concurring in part and dissenting in part:
I agree with the majority that the ASPA has standing to assert its claims. But that is where the majority and I part ways. Even as is, the Complaint states a plausible claim that the City enacted section 25 as a regulatory measure rather than a proprietary one. At this stage, we must-say that this overly broad and facially suspect regulation of labor relations at Los Angeles International Airport (“LAX”)— issued, by the City’s airport commission ostensibly to promote labor peace—contravenes the delicate congressional balancing of national labor relations policy affecting key facilities of interstate commerce, I respectfully dissent.
I
A
It is well established that, in enacting the National Labor Relations Act (“NLRA”), “Congress largely displaced state regulation of industrial relations.” Wis. Dep’t of Indus., Labor, & Human Relations v. Gould Inc., 475 U.S. 282, 286, 106 S.Ct. 1057, 89 L.Ed.2d 223 (1986). “The purpose of the [NLRA] was to obtain ‘uniform application’ of its substantive rules and to avoid the ‘diversities and conflicts- likely to result from a variety of local procedures and attitudes toward labor controversies.’” NLRB v. Nash-Finch Co., 404 U.S. 138, 144, 92 S.Ct. 373, 30 L.Ed.2d 328 (1971) (quoting Garner v. Teamsters Local Union No. 776, 346 U.S. 485, 490, 74 S.Ct. 161, 98 L.Ed. 228 (1963)). To these ends, through the NLRA, Congress erected “a complex and interrelated federal scheme of law, remedy, and administration” and “entrusted administration of the labor policy for the Nation to a centralized administrative agency.” San Diego Bldg. Trades Council v. Garmon, 359 U.S. 236, 242-43, 79 S.Ct. 773, 3 L.Ed.2d 775 (1959).
Two complementary preemption doctrines serve to preserve uniformity in national labor policy. The first, Garmon preemption, “forbids States to ‘regulate activity that the NLRA protects, prohibits, or arguably, protects or prohibits.’ ” Chamber of Commerce v. Brown, 554 U.S. 60, 65, 128 S.Ct. 2408, 171 L.Ed.2d 264 (2008) (quoting Gould, 475 U.S. at 286, 106 S.Ct. 1057). The second, Machinists preemption, “prohibits state and municipal regulation of areas that have been left ‘to be controlled by the free play of economic forces.’ ” Bldg. & Constr. Trades Council of Metro. Dist. v. Associated Builders & Contractors of Mass./ R.I., Inc. (Boston Harbor), 507 U.S. 218, 225, 113 S.Ct. 1190, 122 L.Ed.2d 565 (1993) (quoting Lodge 76, Int’l Ass’n of Machinists & Aerospace Workers v. Wis. Emp’t Relations Comm’n, 427 U.S. 132, 140, 96 S.Ct. 2548, 49 L.Ed.2d 396 (1976)). Together, Garmon and Machinists preempt state and local policies that would otherwise balkanize the “integrated scheme of regulation” and disrupt the balance of power between labor and management embodied in the NLRA. Golden State Transit Corp. v. City of Los Angeles (Golden State I), 475 U.S. 608, 613-14, 106 S.Ct. 1395, 89 L.Ed.2d 616 (1986).
Similarly, the Railway Labor Act (“RLA”) established a centralized system of labor dispute resolution for the railway and airline industries to promote the free flow of interstate commerce. Aircraft Serv. Int’l, Inc. v. Int’l Bhd. of Teamsters, Local 117, 779 F.3d 1069, 1073 (9th Cir.. 2015). Machinists and Garmon preemption also apply in the RLA context. Bhd. of R.R. Trainmen v. Jacksonville Terminal Co., 394 U.S. 369, 380-81, 89 S.Ct. 1109, 22 L.Ed.2d 344 (1969).
B
As the majority correctly notes, a “market participation” exception allows state and local policies to avoid preemption analysis altogether if those policies serve to protect a proprietary interest rather than regulate the labor market. Boston Harbor, 507 U.S. at 229-30, 113 S.Ct. 1190. But by focusing solely on the market participant exception, the majority glosses over a glaring reality: if the City had no proprietary interest in LAX, section 25 would plainly be preempted by the NLRA.
Section 25 requires service providers to enter into a “labor peace agreement” (“LPA”)—a “binding and enforceable” agreement that prohibits affected employees “from engaging in picketing, work stoppages, boycotts, or any other economic interference”—with any labor organization that requests one. If a service provider and requesting labor organization cannot reach a no-strike agreement within sixty days, section 25 requires the parties to submit to binding arbitration. If a service provider refuses to abide by the terms of section 25, the City may revoke its license to do business at the airport.
Section 25 represents precisely the type of local interference in labor-management relations that Machinists preemption forbids. In Golden State I, the Supreme Court held that while the NLRA “requires an employer and a union to bargain in good faith, ... it does not require them to reach agreement,” nor does it demand a particular outcome from labor negotiations. 475 U.S. at 616, 106 S.Ct. 1395; see also 29 U.S.C. § 158(d) (providing that the duty to bargain in good faith “does not compel either party to agree to a proposal or require the making of a concession”). The substance of labor negotiations, and the results therefrom, are among those areas Congress intentionally left to the free play of economic forces when it legislated in the field of federal labor law. See Golden State I, 475 U.S. at 616, 106 S.Ct. 1395 (describing the NLRA as providing only “a framework for the negotiations”).
The facts of Golden State I are instructive—and Los Angeles has been in trouble before for flouting federal labor laws. In that case, the Supreme Court found that Machinists preempted the City of Los An-geles’ refusal to renew a taxi cab company’s license when it failed to reach an agreement with striking union members. Id. at 618, 106 S.Ct. 1395. By conditioning the renewal of the taxi cab franchise on the acceptance of the union’s demands, the City effectively imposed a timeline on the parties’ negotiations and undermined the taxi cab company’s ability to rely on its own economic power to resist the strike. Id. at 615, 106 S.Ct. 1395. The Supreme Court held that the City could not pressure the taxi cab company into reaching a settlement and thereby “destroy[ ] the balance of power designed by Congress, and frustrate!)] Congress’ decision to leave open the use of economic weapons.” Id. at 619, 106 S.Ct. 1395.
Like the taxi cab company in Golden State I, service providers here face a Hob-son’s choice plausibly inferred from the allegations of the Complaint. If a service provider refuses to negotiate an LPA with a requesting labor organization, it loses its right to do business at LAX. But if the service provider negotiates an LPA, the union knows full well that it can hold out for significant concessions in exchange for its members giving up one of their most valuable economic weapons—the power to go on strike. If the union is unsatisfied with the terms the service provider offers, the union can request mediation and binding arbitration. Once forced to arbitrate, the tribunal will dictate the result the service provider must accept. The threat of binding arbitration thus seriously limits service providers’ ability to rely on their own “economic weapons of self-help” to resist a union’s demands.
By forcing unwilling service providers to negotiate and accept LPAs, section 25 compels a result Congress deliberately left to the free play of economic forces. The NLRA does not allow state and local governments to “introduce some standard of properly balanced bargaining power ... or to define what economic sanctions might be permitted negotiating parties in an ideal or balanced state of collective bargaining.” Golden State I, 475 U.S. at 619, 106 S.Ct. 1395 (alteration in original) (quoting Machinists, 427 U.S. at 149-50, 96 S.Ct. 2548). Yet that is exactly what section 25 does. In doing so, it directly contravenes federal law.
II
A
Whether the City can enforce section 25 thus hinges entirely on the applicability of the market participant exception. The majority is willing to conclude—with little examination of what the full effects of section 25 will be—that the City’s proprietary interest in LAX immunizes section 25 from preemption. Supreme Court precedent cautions us against drawing such hasty conclusions, particularly when serious questions persist about whether section 25 advances the City’s proprietary interest.
As a preliminary matter, the Supreme Court has made clear that not every government action escapes preemption simply because it touches a proprietary interest. Gould, 475 U.S. at 287, 106 S.Ct. 1057 (calling “an exercise of the State’s spending power rather than its regulatory power.... a distinction without a difference”). The animating concern of Gould, in the words of Judge Posner, was that “[t]he [state’s] spending power may not be used as a pretext for regulating labor relations.” Metro. Milwaukee Ass’n of Commerce v. Milwaukee County (Metropolitan Milwaukee II), 431 F.3d 277, 279 (7th Cir. 2005) (emphasis added).
The fact that the City of Los Angeles owns and operates LAX through its municipal airport commission, and thus has an interest in minimizing disruptions to air travel, cannot alone qualify section 25 for the market participant exception. Instead we must determine, by examining section 25’s “actual content and its real effect on federal rights,” Livadas v. Bradshaw, 512 U.S. 107, 108, 114 S.Ct. 2068, 129 L.Ed.2d 93 (1994), whether section 25’s “manifest purpose and inevitable effect” is to do more than protect the City’s proprietary interest in running the airport, see Gould, 475 U.S. at 291, 106 S.Ct. 1057.1 Because our inquiry is informed by how section 25 might actually work in practice, it “inevitably is fact-specific,” Roger C. Hartley, Preemption’s Market Participant Immunity—A Constitutional Interpretation: Implications for Living Wage and Labor Peace Policies, 5 U. Pa. J. Lab. & Emp. L. 229, 252 (2003), and deserves more than the surface-level review undertaken by the majority.
The Supreme Court’s decision in Chamber of Commerce v. Brown, 554 U.S. 60, 128 S.Ct. 2408, 171 L.Ed.2d 264 (2008), illustrates the fact-sensitive nature of our analysis. At issue in Brown was California’s Assembly Bill 1889 (AB 1889), which prohibited certain private employers from using state funds to “assist, promote, or deter union organizing.” Id. at 63, 128 S.Ct. 2408 (quoting Cal. Gov’t Code §§ 16645.1-16645.7). The Court found it “beyond dispute that California enacted AB 1889 in its capacity as a regulator rather than a market participant.” Id. at 70, 128 S.Ct. 2408. As one obvious example, the preamble to AB 1889 announced an explicit regulatory purpose. Id.
The heart of the Court’s market participation analysis, however, focused not on AB 1889’s official purpose but on its practical consequences. Significantly, although AB 1889 purported to affect only state funds, the statute’s combination of compliance burdens and litigation risks effectively deterred employers from using any funds, state or otherwise, to exercise speech rights protected under the NLRA. Id. at 72-73, 128 S.Ct. 2408. In light of these realities, the Court held that although California had a “legitimate proprietary interest in ensuring that state funds are spent in accordance with the purposes for which they are appropriated,” in operation; AB 1889 “effectively reache[d] far beyond the use of funds over which California maintains a sovereign interest.” 2 Id. at 70-71, 128 S.Ct. 2408.
With respect to section 25, we must be similarly sensitive to the ordinance’s real-world impacts. We must also construe the allegations in the Complaint in the light most favorable to the party resisting dismissal. Syed v. M-I, LLC, 853 F.3d 492, 499 (9th Cir. 2017). Yet the majority seems content to decide, with little examination of how section 25 might actually operate, that section 25 serves a purely proprietary function. Applying the Cardinal Towing test, the majority makes a conclusory finding that, “like the college in Johnson, the City is attempting to avoid disruption of its business.” And with similarly scant analysis, the majority decides that section 25 is “narrowly tied to [the City’s] specific proprietary problem.” Distinguishing between government as market participant and government as regulator, however, requires a closer look at section 25’s “actual content” and “real effect[s].” See Livadas, 512 U.S. at 108, 114 S.Ct. 2068.
B
Under the first prong of Cardinal Towing, we cannot say that section 25 reflects the City’s interest in the “efficient procurement of needed goods and services,”' as we-might expect from a private entity. Johnson, 623 F.3d at 1023 (quoting Cardinal Towing & Auto Repair, Inc. v. City of Bedford, 180 F.3d 686, 693 (5th Cir. 1999)). At the risk of stating the obvious, the City here is not directly procuring goods and services to execute a discrete project, but rather providing ongoing'licenses permitting a host of service providers handling baggage, assisting passengers, refueling aircraft, serving food and beverages, and otherwise keeping planes operating on schedule to do business at the airport. The City’s proprietary interest here is thus markedly different in kind than that in cases like Boston Harbor and’Johnson, where local governments required project labor agreements that were “specifically tailored to one particular job.” See Boston Harbor, 507 U.S. at 232, 113 S.Ct. 1190.
Furthermore, unlike the project labor agreements in Boston Harbor and Johnson, there is no evidence that a private operator of LAX- would use LPAs as- -a means of ensuring labor peace. See Metro. Milwaukee, II, 431 F.3d at 282. Section 8(e)-(f) of the NLRA specifically authorizes the type of project .labor agreements at issue in Boston Harbor and Johnson, indicating that such agreements “are a tried and true remedy for construction stoppages owing to labor disputes.” Id. at 281-82. Nothing in the record suggests the same is true for LPAs in the private marketplace.
Indeed, if the City’s true purpose here is to minimize work stoppages at LAX, section 25 seems an .ill-fitted tool for the job. Section 25 is both too narrow and too broad as a means of achieving its purported objective. It is too narrow because, by its own terms, section 25 does not even apply to service providers’ employees, but only to the members of a labor organization that requests an LPA. Therefore, if a service provider’s employees currently have no recognized collective bargaining representative, those employees will not be covered by an LPA at all. Nor does section 25 apply to other classes of airport workers who may threaten work stoppages. Section 25 also applies only partial deterrence: it penalizes service providers, but not labor organizations', for violating an LPA.
At the same time, section 25 sweeps more broadly than necessary to achieve its goal. In order for unions to forgo their right to strike, common sense and long experience in labor negotiations tell us we would reasonably expect that service providers will have to make concessions favorable to the unions. These concessions may be totally unrelated to preventing strikes, and may or may not actually promote labor peace. Instead of forcing service providers and labor organizations into LPA negotiátions, the City could’have used other, more targeted mechanisms to prevent labor strife. In Metropolitan Milwaukee II, Judge Posner observed that
[t]he.usual way of dealing with [service interruptions] is to include contract terms that by adding sticks or carrots or both give the provider of the service a compelling incentive to take effective measures to avoid stoppages. The buyer can offer a premium for timely performance and insist on the inclusion of a stiff liquidated-damages provision as a sanction for untimely performance; there is also, as a further incentive to good performance, the implicit threat of refusing to renew the contract if performance is unsatisfactory.
431 F.3d at 280.3 Section 25 is far less straightforward. To summarize, it only covers a service provider’s employees if: (1) those employees are already represented by a labor organization; (2) that labor organization requests an LPA; (3) the labor organization and service provider enter into LPA negotiations; .and (4) the service provider makes concessions acceptable to the union, which may be totally unrelated to preventing strikes. If a service provider’s employees are not already unionized, once a labor organization secures an LPA, the labor organization must then (5) become the certified bargaining representative of the service provider’s employees through NLRB elections. Compared to simple, contract-based incentives, see id., section 25 certainly seems a roundabout way to minimize labor disruptions at LAX.
These tailoring problems suggest that section 25’s “manifest purpose and inevitable effect” may not be to protect the City’s proprietary interest in the airport at all.4 See id. (holding that tailoring problems may indicate a regulatory purpose); see also Hotel Emps. & Rest. Emps. Union, Local 57 v. Sage Hosp. Res., LLC, 390 F.3d 206, 214 (3d Cir. 2004) (noting that “[ojther appellate courts that have examined the regulator/market-participant distinction also focus on the fit between the challenged state requirement and the state’s proprietary interest in a particular project or transaction” (citing Chamber of Commerce v. Reich, 74 F.3d 1322 (D.C. Cir. 1996))). If section 25 does not directly advance the City’s proprietary interest, is it instead a pretext for regulating labor relations? The complaint plausibly alleges as much.
Historical experience with LPAs, which the majority does not bother to consider, also provides useful insight into whether section 25 reflects a proprietary interest or a regulatory one. That experience suggests that section 25’s true purpose is to alter the balance between labor and management. In typical LPAs, in exchange for relinquishing their right to strike, unions gain concessions from employers to support unionization of the employer’s employees. See Hartley, supra, at 246 (summarizing study of over one hundred LPAs). For example, LPAs often require an employer to remain neutral during union organizing drives. Id. LPAs also often require employers to provide unions with employees’ contact information and access to the employer’s physical premises to assist with organizing efforts.5 Id. at 246-47. A review of LPAs in California similarly found that, in most LPAs, “employers must grant workplace access, provide employee information (names, job titles, contact information, etc.) early in the organizing campaign,” “refrain from making disparaging statements about the union,” and/or “require that employers assent to card check recognition and neutrality.”6 Indeed, in this case, counsel for the City admitted at oral argument that unions would likely seek neutrality from service providers as part of LPA negotiations. We should therefore be unsurprised that, as the ASPA has alleged, the Service Employees International Union (SEIU) lobbied heavily for section 25 after it tried unsuccessfully to unionize service provider employees at LAX.
Given that LPAs are generally used to promote union organizing, and given counsel’s own admission at oral argument, we cannot conclude at this stage that section 25 simply reflects the City’s proprietary interest in preventing work stoppages. Moreover, the City has failed to establish that it enacted section 25 to respond to legitimate concerns about work disruptions at LAX, as we might expect from a private operator of the airport. The “manifest purpose and inevitable effect” of section 25 thus appears to be aimed at altering the balance of power between service providers and organized labor. See Gould, 475 U.S. at 291, 106 S.Ct. 1057.
C
Turning to the second prong of the Cardinal Towing test, we again cannot say conclusively at this stage that section 25’s real-world impacts will be sufficiently narrow to qualify for the market participant exception.
Even on cursory facial examination, section 25 does not appear narrowly drawn. In Johnson, we found that the challenged project labor agreement condition in that case was narrow in scope because it was both limited in time and limited to construction projects costing over $200,000. 623 F.3d at 1028. By contrast, section 25 applies to any service provider at LAX, no matter how big or small the service provider’s operations there. And section 25 is unlimited in duration; service providers must comply with its terms as long as they want to remain licensed to do business at LAX.
The practical effects of section 25 must also inform our determination of whether its scope is narrow. A challenged policy exceeds a state’s proprietary interest if the policy effectively reaches employer conduct “unrelated to the employer’s performance of contractual obligations to the state.” Boston Harbor, 507 U.S. at 228-29, 113 S.Ct. 1190; see also Brown, 554 U.S. at 71, 128 S.Ct. 2408. In Metropolitan Milwaukee II, for example, the court held that a Milwaukee County ordinance was preempted because it affected government contractors’ employees regardless of whether they performed work on government contracts. 431 F.3d at 279. The ordinance required government contractors to secure LPAs that would apply to the contractors’ “employees,” without specifying whether “employees” within the meaning of the ordinance was limited to bargaining units that worked on county contracts. Id. The unrestricted language left open the possibility that an employee who performed only some or no work for the county would be covered by an LPA, even for a labor dispute arising out of non-county work. Id.
Here, we have no assurances—besides the word of the City—that section 25 will have no similar spillover effects. The majority confidently asserts that section 25 will not “hamper service providers’ operations elsewhere.” That conclusion apparently rests on the fact that section 25 as a whole is aimed at operations at LAX. But we should be unsurprised that section 25 focuses on LAX, given that the airport authority lacks jurisdiction to directly regulate service providers beyond LAX; the City clearly cannot impose contracting conditions on service providers with whom it has no contractual relationship. The key point, however, is that nothing in section 25 limits private agreements between service providers and unions from extending beyond LAX. Nothing in section 25, for example, dictates that LPAs shall cover only LAX bargaining units. The ordinance provides only that an LPA must apply to a labor organization’s “members,” regardless of whether they perform only some or none of their work at LAX. In LPA negotiations, therefore, labor organizations may seek concessions that affect service provider employees well beyond LAX. And, depending on service providers’ business arrangements, it may be impracticable for service providers to segregate their workforces so that only employees who work exclusively at LAX are covered by an LPA.7 See Metro. Milwaukee II, 431 F.3d at 279-80.
The sheer scale of LAX may also result in spillover effects. According to the City, “LAX is the fourth busiest passenger airport in the world,” and the second busiest in the U.S. L.A. World Airports, General Information, LAX: Los Angeles World Airports, http://www.lawa.org/welcome_ lax.aspx?id=40 (last visited July 14, 2017). Last year, LAX handled over 80.9 million passengers and nearly 700,000 aircraft takeoffs and landings. Id. In Reich, the D.C. Circuit held that an Executive Order affecting all federal contracts over $100,000 served as a regulation,- and not market participation, in part because the federal government is such a large purchaser of goods and-services. 74 F.3d at 1338. Here, “given the size of [LAX’s] portion of the economy,” -labor negotiations at LAX may similarly “alter .., behavior” in the wider market for worldwide airline services. See id. The ASPA should at least be allowed to prove these potential effects.
Ill
If we are to give effect to Congress’ intent to “avoid the ‘diversities and conflicts likely to result from a variety of local procedures and attitudes toward labor controversies,’ ” Nash-Finch Co., 404 U.S. at 144, 92 S.Ct. 373, we cannot allow the market participation exception to become too broad. It is-not enough to simply accept state and local governments’ assurances that they only seek to enforce labor policies as market participants, particularly when those policies would directly interfere with core rights protected by the NLRA, itself the product of careful congressional balancing of national labor policy in industries affecting interstate commerce. Even at this early stáge of litigation, an inquiry into section 25’s “real effect on federal rights,” Livadas, 512 U.S. at 108, 114 S.Ct. 2068, raises serious doubts that the City’s interest in enforcing section 25 is merely about protecting its proprietary interest in running Los Angeles International Airport. Plaintiffs have pled enough to proceed to discovery. I respectfully dissent.
. To be clear, examining a challenged policy's purpose does not involve an investigation into policymakers’ "subjective reasons for adopting a regulation or agreement.” Johnson v. Rancho Santiago Cmty. Coll. Dist., 623 F.3d 1011, 1026 (9th Cir. 2010); see also Chamber of Commerce v. Reich, 74 F.3d 1322, 1336 (D.C. Cir. 1996) (clarifying that it was unnecessary “to question the President’s motivation in order to determine whether the [Executive] Order” demonstrated a regulatory purpose).
. The majority mischaracterizes my analysis of Brown. The critical lesson from Brown is that preemption analysis requires a careful inquiry into the actual effects of a challenged policy. Contrary to the majority’s interpretation, the Court’s analysis did not focus "solely on the text of AB 1889.” Rather, its ultimate preemption holding rested on how the statute, once operationalized, would affect the real-world choices of entities receiving state funds, and the use of funds over which the state could claim no proprietary interest. Id. at 73, 128 S.Ct. 2408 (“AB 1889’s enforcement mechanisms put considerable pressure on an employer either to forgo his ‘free speech right to communicate his views to his employees,’ or else refuse the receipt of any state funds. In so doing, the statute ... chills one side of the ‘robust debate which has been protected under the NLRA,’ ” (citation omitted)).
. The majority distinguishes Metropolitan Milwaukee II on the grounds that the county ordinance at issue in that case “did little to avoid service interruptions” and "imposed several additional conditions favorable to union organizing,” Nothing in the record establishes, however, that section 25 would achieve labor peace any more effectively. And for reasons explained infra, it- is reasonable, to assume that section 25’s practical effect is to impose conditions on service providers aimed at facilitating union organizing.
. The majority suggests that the poor fit between section 25’s actual effects and its purported goals should play no role in our preemption analysis. But tailoring issues are highly relevant to our evaluation of the first prong of the Cardinal Towing test—whether a challenged policy ”reflect[s] the [government] entity's interest in its efficient procurement of needed goods and services.” Johnson, 623 F.3d at 1023 (quoting Cardinal Towing, 180 F.3d at 693). This inquiry is distinct from examining policymakers’ motives, which does not play a role in our analysis, and the narrowness of the challenged policy’s scope, which is relevant to Cardinal Towing prong two.
. Under the NLRA, by contrast, employers may publicly oppose unionization and refuse to give labor organizations access to workplace facilities. 29 U.S.C. § 158(c) (permitting noncoercive employer speech regarding unionization); Lechmere, Inc. v. NLRB, 502 U.S. 527, 538, 112 S.Ct. 841, 117 L.Ed.2d 79 (1992) (upholding general rule that employer may not be compelled to allow nonemployee union organizers onto the employer’s property for the distribution of union literature). Section 25 thus forces service providers to give up statutory rights that would otherwise be protected.
. John Logan, Innovations in State and Local Labor Legislation: Neutrality Laws and Labor Peace Agreements in California, in The State of California Labor 2003 157, 184 (Ruth Milkman ed., 2003), available at http://www.iir. ucla.edu/publications/documents/Stateof ’ CALabor2003.pdf.
. Contrary to the majority’s suggestion, we have no indication that any “natural division” between labor performed at and outside LAX exists. It may be, for example, that some service provider employees perform work both at LAX and at one of the many other regional airports in the greater Los Angeles area. Should such an employee become involved in a labor dispute, she would be bound by an LPA entered into pursuant to section 25 regardless of whether the dispute arose at LAX or elsewhere. This type of spillover concern was central in Metropolitan Milwaukee II, 431 F.3d at 279-80.