Associated General Contractors of Ohio, Inc. v. Drabik

OPINION BOGGS, Circuit Judge. Associated General Contractors of Ohio, and Associated General Contractors of Northwest Ohio (“Plaintiffs-Appellees”), representing Ohio building contractors, sued to stop the award of a construction contract for the Toledo Correctional Facility to a minority-owned business (“MBE”), in a bidding process from which non-minority-owned firms were statutorily excluded under Ohio’s Minority Business Enterprise Act (“MBEA”). Plaintiffs-Ap-pellees claimed the MBEA is unconstitutional, in that it violates the Fourteenth Amendment’s Equal Protection Clause. The district court agreed, and permanently enjoined the state from awarding any construction contracts thereunder. Defendant-Appellant Sandra Drabik, Director of the Ohio Department of Administrative Services (“DAS”), which coordinates and manages state construction projects, and other Defendants-Appellants, appeal the district court’s order. We affirm. I Ohio passed the Minority Business Enterprise Act (“MBEA”) in 1980. This legislation set aside five percent, by value, of all state construction projects for bidding by certified MBEs exclusively. O.R.C. § 123.151(C)(1). Other provisions govern subcontracting to MBEs of work awarded under this scheme. Ohio defines an MBE as a venture owned and controlled, to the extent of fifty-one percent, for at least one year previous, by “members of one of the following economically disadvantaged groups: Blacks, American Indians, His-panies, and Orientals.” O.R.C. § 122.71(E). Other provisions establish procedures for certification and listing as an MBE; in what follows, “MBE” will be understood to refer to such officially certified businesses. As of October 1998, DAS maintained a list of 1,180 MBEs. Pursuant to the MBEA, DAS decided to set aside, for MBEs only, bidding for construction of the Toledo Correctional Facility’s Administration Building, which represents twenty percent of the total project’s value of $50 million. Non-MBEs, many of whom are members of the plaintiff trade associations, will thus be excluded on racial grounds from bidding on that aspect of the project, and will be restricted in their participation as subcontractors. MBEs are, of course, free to bid on, and participate fully in, non-set-aside as well as set-aside contracts. This court ruled in 1983 that the MBEA was constitutional, see Ohio Contractors Ass’n v. Keip, 713 F.2d 167 (6th Cir.1983), overruling Judge Kinneary’s judgment in the district court that Ohio’s scheme was unconstitutional on its face, see Ohio Contractors Ass’n v. Keip, No. C-2-82-446 (S.D.Oh. Dec. 15, 1982). Subsequently, the Supreme Court, in two landmark decisions, explained and applied at length the criteria of strict scrutiny under which such racially preferential set-asides were to be evaluated. See City of Richmond v. J.A. Croson Co., 488 U.S. 469, 109 S.Ct. 706, 102 L.Ed.2d 854 (1989); Adarand Constructors, Inc. v. Pena, 515 U.S. 200, 115 S.Ct. 2097, 132 L.Ed.2d 158 (1995). This court had already, in Michigan Road Builders Ass’n v. Milliken, 834 F.2d 583 (6th Cir.1987), taken note of the trend developing both in the Supreme Court and Circuit Courts (which was to culminate in Croson and Adarand) to apply the Equal Protection Clause strictly to racial discrimination in government contracting. Michigan Road Builders departed from the more relaxed treatment that Keip had accorded to equal protection challenges to state con- *734 trading disputes. See id. at 598 (Lively, C.J., dissenting). Croson also noted that same evolution in this Circuit. See 488 U.S. at 477, 109 S.Ct. 706. Ohio’s MBEA was passed after many years, during the 1970s, of executive and administrative agency task force consideration of complaints regarding, and statistics concerning, minority group participation in state construction contracts. These, and the legislative hearings and debates that immediately preceded passage of the MBEA, are detailed by Judge Kinneary in his 1982 decision striking down that act. In the light of Croson and Adarand, the district court in this case returned to the prescient standards under which the MBEA had been invalidated in. 1982. Reviewing the evidence, Judge Graham, at the close of a hearing held on October 26, 1998 to consider the state’s request for a six-month continuance (which was denied), cites Judge Kinneary’s 1982 analysis, which ' anticipated that of Croson, with complete approval. Judge Graham found the MBEA patently unconstitutional: “I am mindful of the fact that it is certainly unusual for a court to declaré a state statute which has such far-reaching effects unconstitutional from the bench, but I cannot imagine any clearer case than this for the unconstitutionality of the state' statute.” • ■ The district court also reférred to a 1997 ruling from the Ohio Court of Appeals on the MBEA. In that case, a business own! er of Lebanese descent, who was denied certification as an MBE on the grounds he was not an Oriental, claimed his right to equal protection was violated by the MBEA as applied to him. The Court of Appeals affirmed the trial court’s ruling that the MBEA’s per-se race classification is unconstitutional. That ruling has since been overturned by the Ohio Supreme Court, which, in a lengthy review of minority set-aside jurisprudence, concluded that the MBEA was constitutional. See Ritchey Produce Co. v. State of Ohio Dep’t of Administrative Services, 1997 WL 629965 (Ohio App. 10 Dist. Oct.7, 1997), rev’d, 85 Ohio St.3d 194, 707 N.E.2d 871 (1999). At the time of the district court’s ruling, Rit-chey was still pending in the Ohio Supreme Court. Hence, Ohio argued that the district court should have abstained from making a decision. A motion to that effect was filed with the district court, which denied it after an extensive consideration of abstention doctrine. This appeal also argues that the district court’s denial of the abstention motion was error. II A “The constitutionality of a statute is a question of law, reviewable de novo.” Hadix v. Johnson, 144 F.3d 925, 938 (6th Cir.1998) (citing United States v. Brown, 25 F.3d 307, 308 (6th Cir.), cert. denied, 513 U.S. 1045, 115 S.Ct. 640, 130 L.Ed.2d 546 (1994)). Croson reaffirmed the “strict scrutiny” standard of review adopted by the Court for preferential programs based on racial or ethnic criteria in Wygant v. Jackson Bd. of Educ., 476 U.S. 267, 274, 106 S.Ct. 1842, 90 L.Ed.2d 260 (1986). See Croson, 488 U.S. at 494, 109 S.Ct. 706. This requires that such a program be “narrowly tailored” .to satisfy a “compelling governmental-interest.” 476 U.S. at 274, 106 S.Ct. 1842 (quoting Fullilove v. Klutznick, 448 U.S. 448, 480, 100 S.Ct. 2758, 65 L.Ed.2d 902 (1980) (“narrowly tailored”); Palmore v. Sidoti, 466 U.S. 429, 432, 104 S.Ct. 1879, 80 L.Ed.2d 421 (1984) (“compelling governmental interest”)). Ada-rand reiterated this standard for “all racial classifications imposed by whatever federal, state, or local governmental actor....” See 515 U.S. at 227, 115 S.Ct. 2097. In discussing this issue, it is important to identify precisely the compelling state interest that might be able to overcome the general presumption against racial- classification. It is clear that a government “has a compelling interest in as *735 suring that public dollars ... do not serve to finance the evil of private prejudice.” Croson, 488 U.S. at 492, 109 S.Ct. 706. However, statistical disparity in the proportion of contracts awarded to a particular group, standing alone, does not demonstrate such an evil. It is an unfortunate aspect of reality that there is never more than 100% of anything; thus, raising the percentage allocated to some portion of the total population necessarily means a corresponding reduction in what is available to other portions. There is no question that remedying the effects of past discrimination constitutes a compelling governmental interest. See Croson, 488 U.S. at 508, 109 S.Ct. 706; United Black Firefighters Ass’n v. City of Akron, 976 F.2d 999, 1010-11 (6th Cir.1992). However, to make this showing, a state cannot rely on mere speculation, or legislative pronouncements, of past discrimination. Rather, the Supreme Court has told us that the state bears the burden of demonstrating a “strong basis in evidence for its conclusion that remedial action was necessary” by proving either that the state itself discriminated in the past or was a passive participant in private industry’s discriminatory practices. Croson, 488 U.S. at 486-92, 500, 109 S.Ct. 706. Thus, the linchpin of the Croson analysis, for present purposes, is not simply its mandating of strict scrutiny, the requirement that a program be narrowly tailored to achieve a compelling government interest, but above all its holding that governments must “identify discrimination with some specificity before they may use race-conscious relief;” explicit “findings of a constitutional or statutory violation must be made.” 488 U.S. at 497, 109 S.Ct. 706. In ruling against the State of Ohio in 1982, Judge Kinneary had held that the evidence presented by state studies from the middle and latter 1970s was inadequate to support a conclusion of specific historical discrimination calling for remediation, such as might justify the MBEA. More recently, this court has ruled that seventeen-year old evidence of discrimination is “too remote to support a finding of compelling government interest to justify the affirmative action plan,” and struck down a continuing affirmative action program for female firefighters on the ground that outdated evidence does not reflect “prior unremedied or current discrimination.” Brunet v. City of Columbus, 1 F.3d 390, 409 (6th Cir.1993), cert, denied, 510 U.S. 1164, 114 S.Ct. 1190, 127 L.Ed.2d 540 (1994). The MBEA suffers from the same defect. Moreover, Judge Kinneary anticipated Croson’s insistence on explicit findings in the following observation: In all the documentary evidence relating to the progress of [the MBEA] through the legislature, including drafts of bills, Legislative Service Commission summaries, and transcripts of floor debate, there is not one clear, unambiguous statement of a finding of discrimination to be found. When, in 1983, this court overturned Judge Kineary’s ruling in that ease, we did so out of deference to the legislature, giving it the benefit of the doubt that implicit factfinding of discrimination underlay the MBEA. See Keip, 713 F.2d at 170-71. But the Supreme Court has since required more. See Miller v. Johnson, 515 U.S. 900, 923, 115 S.Ct. 2475, 132 L.Ed.2d 762 (1995) (holding legislation adopting racial distinctions to be entitled to no deference); Croson, 488 U.S. at 499, 109 S.Ct. 706 (holding mere assertions of legislative purposes insufficient). Proponents of racially discriminatory systems such as the MBEA have sought to generate the necessary evidence by a variety of means. See, e.g., George Stephanopoulos & Christopher Edley, Jr., Affirmative Action Review: Report to the President (July 19, 1995), repr. in BNA Daily Labor Report, 139 DLR S-l, 1995. However, such efforts have generally focused on mere “underrepresentation” — a lesser percentage of contracts awarded to a particular group than that group’s percentage in the general population. See, *736 e.g., id. at § 9.1.2 (reporting that, in 1986, “minority business received only ... 2.7 percent of the prime contract dollar” and characterizing that situation, without further analysis, as “discrimination”); Cro-son, 488 U.S. at 479-80, 109 S.Ct. 706 (noting that Richmond’s set-aside scheme relied on findings that only .67% of prime city construction contracts had been awarded to minority firms, in a city with a 50% African-American population). Raw statistical disparity of this sort is part of the evidence offered by Ohio in this case. See JA IV at 45 (Defendant’s Memorandum in Opposition to Plaintiffs Motion for a Preliminary Injunction). But such evidence of mere statistical disparities has been firmly rejected as insufficient by the Supreme Court, particularly in a context such as contracting, where special qualifications are so relevant. See Croson, 488 U-.S. at 501-02, 109 S.Ct. 706. And although Ohio’s most “compelling” statistical evidence compares the percentage of contracts awarded to minorities to the percentage of minority-owned businesses in Ohio — thus marshaling stronger statistics than the statistics in Croson — it is still insufficient. The problem with Ohio’s statistical comparison is that the percentage of minority-owned businesses in Ohio (7% as of 1978) did not take into account how many of those businesses were construction companies of any sort, let alone how many were qualified, willing, and able to perform state construction contracts. The statistical evidence that the -Ohio legislature had before it, when the, MBEA was enacted, consisted of four broad categories of data. The first was statistical evidence gathered by DAS for the years 1957 to 1979. This showed that only 0.21 percent of all state construction contracts went to “identifiable minority businesses.” Brief of Defendants-Appellants at 14. The second was a DAS study, cited in Keif, 713 F.2d at 171, showing that from 1959 to 1975, of the $1.14 billion paid out by the state in general construction contracts, only 0.24% went to minority businesses. Id. at 15. The third was a 1977 report, issued by the Ohio Legislative Budget Office, detailing Ohio Department of Transportation (“ODOT”) construction contracts garnered by minorities, showing figures of 0.13, 0.3, and 0.18 percent for the years 1975, 1976, and 1977, respectively. Ibid. Fourth, a 1978 task force established by the Ohio Attorney General to study the problem concluded that during 1975-77 minority businesses comprised seven percent of all Ohio businesses, but minority businesses received only 0.5 percent of ODOT purchasing contracts. Id. at 15-16. Ohio contends that “[t]his is precisely the kind of statistical data lacking in Croson.” Id. at 18. Though this was more data than was submitted in Cro-son, it is not sufficient under that standard. The deficiencies of the data are glaringly clear. Much of it is severely limited in scope (ODOT contracts) or is irrelevant to this case (ODOT purchasing contracts). As noted previously, the data does not distinguish minority construction contractors from minority businesses generally, and a fortiori makes no attempt to identify minority construction contracting firms that are ready, willing, and able to perform state construction contracts of any particular size. And although Ohio insists that its program is “narrowly tailored,” id. at 20-28, it concedes that “AGC showed that the State had not performed a recent study.” Id. at 19. Even statistical comparisons that might be apparently more pertinent, such as with the percentage of all firms qualified, in some minimal sense, to perform the work in question, would also fail to satisfy the Court’s criteria. If MBEs comprise 10% of the total number of contracting firms in the state, but only get 3% of the dollar value of certain contracts, that does not alone show discrimination, or even disparity. It does not account for the relative size of the firms, either in terms of tbeir ability to do particular work or in terms of the number of tasks they have the resources'to complete. Any time two non-minority firms merge, or a minority firm *737 splits in two, the total proportion of minority contracting firms in the state increases; but it would be ludicrous to imagine that such alteration affects the overall degree of discrimination. The only cases found to present the necessary “compelling interest” sufficient to “justiffy] a narrowly tailored race-based remedy” are those that expose, as in the case of the Alabama Department of Public Safety in 1987, “pervasive, systematic, and obstinate discriminatory conduct,” Ada-rand, 515 U.S. at 237, 115 S.Ct. 2097 (citing United States v. Paradise, 480 U.S. 149,107 S.Ct. 1053, 94 L.Ed.2d 203 (1987)). Ohio has made no such showing. A second and separate hurdle for the MBEA is its failure of narrow tailoring. Adarand teaches that a court called upon to address the question of narrow tailoring must ask, “for example, whether there was ‘any consideration of the use of race-neutral means to increase minority business participation’ in government contracting, Croson, [488 U.S.] at 507, 109 S.Ct. 706 ... or whether the program was appropriately limited such that it ‘will not last longer than the discriminatory effects it is designed to eliminate,’ Fullilove, [448 U.S.] at 513, 100 S.Ct. 2758....” Ada-rand, 515 U.S. at 237-38, 115 S.Ct. 2097. A narrowly-tailored set-aside program must be “linked to identified discrimination.” Croson, 488 U.S. at 507, 109 S.Ct. 706. Its criteria and measures of success must be particularized, not reduced to rigid quotas driven by “simple administrative convenience.” Id. at 508, 109 S.Ct. 706. It must also not suffer from “overinclusiveness.” Id. at 506,109 S.Ct. 706. To begin with the last-named factor in narrow tailoring analysis, the MBEA suffers from defects both of over and under-inclusiveness. By lumping together the groups of Blacks, Native Americans, Hispanics, and Orientals (and leaving unclear the exact extent of the last two designations), the MBEA may well provide preference where there has been no discrimination, and may not provide relief to groups where discrimination might have been proven. Thus, the MBEA is satisfied if contractors of, let us say, Thai origin, who might never have been seen in Ohio until recently, receive 10% of state contracts, while African-Americans receive none. Obviously, other possible examples of this sort can be readily imagined. As the Supreme Court remarked, invalidating Richmond’s set-aside program, if it were “ ‘narrowly tailored’ to compensate black contractors for past discrimination, one may legitimately ask why they are forced to share this ‘remedial relief with an Aleut citizen who moves to Richmond tomorrow?” Croson, 488 U.S. at 506, 109 S.Ct. 706. In addition to the foregoing problems, Ohio’s own “underutilization” statistics suffer from a fatal conceptual flaw, as the district court noted: they do not report the actual use of minority firms; they only report the use of minority firms who have gone to the trouble of being certified and listed among the state’s 1,180 MBEs. While it might be true that most or all of the relevant firms would have sought to take advantage of the special minority program, there is simply no examination of whether contracts are being awarded to minority firms who have never sought such preference, whether from principle, oversight, calculation of the worth of the program, or for some other reason, and who have been awarded contracts in open bidding. Narrow tailoring also implies some sensitivity to the possibility that a program might someday have satisfied its purposes. As previously noted, a race-based preference program must be “appropriately limited such that it Vill not last longer than the discriminatory effects it is designed to eliminate.’” Adarand, 515 U.S. at 238, 115 S.Ct. 2097 (quoting Fullilove v. Klutznick, 448 U.S. 448, 513, 100 S.Ct. 2758, 65 L.Ed.2d 902 (1980)). The district court in this case took note of the outdated character of any evidence that might have been marshaled in support of the MBEA and added that even if such *738 data had been sufficient to justify the statute twenty years ago, it would not suffice to continue to justify it forever. During the debate over the bill in 1980, an amendment had been offered to include a three-year “sunset” provision; this was defeated. The MBEA has remained in effect for twenty years and has no set expiration. This despite, at best, marginally adequate evidence of discrimination from 1975,1978, and 1979. As quoted above, supra, at 736, Ohio concedes this deficiency. Finally, as mentioned above, one of the factors Croson identified as indicative of narrow tailoring is whether non-race-based means were considered as alternatives to the goal. See 488 U.S. at 507, 109 S.Ct. 706. Yet, as the district court noted in an opinion written to accompany its denial of a motion to stay its judgment pending this appeal, the historical record contains no evidence “that the Ohio General Assembly gave any consideration to the use of race-neutral means to increase minority participation in state contracting before resorting to race-based quotas.” Associated Geni Contractors of Ohio, Inc., et at. v. Drabik, et at., 50 F.Supp.2d 741, 764 (SD.Oh.1999). B The district court’s denial of a motion for continuance, and its decision to consolidate the preliminary injunction hearing with a hearing on the merits, pursuant to Fed.R.Civ.P. 65(a)(2), were based on its findings of fact, and its view of the sufficiency of the evidence presented; it deemed the additional evidence that the state sought time to muster unlikely to be relevant. A district court’s determinations of relevancy are subject to review for abuse of discretion. See United States v. Seago, 930 F.2d 482, 494 (6th Cir.1991). The denial of a motion for continuance is also reviewed for abuse of discretion. See United States v. Martin, 740 F.2d 1352, 1360 (6th Cir.1984) (citing Avery v. Alabama, 308 U.S. 444, 60 S.Ct. 321, 84 L.Ed. 377 (1940)). “To determine if there has been abuse, we look to see if the defendant suffered any actual prejudice as a result of the denial, [and] whether additional time would have produced more witnesses or added something to the defendant’s case.” Ibid. We also review for abuse of discretion a district court’s decision to consolidate a hearing for a preliminary injunction with a trial on the merits. See, e.g., Berry v. Bean, 796 F.2d 713, 719 (4th Cir.1986); Northern Kentucky Chiropractic v. Ramey, No. 95-5645,1997 WL 35571, at *2 (6th Cir. Jan.29, 1997). The district court, in consolidating the preliminary injunction hearing with a trial on the merits, denied the state’s request for additional time to gather evidentiary support for the MBEA. But the MBEA has been in existence for almost twenty years, while Croson was decided eleven years, and Ada-rand more than four years, ago. Ohio provides no specifics as to the avenues it would pursue to marshal the statistics it apparently never thought relevant until now. The district court found that the supplementation of the state’s existing data which might be offered given a six-month’s continuance would not sufficiently enhance the relevance of the evidence to justify the delay. As Appellees point out in their Brief at 50-51, under Croson, the state must have had sufficient evidentiary justification for a racially conscious statute in advance of its passage; the time of a challenge to the statute, at trial, is not the time for the state to undertake factfinding. See Croson, 488 U.S. at 504, 109 S.Ct. 706 (requiring that governmental entities “must identify that discrimination ... with some specificity before they may use race-conscious relief’ (emphasis added)). The district court also noted that the state had admittedly been lax, to say the least, in maintaining the type of statistics that would be necessary to undergird its affirmative action program. The proper maintenance of current statistics is relevant to the requisite narrow tailoring of such a program, in order to judge its *739 appropriate limits. But, as noted above, the state does not even know how many minority-owned businesses are not certified as MBEs, and how many of them have been successful in obtaining state contracts. The court’s review of these deficiencies showed a firm grasp of the evidence that had been offered. Therefore, it cannot be said that the district court abused its discretion in denying the motion for a continuance or in consolidating the preliminary injunction hearing with a trial on the merits pursuant to Fed.R.Civ.P. 65(a)(2). The district court’s findings of fact, though made from the bench, are sufficient to permit this court to weigh the merits of the ruling and this appeal; oral findings of fact are explicitly contemplated by Fed. R.Civ.P. 52(a). However, we note that in a case of this importance it would have been helpful had the district court reduced to a contemporaneous writing the reasoning behind its decision of November 2, 1998. Its written ruling of May 20,1999, denying a motion for the stay of its decision pending appeal, is not an adequate substitute. See Associated Gen’l Contractors of Ohio, Inc., et al. v. Drabik, et al., 50 F.Supp.2d 741 (S.D.Oh.1999). C The State of Ohio argues that the district court should have abstained from exercising its jurisdiction in this case, on Pullman grounds, given the pendency before the Ohio Supreme Court of Ritchey. We are not persuaded. Pullman abstention is derived from a case in which a Fourteenth Amendment Equal Protection Clause challenge to a Texas railroad personnel regulation was held to have been prematurely adjudicated in federal court, since a state court’s consideration might have rendered the regulation invalid on state law grounds and rendered the federal constitutional question moot. See Railroad Comm’n of Texas v. Pullman Co., 312 U.S. 496, 61 S.Ct. 643, 85 L.Ed. 971 (1941). The Pullman abstention doctrine requires that “when a federal constitutional claim is premised on an unsettled question of state law, the federal court should stay its hand in order to provide the state courts an opportunity to settle the underlying state law question and thus avoid the possibility of unnecessarily deciding a constitutional question.” Harris County Comm’rs Court v. Moore, 420 U.S. 77, 83, 95 S.Ct. 870, 43 L.Ed.2d 32 (1975). A district court’s denial of a motion to abstain is reviewed by this court de novo. See MacDonald v. Village of Northport, Mich., 164 F.3d 964, 967-68 (6th Cir.1999) (citing Traughber v. Beauchane, 760 F.2d 673, 676 (6th Cir.1985) (“Because theories of state and federal law, and expressions of federalism and comity, are so interrelated in the decision to abstain such dispositions are elevated to a level of importance dictating de novo appellate review.”)). Abstention is, however, “a limited exception to the ‘virtually unflagging’ obligation of federal courts to exercise the jurisdiction given them.” Id. at 968 (citing Colorado River Water Conservation Dist. v. United States, 424 U.S. 800, 813, & 817, 96 S.Ct. 1236, 47 L.Ed.2d 483 (1976)). Ritchey, the state case in favor of which the district court declined to abstain, concerned the refusal of the state to certify as an MBE a produce company wholly owned by Namid Ritchey, a naturalized native of Lebanon. Ritchey wanted preference, or if not, wanted no one to get preference. After several rounds of DAS administrative rulings, appeals therefrom, and a final determination by the Director of DAS that Ritchey Produce was not certifiable as an MBE because Namid Ritchey was not “Oriental,” Ritchey took his case to the Ohio courts. The questions presented were (a) whether the MBEA is constitutional, and (b) whether a person of Lebanese origin qualifies as an “Oriental.” The common pleas court, adopting the magistrate judge’s recommendations, held that the race-based MBE program was unconstitutional, and that MBE certification *740 could only survive strict scrutiny by being recast in terms of economic disadvantage. Ohio appealed, whereupon the Ohio Court of Appeals held that “the state’s MBE program is a race per-se classification” that violates the Equal Protection Clause, and did not reach the second question of Ritchey’s racial status. 1997 WL 629965, at *3. One judge concurred in the judgment, but on the grounds that Ritchey was, indeed, an Oriental. See id. at *3-4 (Tyack, J., concurring). The Ohio Supreme Court reversed the lower court’s holding that the MBEA was unconstitutional, and also found “that the term ‘Orientals,’ as that term is used in R.C. 122.71(E)(1), does not include people of Lebanese ancestry.” 85 Ohio St.3d at 272, 707 N.E.2d at 927. It then took note of the apparent conflict between its ruling and that of the federal district court. It sought to minimize the conflict, by insisting that its holding that the MBEA is constitutional was a narrow one: We specifically wish to avoid a direct conflict between the case at bar and the specific requirements of Judge Graham’s order in Associated Gen. Contrs. of Ohio, Inc. Thus, we limit our holding today to the area of state procurement contracting. We do so in the interests of state and federal judicial comity and because the facts of the case at bar are amenable to a limited holding. 85 Ohio St.3d at 274, 707 N.E.2d at 928. Although Ritchey involved not state construction set-asides but the preference for MBEs in purchasing contracts, the statistics and the rationale underlying both those MBE programs are the same, and the constitutionality of the overall MBE scheme was before the state court, as it was before the district court. A federal court owes no duty to abstain in deference to a state court when a federal constitutional question is at issue. See England v. Louisiana Bd. of Medical Examiners, 375 U.S. 411, 415-16, 84 S.Ct. 461, 11 L.Ed.2d 440 (1964) (noting the “primacy of the federal judiciary in deciding questions of federal law”). Moreover, even if the Ohio Supreme Court could have avoided the federal constitutional question in Ritchey by a decision on state law grounds, i.e., that Ritchey is not “Oriental,” such a decision would not render moot the federal constitutional issue presented in the instant case. We note, in closing, that our opinion is not reconcilable with Ritchey, despite the Ohio Supreme Court’s attempt to distinguish the cases. See 50 F.Supp.2d at 744. Ill For the foregoing reasons, the judgment of the district court is AFFIRMED.