SHEPARD, Chief Justice.
A homeowner whose deed contained various covenants applicable to her subdivision rented out her residence, notwithstanding a covenant not to do so. Her homeowners association sued to enforce the prohibition, and she countersued, claiming that the agreement she had made through the covenant violated the Fair Housing Act. Her counter-complaint appeared to include elements of two very different claims — disparate impact and intentional discrimination.
The trial court granted her relief, appearing largely to rely on disparate impact. We conclude that relief on these grounds was erroneous. We remand for reconsideration of the intentional discrimination claims.
Facts and Procedural History
Algy and Edna McGlothin, husband and wife, purchased a home in Villas West II Planned Unit Development of Willowridge Subdivision on August 26, 1996. Villas West II is a 149-lot development located in the City of Kokomo.1
The MeGlothins purchased their home subject to “any and all easements, agreements and restrictions of record.” One of these provisions prohibited owners from leasing their residences:
Lease of Dwelling by Owner. For the purpose of maintaining the congenial and residential character of Villas West II and for the protection of the Owners with regard to financially responsible residents, lease of a Dwelling by an Owner, shall not be allowed. Each Dwelling shall be occupied by an Owner and their immediate family.
(Appellant’s App. at 43.)2
Mrs. McGlothin lived in the home until she broke her hip in 1998 and moved to a [1278]*1278nursing home. Mr. McGlothin lived in the home another five months until he also moved into the nursing home. Mr. McGlothin remained in the nursing home until his death in June 1999. After Mr. McGlothin’s death, the McGIothins’ daughter began leasing out the home.
On August 20, 2002, the Villas West II Homeowners Association notified the McGIothins’ daughter that Mrs. McGlothin was in violation of the no-lease covenant and demanded compliance. In response, Mrs. McGlothin’s lawyer acknowledged the covenant, but argued that the rent payments were necessary to maintain Mrs. McGlothin in the nursing home.3 He also stated that the no-lease provision could be invalid, alleging it had racially discriminatory roots.
Although not unsympathetic to Mrs. McGlothin’s situation, the Homeowners Association declined to acquiesce in her violation of the no-lease covenant, citing concerns “about its residents and the economic consequences the violation could have on the neighborhood and property values as a whole.” (Id. at 92.) The Homeowners Association demanded that the tenant vacate the premises to avoid further legal action and attached a draft complaint seeking an injunction against the renting of the premises, the eviction of the current tenant, attorney fees, and all other damages.
Unable to resolve the matter, the Homeowners Association filed the complaint on October 10, 2002. Mrs. McGlothin subsequently filed her answer, affirmative defense, and counterclaim alleging the Association’s enforcement of the no-lease covenant violated the Fair Housing Act. The Homeowners Association moved for summary judgment, which the trial court denied.
After a bench trial, the court concluded that the covenant violated the Fair Housing Act, finding it had a greater adverse effect on African Americans and racial minorities and finding “no legitimate nondiscriminatory reason” for the no-lease covenant. (Id. at 17-19.) The court entered judgment for Mrs. McGlothin. The Homeowners Association appealed, and the Court of Appeals affirmed. Villas West II of Willowridge, Homeowners Ass’n, Inc. v. McGlothin, 841 N.E.2d 584, 608 (Ind.Ct.App.2006). We grant transfer, reverse, and remand.
I. The Prevalence of Real Estate Covenants
A restrictive covenant is an express contract between grantor and grantee that restrains the grantee’s use of his land. Holliday v. Crooked Creek Villages Homeowners Ass’n, Inc., 759 N.E.2d 1088 (Ind.Ct.App.2001). Covenants control many aspects of land, including what may be built on the land (fence or above ground pool), how the land may be used (private or commercial), and alienability of the land. See, e.g., Robert G. Natelson, Law of Property Owners Associations 56-58, 153-66 (1989).
Restrictive covenants are used to maintain or enhance the value of land by reciprocal undertakings that restrain or regulate groups of properties. Holliday, 759 N.E.2d at 1092. These covenants are [1279]*1279common in condominium or other “common-interest” housing subdivisions. Prior to selling the first unit or plat, the subdivision or condominium owner creates a declaration or master deed that contains all of the restrictions.4 Property owners who purchase their properties subject to such restrictions give up a certain degree of individual freedom in exchange for the protections from living in a community of reciprocal undertakings.
Restrictions found in a declaration (like those found in a master deed) “are clothed with a very strong presumption of validity which arises from the fact that each individual unit owner purchases his unit knowing of and accepting the restrictions to be imposed.” Hidden Harbour Estates, Inc. v. Basso, 393 So.2d 637, 639 (Fla.Dist.Ct.App.1981). Analogizing restrictions in declarations to covenants running with the land, the Basso court held that restrictions in the declaration “will not be invalidated absent a showing that they are wholly arbitrary in their application, in violation of public policy, or that they abrogate some fundamental constitutional right.” Id. at 639-40. Today, as the Basso court anticipated, restrictive covenants function identically in planned subdivisions and condominiums and function identically regardless of whether they are found in a master deed or a declaration. Natelson, supra, at 58-60 (“the integration between the law of condominium and the law of other covenanted subdivisions is well on the way to being complete,” and thus, not proper “to speak or write of ‘condominium law’ and ‘homeowners association law’ as if they were discrete topics”).5
Condominium associations frequently adopt no-lease covenants that bar rental of units and forbid absentee ownership. Jordan I. Shifrin, No-Leasing Restrictions on Condominium Owners: The Legal Landscape, 94 Ill. B.J. 80, 80-81 (2006). These covenants are adopted to address owners’ concerns “about the negative effects of high resident turnover and renters’ perceived lack of attention to the property.” Id. at 80. Some empirical data validates these concerns; a California study showed that “[a] high number of leased units (over 30%) can impair significantly the market position of the subdivision.” Robert G. Natelson, Consent, Coercion, and “Reasonableness” in Private Law: The Special Case of the Property Owners Association, 51 Ohio St. L.J. 41, 73 n. 150 (1990). Courts commonly enforce no-lease restrictive covenants.6 See, e.g., Flagler Fed. Sav. & Loan Ass’n v. Crestview Towers Condo. Ass’n, Inc., 595 So.2d 198 (Fla.Dist.Ct.App.1992) (association may prohibit leasing; restrictions in declaration presumed valid); Seagate Condo. Ass’n, Inc. v. Duffy, 330 So.2d 484 (Fla.Dist.Ct.App.1976) (leasing restrictions not unreasonable restraints on alienation; [1280]*1280restrictions promote residential character of community); Apple II Condo. Ass’n v. Worth Bank & Trust Co., 277 Ill.App.3d 345, 213 Ill.Dec. 463, 659 N.E.2d 93, 99 (1995) (leasing prohibition made part of declaration presumed valid and upheld unless restriction “arbitrary, against public policy or violates some fundamental constitutional right of the unit owners”).
II. Disparate Impact Claims Under the Fair Housing Act
Still, agreements to limit the use of real property are subject to challenge under other law. The challenge here is based on the Fair Housing Act (FHA), enacted as Title VIII of the Civil Rights Act of 1968. FHA claims can take either of two routes: disparate treatment or disparate impact.
Disparate treatment claims require proof of intentionally discriminatory treatment of a protected class. Disparate impact claims, by contrast, require no proof of intent, and can be established if a policy or practice has a discriminatory effect on a protected class, even if the policy or practice is facially nondiscriminatory. Disparate impact recovery was first allowed in employment discrimination eases under Title VII of the Civil Rights Act of 1964. See Griggs v. Duke Power Co., 401 U.S. 424, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971). Title VII makes- it illegal for an employer “to fail or refuse to hire or to discharge” a person “because of’ a prohibited reason such as race. 42 U.S.C.A. § 2000e-2(a)(l) (West 2008). Title VIII uses the same critical language, making it unlawful to “refuse to sell or rent ... a dwelling to any person because of’ race. 42 U.S.C.A. § 3604. Based on this identical language, disparate impact recovery under the FHA has been allowed by all federal circuit courts that have addressed the question.7
There is wide agreement in the federal circuit courts that the FHA allows disparate impact claims, but no consensus about the proper framework for analyzing such a claim.8 The U.S. Supreme Court has not addressed the issue, and circuits have developed a variety of approaches, for the most part derived from three early FHA disparate impact cases, Metropolitan Housing Development Corp. v. Village of Arlington Heights, 558 F.2d 1283 (7th Cir.1977) (Arlington Heights II),9 Resident Advisory Board v. Rizzo, 564 F.2d 126 (3d Cir.1977), and Huntington Branch, NAACP v. Town of Huntington, 844 F.2d 926 (2d Cir.1988).
In Arlington Heights II, the Seventh Circuit held that disparate impact recovery under the FHA is proper when the defendant’s conduct produces a discriminatory [1281]*1281effect and four factors balance in favor of granting relief. The four factors are (1) the strength of the plaintiffs showing of discriminatory effect; (2) evidence of discriminatory intent, though not enough to satisfy the constitutional standard of Washington v. Davis; (3) the defendant’s interest in the challenged conduct; and (4) whether the plaintiff seeks affirmative relief or merely to restrain the defendant from interfering with individual property owners who wish to provide housing. Arlington Heights II, 558 F.2d at 1290.
One month after Arlington Heights II, the Third Circuit adopted a burden-shifting framework similar to that already in use in Title VII employment cases. Rizzo, 564 F.2d at 148. Under this approach, the plaintiff makes a prima facie case by showing that the defendant’s action has a discriminatory effect. Id. The defendant can rebut this by showing a justification which “serve[s], in theory and in practice, a legitimate, bona fide interest” and by showing that “no alternative course of action could be adopted that would enable that interest to be served with less discriminatory impact.” Id. at 149. The Rizzo court did not apply the Arlington Heights II factors but acknowledged that the result would have been the same. Id. at 148 n. 32.
A decade later, the Huntington court sought to merge Rizzo and Arlington Heights II. Under the Huntington framework, the plaintiff establishes a prima facie case by showing that the “challenged practice of the defendant actually or predictably results in racial discrimination.” Huntington, 844 F.2d at 934. The defendant can rebut this case by showing that its “actions furthered, in theory and in practice, a legitimate, bona fide governmental interest and that no alternative would serve that interest with less discriminatory effect.” The court then balances the four Arlington Heights II factors to determine whether to grant relief. Id. at 936.
These competing approaches spawned a variety of others. One source of disagreement was the role, if any, of the Arlington Heights II factors. A few courts have used the factors as part of the prima facie case. See, e.g., Smith v. Town of Clarkton, 682 F.2d 1055, 1065-66 (4th Cir.1982); Snyder v. Barry Realty, Inc., 953 F.Supp. 217, 220 (N.D.Ill.1996): This approach has been widely rejected as making the plaintiffs case too difficult. E.g., Rizzo, 564 F.2d at 148 n. 32; Huntington, 844 F.2d at 935; Hispanics United v. Vill. of Addison, 988 F.Supp. 1130, 1154 n. 14 (N.D.Ill.1997). Other courts applied the Arlington Heights II. factors only in cases involving public defendants. E.g., Betsey v. Turtle Creek, Assocs., 736 F.2d 983, 989 n. 5 (4th Cir.1984). Still others discarded the second Arlington Heights II factor — proof of discriminatory intent — and weighed the remaining three, reasoning that intent was relevant only in disparate treatment cases. E.g., Mountain Side Mobile Estates P’ship v. Sec’y of Hous. & Urban Dev., 56 F.3d 1243, 1252 (10th Cir.1995); Arthur v. City of Toledo, 782 F.2d 565, 575 (6th Cir.1986).
More recently, most federal circuits have abandoned the Arlington Heights II factors altogether. E.g., Charleston Hous. Auth. v. U.S. Dept. of Agric., 419 F.3d 729 (8th Cir.2005); Tsombanidis v. West Haven Fire Dept., 352 F.3d 565 (2d Cir.2003); Lapid-Laurel, L.L.C. v. Zoning Bd. of Adjustment, 284 F.3d 442 (3d Cir.2002); Langlois v. Abington Hous. Auth., 207 F.3d 43, 51 (1st Cir.2000) (“True, one circuit court decision did refer to balancing, but the few later circuit court decisions on point come closer to a simple justification test, and we think this is by far the better approach.”) (citations omitted); Salute v. Stratford Greens Garden Apartments, 136 F.3d 293 (2d Cir.1998). Yet, at least one [1282]*1282circuit and several district courts have continued to balance. Reinhart v. Lincoln County, 482 F.3d 1225, 1229 (10th Cir.2007) (balancing three factors); Thompson v. U.S. Dept. of Hous. & Urban Dev., 348 F.Supp.2d 398, 417 (D.Md.2005) (balancing four factors); Peoria Area Landlord Ass’n v. City of Peoria, 168 F.Supp.2d 917, 921-22 (C.D.Ill.2001) (balancing four factors); Corp. of Episcopal Church v. W. Valley City, 119 F.Supp.2d 1215, 1219 (D.Utah 2000) (balancing four factors).
Huntington’s burden-shifting framework has also morphed. Several courts changed Huntington’s two burdens into three. See, e.g., Darst-Webbe Tenant Ass’n Bd. v. St. Louis Hous. Auth., 417 F.3d 898, 902-03 (8th Cir.2005); Harris v. Itzhaki 183 F.3d 1043, 1051 (9th Cir.1999); Mountain Side, 56 F.3d at 1254; Hispanics United, 988 F.Supp. at 1162. This test still requires plaintiffs to show a discriminatory effect, but rebutting defendants need to provide only a bona fide, nondiscriminatory justification. Once the defendant offers an appropriate justification, the burden shifts back to the plaintiff to show an alternative that would serve the defendant’s interest with less discriminatory effect. Courts adopting this approach explain that requiring the plaintiff to propose an alternative gives the plaintiff the ultimate burden of proving a violation and has the advantage that “neither party is saddled with having to prove a negative (the nonexistence of bona fide reasons or the absence of less discriminatory alternatives).” Hispanics United, 988 F.Supp. at 1162.
Finally, a much smaller number of federal courts have applied a variant of the disparate treatment test from McDonnell Douglas Corp. v. Green, 411 U.S. 792, 93 S.Ct. 1817, 36 L.Ed.2d 668 (1973), a leading employment discrimination case,, to resolve disparate impact FHA claims. Under this approach, the plaintiff must prove a discriminatory effect, which the defendant may rebut by offering a legitimate, nondiscriminatory reason for its rule. Instead of the plaintiff then showing a less discriminatory alternative, the plaintiff shows the defendant’s offered reason is a pretext. United States v. Badgett, 976 F.2d 1176, 1178 (8th Cir.1992); Snyder, 953 F.Supp. at 220.
Federal district courts in the Seventh Circuit are of course obligated to follow Seventh Circuit precedent, including Arlington Heights II. We are not so restricted, and it appears that Snyder, a district court opinion on which Arlington Heights II relied, is out of the mainstream of federal authority.
Moreover, Arlington Heights II seems doctrinally unsound. In employment disparate impact cases, burden-shifting tests are favored over factor-balancing tests. See Connecticut v. Teal, 457 U.S. 440, 446-47, 102 S.Ct. 2525, 73 L.Ed.2d 130 (1982); Dothard v. Rawlinson, 433 U.S. 321, 329, 97 S.Ct. 2720, 53 L.Ed.2d 786 (1977); Albemarle Paper Co. v. Moody, 422 U.S. 405, 425, 95 S.Ct. 2362, 45 L.Ed.2d 280 (1975). Because Title VII and the FHA use the same language in prohibiting discrimination, we should apply the same framework to both. Moreover, mixing the Arlington Heights II factors with the burden-shifting framework produces an unnecessarily complex process and introduces improper considerations. The burden-shifting framework accommodates the two relevant Arlington Heights II factors by considering the strength of the discriminatory effect of the challenged practice and the defendant’s interest necessitating it. The remaining two factors — form of relief and intent — are improper considerations in the disparate impact context.
[1283]*1283The relief sought is an inappropriate consideration because the FHA specifically authorizes courts to award “preventive relief, including a permanent or temporary injunction, restraining order, or other order.” It sets out no higher standard for plaintiffs seeking affirmative relief. 42 U.S.C.A. § 3614(d)(1)(A).
Intent is an improper consideration because in disparate impact cases the disparate impact itself is the violation, irrespective of intent. See Mountain Side, 56 F.3d at 1252. For the same reason, the third burden should address less discriminatory alternatives, not pretext. Pretext is a means of finding subjective intent to discriminate. The more pershasive federal courts have specifically held that current FHA disparate impact law does not involve pretext. See Huntington, 844 F.2d at 939 (“The McDonnell Douglas test, however, is an intent-based standard for disparate treatment cases inapposite to the disparate impact claim asserted here. No circuit, in an impact case, has required plaintiffs to prove that defendants’ justifications were pretextual.”).
These observations lead us to reject the Arlington Heights II approach and, with it, any search for pretext in a disparate impact case. Instead, we adopt the prevailing test, which involves two shifts, not one. Requiring the plaintiff to identify a specific less restrictive alternative is more efficient and fair than requiring the defendant to “guess at and eliminate” all possible alternatives. Hispanics United, 988 F.Supp. at 1162.
In sum, to establish a right to disparate impact recovery under the FHA, a plaintiff must establish a prima facie case by demonstrating that a policy or practice actually or predictably has a significantly adverse or disproportionate impact on a protected class., To rebut this showing, the defendant must demonstrate that its policy or practice has a manifest relationship to a legitimate, nondiscriminatory interest. The plaintiff may then overcome the defendant’s showing by demonstrating that a less discriminatory alternative would serve the defendant’s legitimate interest equally well.
We now turn to application of this disparate impact framework to the record in this case. The no-lease covenant at issue provides:
Lease of Dwelling by Owner. For the purpose of maintaining the congenial and residential character of Villas West II and for the protection of the Owners with regard to financially responsible residents, lease of a Dwelling by an Owner, shall not be allowed. Each Dwelling shall be occupied by an Owner and their immediate family.
(App. at 43.) McGlothin contends that the no-lease covenant violates the FHA because it has an impermissible disparate impact on African Americans. The trial court so found. (Id. at 20.) Evidence presented at trial showed that regardless of income and age, African Americans rent their homes in greater proportion than do whites. (Id. at 16-17.) The decrease in available rental housing caused by the no-lease covenant will predictably and disproportionately affect African Americans. While the evidence supporting this finding leaves something to be desired, we will proceed on the basis that the prima facie ease is established.
The Homeowners Association sought to rebut this prima facie case by demonstrating that its no-lease provision has a manifest relationship to a legitimate, nondiscriminatory interest. The trial court found that the Association’s reason for excluding renters from the subdivision is that “renters do not maintain homes which they rent as well as owners maintain [1284]*1284their homes. Therefore, the exclusion of renters helps maintain property values.” (Id. at 18.) The record Contains ample expert testimony supporting this proposition. (Tr. at 65-67, 230.)
Because the Association asserted a legitimate, nondiseriminatory reason for its no-lease covenant, the burden returned to McGlothin to propose an equally effective, less discriminatory alternative. McGlothin drew the trial court’s attention to several other covenants which the trial court found “more than adequately assured a neat, clean and visually attractive environment, and a high degree of property maintenance.” (App. at 18.) These covenants require homeowners to, among other things, maintain windows, door hardware, patios, and appliances; water lawns and shrubs; keep the exterior free of trash, certain signs, certain communication devises, and certain vehicles;- and “promptly perform all maintenance and repair ... which, if neglected, might adversely affect any other Dwelling, Common Area or the value of the Property.” (Id. at 33-35, 38-42.) Because these additional covenants were in place, the trial court found that “the Plaintiffs justification for the no-rent provision lacks a factual basis, and is mere subterfuge, rendering said provision unnecessary and useless.” (Id. at 18-19.) Ultimately, the trial court found the covenant to violate the FHA.
This record does not support a claim under a disparate impact theory. Although the specific property-maintenance covenants are a less discriminatory alternative to promote maintenance, the covenants are not an equally effective means of maintaining property values. Maintaining property values involves not merely maintaining property but also improving and updating it. Under these covenants, owners are obligated only to maintain the home, not to improve or update it. Owners who occupy their property have an incentive to improve and update because they can both enjoy the improvements and reap the fruits of their labor upon selling the home.
In a rental situation, however, this incentive is weakened because of divided ownership and occupancy. The renter lacks some incentive to improve the property because he or she would only benefit from the current enjoyment, not from an increased market value. Because the additional property-maintenance covenants do not address the problem of divided interests in rental property, they are not an equally effective, less discriminatory alternative to excluding renters. Moreover, although there is nothing in this record directly addressing the point, it seems obvious that an owner-occupant is both psychologically and financially invested in the property to a greater extent than a renter.10 Personal motivation can surely achieve better results than contractual compulsion in many cases. This is not a matter of reweighing evidence. Both parties’ experts testified that owners maintain property better than do renters. (Tr. at 66, 86, 230-31.) Because the evidence on this point is undisputed, to the extent the trial court found these other covenants equally adequate to maintain property values, the trial court findings on this point are clearly erroneous.
Because plaintiff points to no equally effective, less discriminatory alternatives to the defendant’s legitimate nondiscriminatory policy, that policy does not support a disparate impact FHA violation, even if [1285]*1285it has a disparate impact on a protected class.
III. Disparate Treatment
McGlothin’s counterclaim contains certain allegations that are part of a disparate treatment claim under the FHA. In her counterclaim, McGlothin alleged that
20. In making said covenant ... the developer evidenced an intention to make a preference, limitation, or discrimination among persons who could occupy dwellings within the subdivision based on race, color, sex, familial status, or national origin.
21. In seeking to enforce said covenant provision, members of Villas West II or Willowridge Homeowners Association, Inc. evidence an intention to make a preference, limitation, or discrimination among persons who could occupy dwellings within the subdivision based on race, color, sex, familial status, or national origin.
(App. at 55-56.) The trial court’s findings (and the trial record on intentional discrimination) reflect certain contradictions. The court found that there was no conclusive evidence of discriminatory intention behind the Homeowners Association’s attempt to enforce the no-lease covenant. At the same time, it found the no-lease covenant to be a “subterfuge,” suggesting pretext. (Id at 19.) This finding as to the underlying motivation for the covenant flows from a McGlothin witness who said the term “restricted” (used here in a developer’s advertisement) is understood by minorities, especially African Americans, to be synonymous with “segregated.” (Tr. at 304.) Two African American residents of Villas West II also testified on this point. One African American resident testified that she did not think the covenant language indicated a racial preference or exclusion. Another African American resident testified that the language in the covenant and the developer’s advertisements did not convey any racial discrimination and that she welcomed the covenant because renting had an adverse effect on property values. (Id. at 319-21, 324-27.)
The trial court’s findings about the developer’s intent are likewise ambiguous: “Whether the builder of the tract realized it or not, the words ‘restricted,’ as used in its advertisements clearly sent a message to the African American community ... that African Americans (and perhaps other minorities) are not welcome.” (App. at 19-20.) There is no evidence that the Association ever promoted the subdivision as “restricted,” but the no-lease covenant and advertisement did originate with the developer, who recorded the covenants and first controlled the Association. (Tr. at 15-16.) The record does not contain evidence regarding when advertisements using the word “restricted” were disseminated or whether those advertisements were linked to the Villas West II subdivision.
The nature of these findings and the fact that McGlothin’s brief contains arguments mostly but not exclusively pertinent to disparate impact understandably led the Court of Appeals to conclude that “[o]nly disparate impact is at issue here.” Villas West, 841 N.E.2d at 599. We find ourselves unable to discern whether relief is appropriate on McGlothin’s intentional discrimination claim. The need for a fair adjudication suggests the desirability of remanding for further evidence and findings.
Conclusion
We reverse the judgment of the trial court on the claim of disparate impact and remand for reconsideration of the claim of intentional discrimination.
DICKSON and BOEHM, JJ., concur.
RUCKER, J., dissents with separate opinion in which SULLIVAN, J., concurs.