discussed in Sections V.B.5 above). Waiving or adjusting credit screening may also be required as a reasonable accommodation if an applicant’s poor credit history is related to their disability.100 For example, poor credit could be caused by a gap between a loss of employment income and receipt of disability benefits.
Having no credit history or limited credit history (i.e., an absence of credit history, negative or positive) is even less relevant than having poor credit history. Admitting applicants so long as they do not have a negative credit history can be a more precise policy than requiring a positive credit history. In addition, automated screening tools can be programmed to rely on other available financial information as a substitute for an absent credit score.
- Eviction History
Eviction records are one of the most commonly-marketed tenant screening components, and they are standard features of many screening reports.101 Tenant screening companies have built private databases from court records of eviction cases individuals have been involved in.102 The quality of the records in these databases varies, as does whether they speak only to the existence of a filing or include its ultimate resolution.103 Court records of evictions are notably unreliable: 22% of the eviction records evaluated in a large study contained ambiguous information on how the case was resolved or falsely represented a tenant’s eviction history.104
Eviction disproportionately affects Black and Hispanic renters, women, families with children, and individuals with disabilities. Even though fewer than one in five renters are Black, over half of all eviction cases are filed against Black renters.105 Black and Hispanic renters have eviction cases filed against them at higher rates than White renters, with women of those
99 See Louis v. SafeRent Sols., LLC., Case No.1:22-cv-10800-AK, Doc. 116-1 at 14 (Mar. 28, 2024) (approved by parties, awaiting court approval).
100 See, e.g., Schaw v. Habitat for Human. of Citrus Cnty., Inc., 938 F.3d 1259, 1267–69 (11th Cir. 2019) (requiring owner to consider support from a relative, among other financial resources, as a reasonable accommodation); Edwards v. Gene Salter Props. & Salter Constr. Inc., 739 F. App’x 357, 358 (8th Cir. 2018) (waiving policy requiring pay stubs, an offer letter, or tax returns as a reasonable accommodation); Giebeler v. M & B Assocs., 343 F.3d 1143, 1155 (9th Cir. 2003) (allowing tenant to add mother as a cosigner as a reasonable accommodation).
101 TENANT BACKGROUND CHECKS MARKET at 14–16, 29.
102 See, e.g. Tenant Eviction Checks, TRANSUNION, https://www.mysmartmove.com/tenant-screening-services/eviction-check (last visited Apr. 23, 2024).
103 See, Lesly Fair, Eviction Fiction? $15 Million FTC-CFPB Settlement with TransUnion and Tenant Screening Subsidiary Underscores Importance of FCRA’s “Maximum Possible Accuracy” Requirement, FED. TRADE COMM’N (Oct. 12, 2023), https://www.ftc.gov/business- guidance/blog/2023/10/eviction-fiction-15-million-ftc-cfpb-settlement-trans-union-tenant-screening-subsidiary-underscores.
104 Adam Porton, Ashley Gromis & Matthew Desmond, Inaccuracies in Eviction Records: Implications for Renters and Researchers, 31 HOUS. POL’Y DEBATE 377, 382 (2021) (studying over 3.6 million administrative eviction court records from 12 states).
105 Nick Gretz, Carl Gershenson, Peter Hepburn & Matthew Desmond, Who is Evicted in America, EVICTION LAB (Oct. 3, 2023), https://evictionlab.org/who-is-evicted-in-america/; see, e.g., also Hepburn, Renee Louis & Matthew Desmond, Racial and Gender Disparities Among Evicted Americans, 7 SOC. SCI. 649, 658 (2020).
groups bearing even higher disparities.106 Evictions are filed against adults living with children at more than twice the rate than for adults living without children.107 Some groups of renters with disabilities disproportionately report a high likelihood of being evicted in the next two months.108
Because of these disparities, overbroad screenings for eviction history may have an unjustified discriminatory effect. Therefore, tenant screening companies and housing providers should not rely on eviction records that are old, incomplete, irrelevant, or where a better measure of an applicant’s behavior is available. If this information about an eviction record is known before a screening, the record should not be used. Otherwise, applicants should get the chance afterwards to have the record disregarded and corrections made (as discussed in Sections V.A.2 and V.B.5 above). Note that under the Fair Credit Reporting Act, tenant screening companies must include existing disposition information for eviction records.109
Housing providers and tenant screening companies should not deny or recommend the denial of housing to applicants based on eviction proceedings where the tenant prevailed, a settlement was reached, or the matter was dropped. Likewise, most jurisdictions allow for “no fault” evictions, for example, where the housing provider removes their property from the rental market. Models that use eviction records, irrespective of complexity, should be programmed to distinguish between different types of evictions and varied outcomes.
An eviction that occurred long ago or under circumstances that are no longer relevant does not bear upon the applicant’s future performance as a tenant. For example, an eviction for non-payment of rent from a market-rate unit is not relevant to whether an applicant who has since begun receiving rental assistance (e.g. from a government agency) will pay rent on time.
In certain circumstances, it would be particularly problematic to hold a past eviction against a tenant, for example an eviction filed against a tenant in retaliation for asserting their rights (e.g., to remedy conditions violations). An eviction that is due to an underlying
106 Sophie Beiers, Sandra Park & Linda Morris, Clearing the Record: How Eviction Sealing Laws Can Advance Housing Access for Women of Color, ACLU (Jan. 10, 2020), https://www.aclu.org/news/racial-justice/clearing-the-record-how-eviction-sealing-laws-can-advance-housing-access-for- women-of-color. See also TIMOTHY A. THOMAS, OTT TOOMET, IAN KENNEDY & ALEX RAMILLER, THE STATE OF EVICTIONS: RESULTS FROM THE UNIVERSITY OF
WASHINGTON EVICTIONS PROJECT, § 4.5 (2020), https://evictionresearch.net/washington/index.html ; BRIAN J. MCCABE & EVA ROSEN, EVICTION IN WASHINGTON, DC: RACIAL AND GEOGRAPHIC DISPARITIES IN HOUSING INSTABILITY 17–18 (2020), https://georgetown.app.box.com/s/8cq4p8ap4nq 5xm75b5mct0nz5002z3ap; Matthew Desmond, Eviction and the Reproduction of Urban Poverty, 119 AM. J. SOC. 88, 91, 99 (2012); Peter Hepburn, Renee Louis & Matthew Desmond, Racial and Gender Disparities Among Evicted Americans, 7 SOC. SCI. 649, 659 (2020).
107 Nick Gretz, Carl Gershenson, Peter Hepburn & Matthew Desmond, Who is Evicted in America, EVICTION LAB (Oct. 3, 2023), https://evictionlab.org/who-is-evicted-in-america/.
108 Carli Friedman, Housing Insecurity of Medicaid Beneficiaries with Cognitive Disabilities During the COVID-19 Pandemic, 16 DISABILITY & HEALTH
- 1, 2 (2023); Jaboa Lake, Valerie Novack & Mia Ives-Rublee, Recognizing and Addressing Housing Insecurity for Disabled Renters, CTR. FOR AM. PROGRESS (May 27, 2021), https://www.americanprogress.org/article/recognizing-addressing-housing-insecurity-disabled- renters/#:~:text=High%20housing%20insecurity%20among%20renters%20with%20disabilities&text=Overall%2C%207%20million%20renters%2 0with,more%20likely%20to%20face%20eviction.
109 Advisory Opinion, Fair Credit Reporting; Background Screening, 89 Fed. Reg. 4171, 4174 (Jan. 23, 2024); see also Complaint ¶ 24, Fed. Trade Comm’n v. TransUnion Rental Screening Sols., Inc., Civ. No. 1:23-cv-2659 (D. Colo. filed Oct. 12, 2023).
experience of domestic violence, dating violence, sexual assault, or stalking (e.g., for related noise violations) should not be held against the survivor.110 In addition, reasonable accommodations to the screening policy may be necessary if the eviction was related to the applicant’s disability, for example, an eviction for late payment of rent because of the timing of an SSI or SSDI payment or a medical emergency.
- Criminal Records
Persons who have been involved with the criminal justice system are disproportionately individuals with disabilities and Black and Brown persons, and therefore overbroad criminal records screening policies are likely to have an unjustified discriminatory effect.111 “Research shows that these disparities cannot be simply attributed to certain groups committing more crimes and are better explained by biases in the criminal justice system.”112
As discussed at length in prior guidance from HUD, overbroad criminal records screenings are likely to have an unjustified discriminatory effect because of these disparities; the principles discussed in these prior documents apply to tenant screening companies, as well as to housing providers.113
Overbroad criminal records screenings include those that do not differentiate between offenses based on their nature, severity, or how long ago they occurred; those that consider records, such as arrest records, that did not result in a conviction; and those that do not provide an opportunity for the applicant to provide evidence of rehabilitation or other mitigating factors.114 As a best practice, tenant screening companies should ensure that screening reports, recommendations, grades, and algorithmic models differentiate between criminal records on these bases — such as by excluding records that are old or for offenses not directly relevant to tenancy — and housing providers should account for these considerations when formulating screening policies and reviewing applications.
When considering the criminal record of an individual with a disability, a reasonable
110 In addition to Fair Housing Act requirements, when screening tenants under certain housing programs VAWA prohibits taking into account evictions that were due to an underlying experience of VAWA violence/abuse. 24 C.F.R. § 5.2005(b).
111 See HUD 2016 Criminal Records Guidance at 2; HUD 2022 Criminal Records Guidance at 2; see also SECRETARY MARCIA L. FUDGE, ELIMINATING BARRIERS THAT MAY UNNECESSARILY PREVENT INDIVIDUALS WITH CRIMINAL HISTORIES FROM PARTICIPATING IN HUD PROGRAMS 1 (Apr. 12, 2022),
https://www.hud.gov/sites/dfiles/Main/documents/Memo_on_Criminal_Records.pdf; see also Reducing Barriers to HUD-Assisted Housing, 89 Fed. Reg. 25,333, 25,346 (proposed Apr. 10, 2024). The 2024 Notice of Proposed Rulemaking proposes a number of policy responses to these issues with respect to housing providers that participate in HUD programs. HUD will carefully consider all comments on the NPRM before finalizing a rule.
112 HUD 2022 Criminal Records Guidance at 2.
113 HUD 2016 Criminal Records Guidance at 2–7; HUD 2022 Criminal Records Guidance at 2–3.
114 For example, HUD deemed overbroad a criminal records screening policy that required the denial of all applicants and the eviction of all tenants who had any felony record, regardless of the nature or recency of the offense; under this policy, a tenant in good standing was evicted for a fifteen-year-old forgery conviction for which he served two years in prison. Charge of Discrimination ¶¶ 12–16, 20, 22–23, Sec’y of Dep’t of Hous. & Urb. Dev. ex rel. Loveless v. Wesley Apt. Homes, LLC, No. 17-AF-0046-FH-004 (H.U.D. O.H.A. filed Jan. 18, 2017).
accommodation may be required.115 This can include making exceptions to admissions policies and disregarding certain criminal records. Such an accommodation may be required, for example, if the individual’s disability makes it unlikely that they would reoffend (e.g., a record of assault by someone who has since developed a severe mobility impairment). An accommodation may also be required if the criminal activity was related to the individual’s disability and is unlikely to recur (e.g., criminal activity related to a mental health disability that has since been treated).116 In addition, criminal activity should not be taken into account if it was due to an underlying experience of domestic violence, dating violence, sexual assault, or stalking (e.g., property damage during a violent incident, or a victim being forced to write bad checks).
- Conclusion
Housing providers and tenant screening companies both have a role to play in ensuring that tenant screenings are transparent, accurate, and fair. This is the case even with the use of advanced technologies, such as machine learning and other forms of AI. Doing so not only promotes compliance with the Fair Housing Act, but also ensures that all applicants are given an equal opportunity to be evaluated on their own merit when seeking to fulfill a need as fundamental as housing.
115 See FED. TRADE COMM’N, CONSUMER FIN. PROT. BUREAU, DEP’T OF HOUS. & URB. DEV. & DEP’T OF JUSTICE, TENANT BACKGROUND CHECK AND YOUR RIGHTS 5
(2024), https://www.hud.gov/sites/dfiles/FHEO/documents/HUD_Tenant_Background_Checks_and_Your_Rights.pdf.
116 See, e.g., Simmons v. T.M. Assocs. Mgmt., Inc., 287 F. Supp. 3d 600, 604 (W.D. Va. 2018); see also, e.g., Hunt v. Aimco Props., L.P., 814 F.3d 1213, 1226–27 (11th Cir. 2016); Roe v. Sugar River Mills Assocs., 820 F. Supp. 636, 639–40 (D.N.H 1993); Roe v. Hous. Auth. of City of Boulder, 909 F. Supp. 814, 822 (D. Colo. 1995); Bos. Hous. Auth. v. Bridgewaters, 898 N.E. 2d 848, 850 (Mass. 2009); DEP’T OF HOUS. & URB. DEV. & DEP’T OF
JUSTICE, REASONABLE ACCOMMODATIONS UNDER THE FAIR HOUSING ACT 4–6 (2004), https://www.hud.gov/sites/documents/huddojstatement.pdf.