Alvin Pearson, Brenda Curtis, and Century 21 Pearson, Inc. Realtors v. James R. Edgar and James E. Ryan
153 F.3d 397
7th Cir.
1998
Check Treatment
KANNE, Circuit Judge.

In this opinion again we address the constitutionality of an Illinois statute regulating certain real estate sales practices regarding solicitation. The district court held that the statute violates the First Amendment and the Equal Protection and Due Process Clauses of the Fourteenth Amendment. We agree that the statute violates the First Amendment, but we disagree with the district court’s other conclusions.

I. HISTORY

The Illinois General Assembly enacted 720 ILCS 590/1 — 1(d) (“the statute”) to try to prevent blockbusting, or panic peddling, by real estate agents. Blockbusting, or panic peddling, is a real estate practice in which real estate agents encourage homeowners to put their homes on the market by exploiting fears of change in the racial composition of the neighborhood and the ensuant declining home values. The statute allows homeowners to notify real estate agents that they do not wish to be solicited. Solicitation of one of these homeowners by an agent who has notice of the homeowner’s contrary wishes is a criminal offense. The statute reads as follows:

It shall be unlawful for any person or corporation knowingly:
(d) To solicit any owner of residential property to sell or list such residential property at any time after such person or corporation has notice that such owner does not desire to sell such residential property. For the purpose of this subsection, notice must be provided as follows:
(1)The notice may be given by the owner personally or by a third party in the owner’s name, either in the form of an individual notice or a list, provided it complies with this subsection.
(2) Such notice shall be explicit as to whether each owner on the notice seeks to avoid both solicitation for listing and sale, or only for listing, or only for sale, as well as the period of time for which any avoidance is desired. The notice shall be dated and either of the following shall apply: (A) each owner shall have signed the notice or (B) the person or entity preparing the notice shall provide an accompanying affidavit to the effect that all the names on the notice are, in fact, genuine as to the identity of the persons listed and that such persons have requested not to be solicited as indicated.
(3) The individual notice, or notice in the form of a list with the accompanying affidavit, shall be served personally or by certified or registered mail, return receipt requested.

The remainder of the statute makes blockbusting itself illegal; that part of the statute is not at issue in this case.

After passage of the statute, a community group in the Beverly neighborhood of Chicago known as the Beverly Area Planning Association canvassed the neighborhood to find homeowners who did not want to be solicited by real estate agents. The Association compiled a list of names and served the list on local real estate agents.

The plaintiffs are Century 21 Pearson, Inc. Realtors, Alvin Pearson, and Brenda Curtis. Pearson owns the Century 21 agency, which employs plaintiff Curtis as well as Mardie Brown. Brown made a solicitation call to a homeowner in the Beverly neighborhood after the Century 21 agency had been served with a list containing that homeowner’s name. Pearson, Brown, and the Century 21 agency were charged with violating the statute; the criminal complaints were dismissed but not before each was sentenced to a $100 fine and placed under court supervision. On March 31, 1986, the plaintiffs filed this civil rights suit alleging that the statute violates the First Amendment. They also alleged *400 that the statute is unconstitutionally vague and violates the Equal Protection Clause.

The plaintiffs moved for a preliminary injunction to halt application of the statute during litigation. The district court found that the complaint failed to state a claim under the First Amendment and that the plaintiffs had no likelihood of success on the merits; the district court therefore denied the motion for a preliminary injunction. Only plaintiff Curtis pursued an interlocutory appeal. We affirmed the district court’s denial of the preliminary injunction and remanded the case for further proceedings in an opinion that became the key to this long running saga. See Curtis v. Thompson, 840 F.2d 1291 (7th Cir.1988).

The district court read our opinion in Curtis to preclude any chance of success for any of the plaintiffs and to establish the law of the case, and the district court dismissed the case in its entirety. See Pearson v. Thompson, No. 86 C 2181, 1989 WL 88367 (N.D.Ill. July 26, 1989). All the plaintiffs then appealed, and we affirmed on law of the case grounds. See Pearson v. Thompson, 955 F.2d 46, 1992 WL 25349 (7th Cir. Feb. 13, 1992), vacated, Pearson v. Edgar, 507 U.S. 1015, 113 S.Ct. 1809, 123 L.Ed.2d 441 (1993).

The plaintiffs petitioned for a -writ of cer-tiorari. The Supreme Court granted the writ, vacated our second opinion, and remanded the case to us “for further consideration in light of City of Cincinnati v. Discovery Network, Inc., 507 U.S. 410, 113 S.Ct. 1505, 123 L.Ed.2d 99 (1993).” Pearson v. Edgar, 507 U.S. 1015, 113 S.Ct. 1809, 123 L.Ed.2d 441 (1993). We, in turn, remanded the case to the district court “for consideration of the impact of Cincinnati v. Discovery Network.... The district court should conduct an evidentiary hearing to allow the parties to create the appropriate record for determining the constitutionality of Ill.Rev. Stat. ch. 38 7051(d) under the new standards set out by Discovery Network.” Pearson v. Thompson, 4 F.3d 997, 1993 WL 315601 (7th Cir. Aug. 12, 1993).

On remand, the district court conducted that evidentiary healing but then decided that further proceedings would help it make a better decision about the constitutionality of the statute. To that end, the district court ordered the parties to prepare for a bench trial.

At the conclusion of the trial, the district court analyzed the state’s two asserted justifications for the statute: preventing blockbusting and protecting residential privacy. Some of the key findings of fact relating to those justifications are:

37. There is no evidence that standard real estate marketing materials cause rapid racial change or contribute to panic selling.
38. Panic peddling and blockbusting did occur in Chicago during the 1960s and early 1970s.
39. However, blockbusting and panic peddling rarely, if ever, occur in Illinois today.
49. During the bench trial, the defendants produced no evidence in this case that real estate solicitation harms or threatens to harm residential privacy.

Pearson v. Edgar, 965 F.Supp. 1104, 1108-09 (N.D.Ill.1997) (citations to record omitted). Relying on these key findings, the district court applied the Central Hudson test for restrictions on commercial speech. See Central Hudson Gas & Elec. Corp. v. Public Serv. Comm’n, 447 U.S. 557, 566, 100 S.Ct. 2343, 65 L.Ed.2d 341 (1980). Regarding both blockbusting and residential privacy, the district court held that the state failed to satisfy the Central Hudson test. Specifically, the district court held that the state failed to prove that the statute directly advances its asserted goals and also failed to prove that the statute is not more extensive than necessary to serve those interests. The district court held the statute unconstitutional and granted relief to the plaintiffs. The state appeals from that decision.

II. ANALYSIS

A. Standards of Review

We review the findings of fact for clear error. See Fed.R.Civ.P. 52(a) (“Findings of fact ... shall not be set aside unless clearly erroneous....”); Thornton v. *401 Brown, 47 F.3d 194, 196 (7th Cir.1995). A factual finding is clearly erroneous only if we are “left with the definite and firm conviction that a mistake has been committed.” Thornton, 47 F.3d at 196 (citations omitted). We review the district court’s conclusions of law de novo. See McFarlane v. Life Ins. Co. of North Am., 999 F.2d 266, 267 (7th Cir.1993). And in First Amendment jurisprudence, “[t]he party seeking to uphold a restriction on commercial speech carries the burden of justifying it.” Bolger v. Youngs Drug Prods. Corp., 463 U.S. 60, 71 n. 20, 103 S.Ct. 2875, 77 L.Ed.2d 469 (1983). “This burden is not satisfied by mere speculation or conjecture; rather, a governmental body seeking to sustain a restriction on commercial speech must demonstrate that the harms it recites are real and that its restriction will in fact alleviate them to a material degree.” Edenfield v. Fane, 507 U.S. 761, 770-71, 113 S.Ct. 1792, 123 L.Ed.2d 543 (1993).

B. Curtis v. Thompson

We begin our analysis by reviewing our decision in Curtis. As noted in Curtis we upheld the district bourt’s denial of a preliminary injunction. We observed that the parties had not had a chance to present affidavits or other evidence. See id. at 1295. However, we determined on the record before us that the statute properly advanced the state’s interest in residential privacy. See id. at 1299, 1302. That being the case, we agreed with the district court that Curtis had no likelihood of success on the merits and therefore a preliminary injunction was not warranted.

In reaching that result, we applied the four-part Central Hudson test for restrictions on commercial speech. We asked:

1. Whether the speech concerns lawful activity and is not misleading;
2. “[Whether t]he asserted governmental interest is substantial”;
3. “[W]hether the regulation directly advances the governmental interest asserted”;
4. “[W]hether it is not more extensive than is necessary to serve that interest.”

Central Hudson, 447 U.S. at 566, 100 S.Ct. 2343. Steps three and four of the Central Hudson test examine the fit between the restriction on speech and the government’s justification for that restriction. See Rubin v. Coors Brewing Co., 514 U.S. 476, 486, 115 S.Ct. 1585, 131 L.Ed.2d 532 (1995).

We applied that test first to the state’s asserted interest in preventing blockbusting and noted a lack of evidence at that stage of the proceedings to show whether the statute met the third prong of the test. See Curtis, 840 F.2d at 1299. We advanced to the state’s other asserted interest, protecting residential privacy. We held unequivocally that the statute met all four parts of the Central Hudson test for the state’s residential privacy interest. Specifically, we held that 1) real estate solicitation concerns lawful activity and is not misleading; 2) the state’s interest in protecting residential privacy is substantial; 3) the statute directly advances the cause of protecting residential privacy; and 4) the statute is not more extensive than necessary to protect residential privacy. See id. at 1298-1303.

This Court has twice had the opportunity to revisit Curtis. As discussed above, on remand the district court dismissed the entire ease reasoning that Curtis established the law of the case and mandated such disposition. See Pearson v. Thompson, No. 86 C 2181, 1989 WL 88367 (N.D.Ill. July 26, 1989). The plaintiffs appealed the district court’s decision, and we affirmed on law of the case grounds. See Pearson v. Thompson, 955 F.2d 46, 1992 WL 25349 (7th Cir. Feb. 13, 1992), vacated, Pearson v. Edgar, 507 U.S. 1015, 113 S.Ct. 1809, 123 L.Ed.2d 441 (1993). We did not disturb Curtis’s reasoning or holding.

Our second occasion to apply Curtis came in South-Suburban Housing Ctr. v. Greater South Suburban Bd. of Realtors, 935 F.2d 868 (7th Cir.1991). South-Suburban involved a municipal ordinance similar to the statute at issue in Curtis and again here. The government asserted the interests of preventing blockbusting and protecting residential privacy. See id. at 876. We found Curtis to be binding precedent and fully embraced it, holding that the government’s *402 asserted interest in protecting residential privacy adequately supported the restrictions on commercial speech that the ordinance occasioned. See id. at 894.

Absent any intervening Supreme Court decisions, Curtis and South-Suburban would be binding precedent on this issue, and Curtis would also be the law of the case. But Discovery Network has intervened, and our task today is to decide whether it has changed the law of commercial speech such that Curtis and Southr-Suburban no longer rest on solid ground. Both Curtis and Southr-Suburban dealt largely with the state’s interest in residential privacy. The state here asserts two justifications: preventing blockbusting and protecting residential privacy. As either justification could potentially support the statute’s restrictions on commercial speech, we address them both.

C. Anti-Blockbusting Interest

At trial, the state reasserted its interest in preventing blockbusting as a justification for the statute. As we noted above, the district court found as a fact that blockbusting no longer occurs with any frequency in Illinois. After an independent review of the record, we cannot call that finding clearly erroneous. Without evidence that blockbusting is a problem in Illinois, the state cannot satisfy the third and fourth prongs of Central Hudson, that “the regulation directly advances the governmental interest asserted,” and that the regulation “is not more extensive than is necessary to serve that interest.” Central Hudson, 447 U.S. at 557, 100 S.Ct. 2343. In other words, the state did not show that “the harms it recites are real and that its restriction will in fact alleviate them to a material degree.” Edenfield, 507 U.S. at 770-71, 113 S.Ct. 1792. The district court properly held that the state’s interest in anti-blockbusting cannot support its restriction on real estate solicitation. 1

D. Residential Privacy Interest

Addressing the state’s asserted interest in protecting residential privacy is more complicated because our decisions in Curtis and Southr-Suburban relied on this justification.

1.

As we previously mentioned, the Supreme Court vacated our decision in Pearson v. Thompson, 955 F.2d 46, 1992 WL 25349 (7th Cir. Feb. 13, 1992) (where we affirmed the district court’s dismissal of the complaint on the grounds that our decision in Curtis formed the law of the case) for further consideration in light of Discovery Network, 507 U.S. at 410, 113 S.Ct. 1505. In that case, the Supreme Court considered a restriction on commercial speech imposed by the City of Cincinnati. Cincinnati banned the distribution of commercial handbills from freestanding newsracks but it imposed no similar ban on “regular” newspapers. See id. at 414, 113 S.Ct. 1505. Cincinnati cited safety and esthetics as justifications for the ban on news-racks. See id. at 416, 113 S.Ct. 1505. The Supreme Court accepted those interests as substantial but nonetheless held that the distinction Cincinnati made between commercial speech and noncommercial speech violated the First Amendment. See id. at 424, 113 S.Ct. 1505. Two major concepts permeate Discovery Network.

First is the importance of “reasonable fit” between the restriction on speech and the goal to be achieved by that restriction. Discover Network, 507 U.S. at 417, 113 S.Ct. 1505 (quoting Board of Trustees of State Univ. of N.Y. v. Fox, 492 U.S. 469, 480, 109 S.Ct. 3028, 106 L.Ed.2d 388 (1989)). The Court noted that commercial and noncommercial newsracks posed equal safety problems and were equally ugly, yet the distinction between commercial and noncommercial speech bore “no relationship whatsoever to the particular interests that the city has as *403 serted,” safety and esthetics. Id. at 424, 113 S.Ct. 1505. Nothing in the content of the commercial handbills within the racks worsened the racks’ safety or visual appeal. Furthermore, the Court was troubled by the inefficacy of Cincinnati’s ban on commercial newsracks. Commercial newsracks made up a small fraction of the total newsracks on Cincinnati’s streets, yet Cincinnati banned them to try to solve a problem caused by all newsracks equally. Because banning such a small number of newsracks could bring no major improvement in safety or esthetics, the Court concluded that the restriction on commercial newsracks fit the asserted justifications poorly and therefore violated the First Amendment. See id. at 418, 113 S.Ct. 1505.

Second, the Court stressed the value of commercial speech. Cincinnati undervalued commercial speech; it placed too much emphasis on the difference between commercial and noncommercial speech in tailoring its ban. The Court said that the ban on all commercial newsracks without any restrictions on noncommercial newsracks “seriously underestimates the value of commercial speech.” Id. at 419, 113 S.Ct. 1505. While commercial speech enjoys less protection than noncommercial speech, it is not entitled to so much less protection that it may be banned without adequate justification.

2.

When we apply these two delineated concepts to Curtis, it becomes apparent that our prior decision has been rendered inconsistent with the approach announced by the Supreme Court in Discovery Network. It is true that the state has carried its burden as to the first and second parts of the Central Hudson test: the speech involved here is lawful and not misleading, and the state’s interest in protecting residential privacy is substantial. We are compelled, nevertheless, to determine that the state has not carried its burden on the third and fourth parts of that test by showing a “reasonable fit” between the limited ban on real estate solicitation and the state’s interest in protecting residential privacy.

The concept of “fit” predates Discovery Network and has been applied to many commercial speech cases. We examine Curtis for proper fit using the Discovery Network standard even though Curtis involves a distinction between kinds of commercial speech and not specifically a distinction between commercial and noncommercial speech. See Fox, 492 U.S. at 480, 109 S.Ct. 3028.

Other cases before and after Discovery Network have tried to define the precision of fit that the First Amendment requires between restrictions on commercial speech and the government’s asserted justification. The Supreme Court’s decisions require “a ‘fit’ between the legislature’s ends and the means chosen to accomplish those ends—a fit that is not necessarily perfect, but reasonable; that represents not necessarily the single best disposition but one whose scope is ‘in proportion to the interest served; that employs not necessarily the least restrictive means but, as we have put it in the other contexts discussed above, a means narrowly tailored to achieve the desired objective.’ ” Fox, 492 U.S. at 480, 109 S.Ct. 3028 (citations omitted). We must be satisfied that the state’s restriction on speech “advances [the state’s] interests in a direct and material way.” Edenfield, 507 U.S. at 767, 113 S.Ct. 1792. In sum, we require a “reasonable fit.” Discovery Network, 507 U.S. at 417, 113 S.Ct. 1505 (quoting Fox, 492 U.S. at 480, 109 S.Ct. 3028).

In Curtis, we examined the peculiar relationship between the statute and the state’s proffered interest in residential privacy. We noted that if the state were concerned about the intrusion of solicitation on residential privacy, it could have addressed the problem more effectively by allowing homeowners to reject solicitation of any commercial enterprise, not just of real estate agents. We upheld the state’s decision to attack the smaller problem of real estate solicitation, though, because the Central Hudson test is concerned with overextensive, not underextensive statutes. See Curtis, 840 F.2d at 1303. We said that “[w]hile the Illinois statute could have been more extensive, the state’s decision to limit the scope of the statute does not translate into a constitutional problem.” Id. at 1303.

*404 The limited ban at issue in Discovery Network was underinclusive — it only attacked part of the problems of safety and esthetics caused by newsracks. The limited ban here is also underinclusive. While in Curtis we were not concerned too much with this fact, now we must be, for in Discovery Network, underinclusiveness indicated a lack of reasonable fit. Here too we think that the underin-clusive nature of the statute indicates unreasonable fit. The district court found as a fact that the state produced “no evidence in this ease that real estate solicitation harms or threatens to harm residential privacy.” Pearson v. Edgar, 965 F.Supp. 1104, 1109 (N.D.Ill.1997). Absént such evidence, a mechanism whereby homeowners can reject real estate solicitation but not other kinds of solicitation cannot be said to advance the interest in residential privacy “in a direct and material way.” Edenfield, 507 U.S. at 767, 113 S.Ct. 1792. Severe underinclusiveness is not forgivable after Discovery Network’s condemnation of unjustified limited bans if the underinclusiveness indicates poor fit.

In Curtis, we also stressed the importance of deference to the legislative branch of government. We declined to “second-guess” the Illinois General Assembly’s judgment that real estate solicitation poses a special problem to residential privacy because “‘[ejven on matters touching the First Amendment, courts must accept plausible judgments by other governmental actors.’” Id. (quoting American Jewish Congress v. City of Chicago, 827 F.2d 120, 130 (7th Cir.1987) (Easterbrook, J., dissenting)). However, the distinction between real estate solicitation and other types of solicitation is not plausible absent evidence that real estate solicitation poses a particular threat to residential privacy. The restriction on speech does not reasonably fit the reason for the restriction.

In Curtis, we relied on Rowan v. United States Post Office Dept., 397 U.S. 728, 90 S.Ct. 1484, 25 L.Ed.2d 736 (1970). 2 That reliance too is weakened by Discovery Network’s emphasis on reasonable fit. Rowan involved a federal statute that permitted homeowners to contact the Postmaster General if the homeowner received any mailings that the homeowner considered to be erotic or sexual. See id. at 730, 90 S.Ct. 1484. The Postmaster would then contact the sender and order that entity to delete the homeowner’s name from its mailing list. See id. The Court upheld the statute, proclaiming that “a mailer’s right to communicate must stop at the mailbox of an unreceptive addressee.” Id. at 736-37, 90 S.Ct. 1484. In Rowan, a homeowner could prevent any material from entering his home, even a “dry goods catalog,” if the homeowner deemed the material offensive. Rowan, 397 U.S. at 737, 90 S.Ct. 1484. Here, the homeowner cannot ban any bothersome solicitation but only real estate solicitation. The distinction between real estate solicitation and other kinds of solicitation is a by-product of the statute’s original purpose of preventing blockbusting and cannot, we think, satisfy the Discovery Network requirement of a “ ‘fit’ between [the state’s] goals and [the state’s] chosen means.” Discovery Network, 507 U.S. at 428, 113 S.Ct. 1505. Rowan involved no governmental distinction between one kind of offensive material and another — all distinctions were made by the homeowner. Here, the state, not the homeowner, has made the distinction between real estate solicitations and other solicitations without a logical privacy-based reason. We can no longer, after Discovery Network, place the interest in residential privacy above the interest in logical distinctions in speech restrictions absent some showing that the restriction reasonably fits the justification.

Second, we examine Curtis in light of Discovery Network’s emphasis on the value of commercial speech. We relied considerably on the fact that commercial speech is of less value than noncommercial speech: “In the area of noncommercial speech, content-based restrictions (such as the one at hand) are sustained only in the most extraordinary circumstances .... By contrast, regulation of commercial speech based upon content is less *405 problematic.” Curtis, 840 F.2d at 1297-98. While that statement remains true today as the continued viability of the Central Hudson test shows, we cannot say that the difference in value between commercial speech and noncommercial speech is as great as it was before Discovery Network. Cincinnati, in that case, “seriously underestimate[d] the value of commercial speech,” Discovery Network, 507 U.S. at 419, 113 S.Ct. 1505, and we do not wish to be guilty of the same mistake.

Our reliance on the lower value of commercial speech is apparent in our analysis of the relationship between the homeowner’s interest in privacy and the business person’s interest in promoting wares. We decided in Curtis that “[w]hen the fundamental right to privacy clashes with the right of free expression, the interest in privacy does not play second fiddle when the speech is merely intended to propose a commercial transaction.” Curtis, 840 F.2d at 1300. That is, we found the interest in privacy to be absolutely greater than the interest in commercial speech. We cannot say that after Discovery Network this is still the state of affairs. While residential privacy is unquestionably a very substantial interest, see Frisby v. Schultz, 487 U.S. 474, 108 S.Ct. 2495, 101 L.Ed.2d 420 (1988); Carey v. Brown, 447 U.S. 455, 100 S.Ct. 2286, 65 L.Ed.2d 263 (1980), so is the real estate agent’s interest in using commercial speech to solicit business, see Discovery Network, 507 U.S. at 419, 113 S.Ct. 1505.

In sum, we stand behind our analyses in Curtis and SoutE-Suburban insofar as they are not affected by Discovery Network, but Discovery Network has changed the law of commercial speech on which those cases were built. Because the state has been unable to explain how this ban on real estate solicitation directly and materially advances its interest in protecting against blockbusting and protecting residential privacy, we conclude that the statute does not reasonably fit the asserted state interests and cannot pass the Central Hudson test for restrictions on commercial speech. We therefore hold that the statute violates the First Amendment.

E. Equal Protection and Vagueness

We remanded this ease to the district court “for consideration of the impact of Cincinnati v. Discovery Network, 507 U.S. 410, 113 S.Ct. 1505, 123 L.Ed.2d 99 (1993). The district court should conduct an evidentiary hearing to allow the parties to create the appropriate record for determining the constitutionality of Ill.Rev.Stat. ch. 38 7051(d) under the new standards set out by Discovery Network.” Pearson v. Thompson, 4 F.3d 997, 1993 WL 315601 (7th Cir. Aug. 12, 1993). We said nothing about arguments relating to equal protection or vagueness, however, the district court proceeded to evaluate and decide these questions.

The direction we gave bound the district court. When a remand is limited to a specific purpose, the district court may not venture into other areas. See United States v. Polland, 56 F.3d 776, 777 (7th Cir.1995); Ohio-Sealy Mattress Mfg. Co. v. Sealy, Inc., 669 F.2d 490, 493 (7th Cir.1982). We affirmatively decided the issues of equal protection and vagueness in Curtis, 840 F.2d at 1304 n. 12, and in Pearson v. Thompson, 955 F.2d 46, 1992 WL 25349 (7th Cir. Feb. 13, 1992), vacated, Pearson v. Edgar, 507 U.S. 1015, 113 S.Ct. 1809, 123 L.Ed.2d 441 (1993). Even though the Supreme Court vacated our entire judgment in Pearson, Curtis still establishes the law of the case on equal protection and vagueness. Discovery Nehvork did not affect either of these areas of the law. Because the district court had no authority to revisit equal protection and vagueness, its decision with respect to those issues is without effect.

For the foregoing reasons, we Afferm the judgment of the district court regarding the First Amendment and we Vacate the judgment of the district court regarding equal protection and vagueness.

Notes

1

. Two other courts have decided this issue the same way, although in each case the statute prevented real estate solicitation of all residents in an area, not only residents who affirmatively rejected solicitation. Cf. New York State Ass'n of Realtors, Inc. v. Shaffer, 27 F.3d 834, 844 (2d Cir.1994) (banning solicitation in entire zones not narrowly tailored to legitimate interest in preventing blockbusting); Greater Baltimore Bd. of Realtors, Inc. v. Baltimore County, Md., 752 F.Supp. 193 (D.Md.1990) (holding that the state failed to show blockbusting is a problem today and that the ban is not narrowly tailored).

2

. Our Rowan analysis in Curtis did not garner a majority of votes. However, we reaffirmed the Rowan analysis in South-Suburban by a unanimous vote, so we will treat the Rowan analysis as if it were established by a majority in Curtis.