Michigan Building & Construction Trades Council v. Snyder

729 F.3d 572, 2013 WL 4767011, 196 L.R.R.M. (BNA) 2825, 2013 U.S. App. LEXIS 18580

• Court: Court of Appeals for the Sixth Circuit
• Decided: September 6, 2013
• Docket: 12-1246, 12-2548
• Status: Published

Opinions

ROGERS, J., delivered the opinion of the court, in which SILER, J., joined. MOORE, J., (pp. 582-93), delivered a separate dissenting opinion.

OPINION

ROGERS, Circuit Judge.

Project labor agreements are contracts typically used in the construction industry to set common terms and conditions of employment for large projects involving multiple subcontractors and unions. The question on this appeal is whether the State of Michigan—with respect to the construction of public projects—can make an across-the-board determination not to require that its contractors enter into such agreements. Such an across-the-board determination could be made by a private developer. Michigan can do the same because in this respect the state is acting as a market participant rather than as a regulator.

Michigan passed the first version of the Fair and Open Competition in Governmental Construction Act in 2011. Plaintiffs— state and local trades councils—claimed that the act was preempted by the. National Labor Relations Act (NLRA) and asked for an injunction. The district court granted the injunction in February 2012. That version of the act has been entirely superseded by an amended version of the act, passed in 2012, rendering the Governor’s appeal of that injunction moot. The district court subsequently enjoined the current version of the act, finding it preempted by the NLRA. However, the act furthers Michigan’s proprietary goal of improving efficiency in public construction projects, and the act is no broader than is necessary to meet those goals. Thus, the law is not preempted by the NLRA.

Both versions of the act restrict the use of Project Labor Agreements (PLAs) on publicly funded construction projects. A PLA sets out the terms and conditions of employment on a specific construction project. On a public construction project, the PLA can be entered into by the governmental unit paying for the project or by a general contractor the governmental unit hires. The other party to the PLA is the relevant labor organization. Once a PLA is in force, every lower-level contractor must abide by it to be able to work on the project. Thus, if the governmental unit itself enters into a PLA, all contractors bidding on the project must agree to abide by the PLA. If a general contractor enters into a PLA, all its subcontractors on that project must agree to abide by the PLA. The PLAs will often incorporate terms from individual local union collective bargaining agreements, but the PLA will supersede those agreements.

There has been debate over whether PLAs increase the costs of government projects. Opponents of PLAs argue that PLAs discourage nonunion contractors and subcontractors from bidding on government contracts and that the rules included in PLAs increase construction costs. The Governor cites reports that found that PLAs add 12-18% to the costs of public projects.¹ PLA proponents, like the trades councils, counter that PLAs enhance job-site cooperation and reduce labor disputes, thus preventing delays and cost overruns on public projects. The trades councils cite several reports supporting their position.² The federal government has gone back and forth on whether PLAs should be permissible for federally funded projects. President George H.W. Bush used an executive order to forbid the use of PLAs on federally funded projects. See Exec. Order No. 12,-818, 57 Fed.Reg. 48713 (Oct. 23, 1992). President Clinton rescinded that order and encouraged the use of PLAs on construction projects over $5 million. See Exec. Order No. 12,836, 58 Fed.Reg. 7045 (Feb. 1, 1993). President George W. Bush reinstated the ban on the use of PLAs on federally funded projects, see Exec. Order No. 13,302, 66 Fed.Reg. 11225 (Feb. 17, 2001), only to see President Obama again lift the ban and allow agencies to use PLAs for construction projects that cost the government over $25 million, see Exec. Order No. 13,502, 74 Fed.Reg. 6985 (Feb. 6, 2009).

The Michigan legislature stepped into this debate in 2011 by passing S.B. 165, the Fair and Open Competition in Governmental Construction Act, 2011 Mich. Pub. Acts 98. The introduction to the act stated that its goal was

to provide for fair and open competition in governmental construction contracts, grants, tax abatements, and tax credits; to prohibit requirements for certain terms in government contracts and contracts supported through government grants and tax subsidies and abate-ments; to prohibit expenditure of public funds under certain conditions; to prohibit certain terms in procurement documents for certain expenditures involving public facilities; and to provide for powers and duties of certain public officers, employees, and contractors.

Id. The act barred governmental units from entering or expending funds on a project if the contract or any subcontract contained a PLA. Id. § 5. It also forbade the governmental units from awarding grants, tax abatements, or tax credits while under a PLA, id. § 7, and forbade governmental units and their agents from placing any PLA terms in bid specifications, project agreements, or other controlling documents, id. § 9.

The Michigan Building and Construction Trades Council, AFL-CIO, and Genesee, Lapeer, Shiawassee Building and Construction Trades Council, AFL-CIO, both associations of labor organizations, filed suit. The trades councils argued, among other things, that the new law was preempted by the NLRA, which permitted the use of PLAs. The district court agreed and issued a preliminary injunction. The court rejected the Governor’s argument that the state was acting in its proprietary capacity when it passed the law. The court determined that the law was regulatory and that it was preempted by Sections 7 and 8 of the NLRA. Mich. Bldg, and Const. Trades Council, AFL-CIO v. Snyder, 846 F.Supp.2d 766, 783 (E.D.Mich.2012).

The Governor appealed. While that appeal was pending in this court, the Michigan legislature amended the act. The legislature clarified that it intended the act “to provide for more economical, nondiscriminatory, neutral, and efficient procurement of construction-related goods and services by this state and political subdivisions of this state as market participants,” and that “providing for fair and open competition best effectuates this intent.” 2012 Mich. Pub. Acts 238 § 2 (codified at Mich. Comp. Laws § 408.872). The amended act replaced Section 5, which had previously barred government spending on any project that included a contract or a subcontract that contained a PLA. The new Section 5 only barred governmental units from entering into PLAs themselves. Id. § 5 (codified at Mich. Comp. Laws § 408.875). It also forbade governmental units from discriminating against bidders on public projects based on whether the bidder had entered into a PLA. Id. The legislature also added a new section to the act stating that the act “does not prohibit a governmental unit from awarding a contract, grant, tax abatement, or tax credit to a private owner, bidder, contractor, or subcontractor who enters into or who is party” to a PLA so long as entering into that PLA “is not a condition for award of the contract, grant, tax abatement, or tax credit....” Id. § 8 (codified at Mich. Comp. Laws § 408.878).

These changes satisfied neither the trades councils nor the district court. On the trades councils’ request, the district court enjoined the new version of the act as well. Once again, the district court found the act preempted by Sections 7 and 8 of the NLRA because the act still prohibited governmental units from requiring their contractors to adhere to PLAs as a condition of a contract award. The court also found, contrary to the legislature’s statement that the state was acting as a market participant, that the act is tantamount to regulation because it is broad in scope and “does not reflect the State’s interest in efficient procurement of goods or services.” The court noted that if the state were acting as a private proprietor would, it would consider PLAs on a case-by-case basis and would not issue a blanket prohibition. Mich. Bldg. & Constr. Trades Council v. Snyder, No. 12-13567, 2012 WL 6155964, at *7 (E.D.Mich. Nov. 15, 2012).

The Governor also appealed the second injunction. Upon the Governor’s unopposed request, this court consolidated the two appeals.

The Governor’s appeal of the injunction of the original version of the act is now moot. That version has been replaced entirely by the current version, enacted in 2012, and removing the injunction would have no effect. A case becomes moot “when it is impossible for a court to grant any effectual relief whatever to the prevailing party.” Knox v. Serv. Emps. Int’l Union, Local 1000, — U.S. -, 132 , 183 L.Ed.2d 281 (2012) (internal quotation marks omitted). “[I]f an event occurs while a case is pending on appeal that makes it impossible for the court to grant ‘any effectual relief whatever’ to a prevailing party, the appeal must be dismissed.” Church of Scientology of Cal. v. United States, 506 U.S. 9, 12, 113 S.Ct. 447, 121 L.Ed.2d 313 (1992). The enactment of the current version of the act is such an event. Therefore, the Governor’s appeal of the first injunction must be dismissed.

Both parties argue that the appeal of the original injunction is not moot, but their arguments are unavailing. The Governor argues that the appeal is not moot because the councils make the same preemption arguments with regard to the current version of the act as they did against the first version. However, those arguments are appropriately considered in light of the language of the law as it stands now, not as it was before the amendments. The councils note that a reversal of the first injunction will jeopardize PLAs signed in the period while the original act was enjoined but before the amended act went into effect. However, the Governor has given no indication that the state intends to challenge contracts entered into by governmental units before the enactment of the current version of the act, and no such challenges have been identified. While the original act barred governmental units from entering into or expending funds under contracts that contained PLAs, see 2011 Mich. Pub. Acts 98 § 5, the current version only bars governmental units from entering into new PLAs on or after the effective date of the amendatory act, see 2012 Mich. Pub. Acts 238 § 5 (codified at Mich. Comp. Laws § 408.875). The Governor has repeatedly contended before this court that, contrary to its apparent language, the original act’s exceptions meant that it did not bar governmental units from entering into contracts with contractors or subcontractors that had PLAs. See Reply Br. in No. 12-1246 at 5-6. Thus, we take these contentions as a representation that the state will not seek to invalidate contracts that would be legal under the current version of the act on the basis of illegality under the original version.

The parties agree that if the state’s statutes are proprietary rather than regulatory, the councils’ arguments based on the NLRA fail. See Bldg. & Constr. Trades Council v. Associated Builders & Contractors, 507 U.S. 218, 227, 113 S.Ct. 1190, 122 L.Ed.2d 565 (1993). The current version of the act is proprietary. The state is seeking to preserve taxpayer resources by encouraging open competition among potential contractors and subcontractors. It is not banning PLAs, and contractors who enter into PLAs can compete on equal footing with non-PLA contractors for public contracts. Private entities, including contractors working on government projects, remain free to enter into PLAs. The law’s effect is limited to forbidding governmental units from entering into PLAs and then forcing the terms and conditions found within on bidders, contractors, and subcontractors. Such a limited action is similar to those found to be proprietary by the Supreme Court, this court, and other circuits.

The statements of intent within the legislation and its legislative history provide evidence that the legislation was passed in an effort to improve efficiency in government projects, not to regulate. The act specifically states that it is intended “to provide for more economical, nondiscriminatory, neutral and efficient procurement of construction-related goods and services by this state and political subdivisions.” 2012 Mich. Pub. Acts 238 § 2 (codified at Mich. Comp. Laws § 408.872). The legislative history further demonstrates this intent. Senator Moolenaar explained that the act is intended to “guarantee[] the equal opportunity and fiscal accountability that taxpayers expect from government.” S. Journal 48, 96th Leg., at 867 (Mich. 2012). ' Even the act’s opponents believed that the act was intended to save resources, with Senator Gleason arguing that the act constitutes a terrible decision to “go cheap on labor” because it would lead to the use of lesser-skilled workers and shabbily built projects. He noted that it would be better to have “high standards and high qualifications, not low cost.” Id. at 866-67.

The limits of the act demonstrate its proprietary nature. The act affects only the actions of the state and political subdivisions of the state. It has no effect on private projects. Furthermore, even its effect on public projects is limited. The act forbids governmental units and their agents from entering into PLAs. It does not forbid the use of PLAs on public projects. If a governmental unit uses a general contractor on a project, and that general contractor is responsible for all subcontracting, the general contractor could enter into a PLA that would cover the entire project. Under the act, bidding on public construction projects must be open to contractors whether or not they are parties to PLAs, and a governmental unit cannot discriminate in favor of or against a contractor because it is party to a PLA.2012 Mich. Pub. Acts 238 § 5(b) (codified as Mich. Comp. Laws § 408.875(b)).³ The councils argue that the act severely restricts the use of PLAs because PLAs are ineffective if they do not apply to every contractor on the project, and some contracts involve more than one general contractor. However, all that a governmental unit needs to do to encourage bids that include effective PLAs is choose to give all subcontracting power to one general contractor. Furthermore, for those projects where a PLA is the most efficient way to proceed, the general contractors are free to come together to enter a PLA.

The councils argue that the act is too broad to be proprietary because it does not consider projects on a case-by-case basis. But private proprietors can and do act on an across-the-board basis without somehow becoming regulators. The legislature permissibly decided that public resources would best be preserved by taking the PLA decision out of the hands of governmental units and leaving it to private contractors.

The trades councils also argue that the statute is too broad because it extends coverage to governmental projects funded by private or federal funds. However, this argument presumes that the state does not have a proprietary interest in the efficiency of construction projects financed by private or federal money. Using such funds efficiently allows for more to be done with limited funding and increases the likelihood that more projects will be funded. This is a proprietary interest that the legislation directly furthers. Moreover, even private developers could reasonably impose such restrictions notwithstanding the presence of outside funding. In contrast, by focusing on state action instead of state funding, the act allows private projects receiving state grant funds to use PLAs if those private entities so desire. See 2012 Mich. Pub Acts 288 § 8 (codified at Mich. Comp. Laws § 408.878). This is a limit President George W. Bush’s executive order did not contain, see Exec. Order No. 18,302, 66 Fed.Reg. 11225 (Feb. 17, 2001), and it further demonstrates that the statute’s goal is efficiency in government construction and not the wholesale elimination of PLAs.

The proprietary nature of the act is directly supported by the Supreme Court’s holding in Building & Construction Trades Council v. Associated Builders & Contractors, 507 U.S. 218, 113 S.Ct. 1190, 122 L.Ed.2d 565 (1993) (Boston Harbor). In Boston Harbor, the Court held that the government-mandated requirement of a PLA for the Boston Harbor construction project was proprietary. In Boston Harbor, the Massachusetts Water Resources Authority enforced a PLA negotiated by its agent and a local trades council in an effort to avoid delays from labor disputes. A contractors’ association complained, arguing that the agreement violated the NLRA. The Supreme Court held that preemption did not apply because the Authority’s action was proprietary. “When a State owns and manages property,” the Court explained, “it must interact with private participants in the marketplace. In so doing, the State is not subject to preemption by the NLRA, because pre-emption doctrines apply only to state regulation.” Id. at 227, 113 S.Ct. 1190 (emphasis in original). The Court noted that “[t]o the extent that a private purchaser may choose a contractor based upon that contractor’s willingness to enter into a prehire agreement, a public entity as purchaser should be permitted to do the same.” Id. at 231, 113 S.Ct. 1190 (emphasis in original).

That lesson must be applied to this case. Just as a private purchaser can choose not to enter into PLAs, believing them to be inefficient, a state legislature, sharing that same belief, can decide that public money should not to be used for PLA projects. In Boston Harbor, the Authority “was attempting to ensure an efficient project that would be completed as quickly and effectively as possible at the lowest cost.” Id. at 232, 113 S.Ct. 1190. The Michigan legislature has stated a similar goal. Although the legislature acted on all projects at once, unlike the Authority in Boston Harbor, that is not alone sufficient to make an action regulatory. While some, such as Senator Gleason, may argue that banning the use of PLAs actually sacrifices speed and efficiency in favor of cost savings, it is not the job of this court to question whether the legislature made the correct decision. It is a permissible decision.

This type of action is materially different from the challenged action in Wisconsin Department of Industry v. Gould, 475 U.S. 282, 106 S.Ct. 1057, 89 L.Ed.2d 223 (1986). In that case, Wisconsin maintained a list of repeat NLRA-violators and forbade state procurement agents from doing business with any person or firm on that list. Id. at 283-84, 106 S.Ct. 1057. The Court held that the statute requiring this practice was regulatory and preempted by the NLRA, explaining that “on its face the debarment statute serves plainly as a means of enforcing the NLRA.... [Tjhe point of the statute is to deter labor law violations and to reward ‘fidelity to the law.’ ” Id. at 287, 106 S.Ct. 1057. Unlike in Boston Harbor, “[n]o other purpose could credibly be ascribed” to the statute. Id. As the Court explained in Boston Harbor, the problem with Wisconsin’s action in Gould was that it constituted “a state agency’s attempt to compel conformity with the NLRA.” 507 U.S. at 228, 113 S.Ct. 1190. The statute “addressed employer conduct unrelated to the employer’s performance of contractual obligations to the State, and ... the State’s reason for such conduct was to deter NLRA violations.” Id. at 228-29, 113 S.Ct. 1190. Here, the act purportedly conserves state resources. The state did not restrict its choices to promote the enforcement of a different labor policy. This act is closer to the action the Court found proprietary in Boston Harbor than to the statute found to be regulatory in Gould.4

This court’s precedent demonstrates this dividing line. In Petrey v. City of Toledo, we held that provisions of a city towing ordinance that chose certain tow companies for police towing were proprietary because they “do not constitute attempts on the part of the City to regulate the towing industry as a whole, or to advance some general societal goal.” 246 F.3d 548, 558-59 (6th Cir.2001), abrogated on other grounds by City of Columbus v. Ours Garage and Wrecker Serv., Inc., 536 U.S. 424, 122 S.Ct. 2226, 153 L.Ed.2d 430 (2002). Instead, the provisions were meant to ensure administrative efficiency and city monitoring. Id. at 558. Similarly, the Michigan act does not regulate the construction industry as a whole and does not aim to eliminate PLAs altogether. Instead, it is meant to ensure efficiency and save taxpayer money.

Other circuits’ decisions support this analysis. Building and Construction Trades Department v. Allbaugh, 295 F.3d 28 (D.C.Cir.2002), is directly on point. Allbaugh was a challenge to President George W. Bush’s ban on federal agencies’ requiring PLAs for federally funded projects. See Exec. Order No. 13,302, 66 Fed.Reg. 11225 (Feb. 17, 2001). The D.G. Circuit considered both Boston Harbor and Gould and found that the challenged executive order constituted proprietary action. The D.C. Circuit held the relevant distinction to be whether the government “ ‘acts just as a private contractor would act’ ... or instead seeks to affect conduct ‘unrelated to the employer’s performance of contractual obligations.’ ” 295 F.3d at 34-35 (quoting Boston Harbor, 507 U.S. at 233, 229, 113 S.Ct. 1190). The court noted that “the Government unquestionably is the proprietor of its own funds, and when it acts to ensure the most effective use of those funds, it is acting in a proprietary capacity.” Id. at 35. The D.C. Circuit also addressed the same case-by-case vs. blanket-rule argument that the councils make here, and concluded that “there simply is no logical justification for holding that if an executive order establishes a consistent practice regarding the use of PLAs, it is regulatory even though the only decisions governed by the executive order are those that the federal government makes as a market participant.” Id. (internal quotation marks omitted). “Because the Executive Order does not address the use of PLAs on projects unrelated to those in which the Government has a proprietary interest, the Executive Order establishes no condition that can be characterized as ‘regulatory.’ ” Id. at 36. The executive order at issue in Allbaugh is almost identical to the Michigan act in its effect on PLAs.

Allbaugh is the closest parallel to this case. The councils argue that Allbaugh is no longer good law; however, the cases they point to have not criticized Allbaugh and each is distinguishable. UAW-Labor Employment and Training Corp. v. Chao, 325 F.3d 360 (D.C.Cir.2003), was a challenge to an executive order requiring contractors who obtain large government contracts to post notices at all facilities informing employees of their rights not to join a union or pay certain union dues. The D.C. Circuit held the order to be regulatory instead of proprietary. The challenged order, like the statute in Gould and unlike the authority’s action in Boston Harbor and the order in Allbaugh, had no effect on government resources. Its sole purpose was to use government contracts as a tool to promote a certain labor policy. The government did not even explicitly argue that its actions were proprietary. 325 F.3d at 363. UAW-Labor was a straightforward application of Gould and does not say anything about the continued validity of Allbaugh.

Precedent from other circuits follows this same pattern. The Fifth Circuit, considering Gould and Boston Harbor, held that a city’s grant of an exclusive towing contract to a company was proprietary. Cardinal Towing & Auto Repair, Inc. v. City of Bedford, 180 F.3d 686, 693 (5th Cir.1999). The city’s action ensured efficient towing, clarified responsibility, minimized administrative confusion, and allowed for easy supervision. Furthermore, there was no indication that the city’s goal was to encourage private towing companies to conduct their business in a specified manner. Id. Likewise, the Third Circuit has noted that the “pivotal difference” between Gould and Boston Harbor “is that in the former case the state deployed its spending authority to achieve a goal far broader than merely protecting or fostering its own investment or proprietary interest, while in the latter instance the public agency limited its spending conditions to the protection of its investment or proprietary interest.” Hotel Emps. & Rest. Emps. Union, Local 57 v. Sage Hospitality Res., 390 F.3d 206, 214 (3d Cir.2004). The Third Circuit held that an ordinance conditioning tax-incentive funding on an employer’s acceptance of a labor neutrality agreement was intended to promote and protect the city’s interest in tax revenues from the project and was therefore proprietary. Id. at 217. Notably, in reaching that decision, the court considered All-baugh in its survey of other circuits, and did not criticize it. Id. at 215. Finally, the' Seventh Circuit held a Milwaukee ordinance requiring city contractors to reach “labor peace agreements” with unions to be regulatory because the ordinance would have significant spillover effects on private contracts and because the city passed up easy alternatives to achieve the same desired effect without the spillover effects. Metro. Milwaukee Ass’n of Commerce v. Milwaukee Cnty., 431 F.3d 277, 280 (7th Cir.2005). These cases all draw a similar line: to be proprietary, the government action must be aimed to achieve a proprietary goal and must be limited to the furtherance of that goal. The Michigan statute falls on the proprietary side of that line.

Michigan’s statute advances the proprietary interest of efficient use of resources and is limited enough to advance only that interest. Accordingly, it is proprietary, and not regulatory, and therefore is not preempted by the NLRA.

In case No. 12-1246, the appeal is dismissed as moot. In case No. 12-2548, the district court’s judgment is reversed and the injunction is vacated.

1 . These reports include Vazques, Glaser & Bruvold, Measuring the Cost of Project Labor Agreements on School Construction in California, available at http://www.nusinstitute.org/ assets/resources/pageResources/Measuring-the-Costof-Project-Labor-Agreements-on-School-Construction-in-California.pdf (last visited August. 08, 2013) (finding that PLA school construction projects cost 13% more per square foot than comparable non-PLA projects); Tuerck, Why Project Labor Agreements Are Not in the Public Interest, 30 CATO J. 45 (Winter 2010), available at http://www. cato.org/sites/cato.org/files/serials/files/cato-journal/2010/l/cj30nl-3.pdf (last visited August 08, 2013) ("By one estimate, PLAs add 12-18 percent to the cost of public projects”); and Tuerck, Glassman & Bachman, Project Labor Agreements on Federal Construction Projects: A Costly Solution in Search of a Problem, available at http://www.beaconhill. org/BHIStudies/PLA2009/PLAFinal090923. pdf (last visited August 08, 2013) (noting that PLAs increased construction costs on state school construction projects).

2 . The councils cite, among others, Dale Bel-man and Matthew M. Bodah, Building Better: A Look at Best Practices for the Design of Project Labor Agreements (Economic Policy Institute Briefing Paper #274, Aug. 10, 2010), available at http://epi.3cdnnet/179fd 74170130cd540_ibm6ib3kd.pdf (last visited August 08, 2013) (discussing benefits of PLAs, including quick resolution of labor disputes and improved working conditions); Dale Bel-man et al., Project Labor Agreements’ Effects on School Construction Costs in Massachusetts, 49 Indus. Rel. 44 (2010), available at https://www.msu.edu/3rdale/Publications/ Construction%20&%20PLAs/Project %20Labor%20Agreements’%20Effect%20on %20School%20Construction%20Costs%20IR %20-%20Copy.pdf (last visited August 08, 2013) (arguing that PLAs do not increase construction costs).

3 . During oral argument, counsel for the Governor stated that a governmental unit could not choose a contractor who had a PLA if it had put in the lowest bid for a project. The Governor later clarified that that reading of the law is incorrect. The opposite is true—if a contractor with a PLA submits the lowest bid, it would be a violation of the act for the governmental entity not to choose that contractor because of the existence of the PLA.

4 . Chamber of Commerce v. Brown, 554 U.S. 60, 128 S.Ct. 2408, 171 L.Ed.2d 264 (2008), is like Gould. The Supreme Court held to be pre-empted a California statute ("AB 1889”) prohibiting employers that receive state grants or program funds from using the funds "to assist, promote, or deter union organizing,” and establishing a "formidable” scheme to enforce the statute. Id. at 63, 128 S.Ct. 2408. The Supreme Court reasoned that California had acted in its regulatory capacity not only because the statute was not “specifically tailored to one particular job” but also because the statute was not a "legitimate response to state procurement constraints or to local economic needs,” id. at 70, 128 S.Ct. 2408. citing Gould. In particular,

the legislative purpose is not the efficient procurement of goods and services, but the furtherance of a labor policy. Although a State has a legitimate proprietary interest in ensuring that state funds are spent in accordance with the purposes for which they are appropriated, this is not the objective of AB 1889. In contrast to a neutral affirmative requirement that funds be spent solely for the purposes of the relevant grant or program, AB 1889 imposes a targeted negative restriction on employer speech about unionization. Furthermore, the statute does not even apply this constraint uniformly. Instead of forbidding the use of state funds for all employer advocacy regarding unionization, AB 1889 permits use of state funds for select employer advocacy activities that promote unions.

Id. at 70-71, 128 S.Ct. 2408 (citations omitted). The Michigan statute at issue in this case has nothing like these facially regulatory aspects of AB 1889.