HOWARD, Circuit Judge.

In October 2011, the State of Maine enacted L.D. 1326, "An Act To Allow School Administrative Units To Seek Less Expensive Health Insurance Alternatives," pursuant to which health insurers must disclose, upon written request from a public school district, aggregate loss information pertaining to any group policies held by the district's employees.1 L.D. 1326, 125th Leg., 1st Reg. Sess. (Me. 2011). Shortly thereafter, plaintiff-appellant Maine Education Association Benefits Trust (the "Trust") — which manages a statewide health insurance plan for a substantial segment of Maine's public school work force — filed suit in the district court, seeking to permanently enjoin the law prior to its enforcement. The Trust alleged, inter alia, that because its loss information constitutes a confidential trade secret, the Act's disclosure requirement results in an uncompensated taking proscribed by the Fifth Amendment. The district court denied the Trust's motion for a preliminary injunction, and the Trust now challenges that denial in this timely appeal. After careful consideration, we affirm.

I.

A brief overview of the Trust, its health insurance plan, and the statutory scheme at issue is necessary for an understanding of the claims presented on appeal. The district court's thoughtful and comprehensive order, see Maine Educ. Ass'n Benefits Trust v. Cioppa, 842 F.Supp.2d 386 (D.Me. 2012), contains a detailed exposition of the undisputed facts, from which we borrow.

A. The Trust and Its Health Insurance Plan

Since 1993, the Trust has provided health insurance to the bulk of Maine's public school employees and their dependents through a plan underwritten by various insurers, most recently Anthem Blue Cross Blue Shield of Maine ("Anthem"). The insurance plan (the "Plan"), which currently covers nearly 67,000 members from 99 percent of Maine's school districts, is community-rated; that is, the price of coverage is negotiated on the basis of group-wide utilization costs, and accounts for neither geographic variation nor an individual employer's demographic mix, prior utilization, or loss experience. This community-rated plan is designed in part to subsidize, through members who are actuarially favorable, the premiums paid by members who are actuarially less attractive to insurers. The Plan as designed economically benefits employees of educational institutions whose work forces are older or less healthy than other members of the group, or who reside in regions — typically Northern and Eastern Maine — with higher health care costs and, on average, lower salaries than their Southern Maine counterparts. The Plan is thus structured, in part, to help mitigate this disparity.

Eligibility for enrollment in the Plan is determined by the collective bargaining agreements negotiated between local bargaining units and individual employers, predominantly school districts. The employees of a given school district are eligible to participate in the Plan if the largest collective bargaining unit in that district is represented by the Maine Education Association ("MEA"), the statewide teachers' union that founded the Trust. Once eligibility has been established, the school board and the employees decide together, by a collaborative vote, whether the employees will be offered the Plan. Those who receive the offer and elect to enroll do so directly with the Plan's insurer, Anthem. The Trust has no contracts with individual educational institutions, and those institutions are not considered to be sponsors of the Plan. Rather, based on the amount agreed to in its own collective bargaining agreement, each school district pays to the Plan a percentage of its employees' health insurance premiums, and the employees are responsible for the remainder.

For the most recent plan year for which there was evidence at the preliminary injunction hearing, the Plan's annual premium was nearly $370,000,000, resulting in an average monthly cost of approximately $460 per member. The Trust itself maintains a reserve fund that, according to its last available audit, held in excess of $87,000,000. The Trust uses the reserve fund to buy down rate increases, thereby avoiding inflation in the monthly cost charged to Plan participants.

B. The Statutory Scheme

Similar to the other states, Maine heavily regulates the insurance industry. "Title 24-A, the `Maine Insurance Code,' has 85 separate chapters and fills almost two complete volumes of the Maine Revised Statutes Annotated. There is [also] an entire department of State Government, the Insurance Bureau, devoted to regulating the business of insurance." Lessard v. Allstate Ins. Co., No. Civ. A. cv-98-162, 2001 WL 1712653, at *4 (Me.Super.Ct. March 12, 2001).

Under Maine's regime, the state has historically compelled the disclosure of basic "loss information," generally defined as the aggregate claims experience of a given policy, or more specifically, the ratio of premiums charged to claims paid — in short, a simple equation, and the applicable data, for deciphering the purchaser's own healthcare costs. For example, Maine law has entitled holders of property and casualty insurance policies to obtain such loss information from their carriers since 1989, see 1989 Me. Laws, Ch. 696, "An Act Requiring the Availability of Insurance Loss Information" (codified at Me. Rev. Stat. Ann. tit. 24-A, §§ 2910, 3042). In 1995, this access was expanded to cover current and former policyholders of group and blanket health insurance, see 1995 Me. Laws, Ch. 71, "An Act to Require Insurance Companies to Provide Loss Information to Insured Groups" (codified at Me. Rev. Stat. Ann. tit. 24-A, § 2803-A) ("Disclosure of Loss information. Upon written request, every insurer shall provide loss information concerning a group policy or contract to its policyholder ...." (emphasis added)). L.D. 1326 is incidentally more expansive, extending the right of access not just to "policyholders," but also to the small subgroup of public school districts whose employees are enrolled in a shared health insurance policy.

In seeking preliminary injunctive relief, the Trust mounted challenges to the two provisions of the Act that deal with loss information, codified at 20-A M.R.S.A. § 1001(14)(E) and 24-A M.R.S.A. § 2803-A(2). As amended, 20-A M.R.S.A. § 1001(14)(E) states, in pertinent part:

24-A M.R.S.A. § 2803-A(2) reads, in turn:

The "loss information" at the center of this controversy is statutorily defined as "the aggregate claims experience of the group insurance policy or contract," including "the amount of premium received, the amount of claims paid[,] and the loss ratio," but not including "any information or data pertaining to the medical diagnosis, treatment[,] or health status that identifies an individual covered under the group contract or policy." Me. Rev. Stat. Ann. tit. 24-A, § 2803-A(1)(B).

C. The Trust's Past Treatment of Loss Information

The Trust has structured itself, as well as its relationships with insurers, to keep the Plan's loss information confidential. Since the Trust's formation in 1993, its Trust Agreement has provided, in relevant part:

Additionally, since 2005, the Trust's Group Agreements with Anthem have included the following confidentiality clause:

In accordance with these provisions, the Trust has always considered the experience ratings and claims history of individual school districts to be a proprietary, confidential trade secret. The information is not conveyed to anyone outside of Anthem (including to members of the Trust itself), and the Trust has undertaken reasonable efforts to ensure the information's confidentiality by neither collecting it nor allowing Anthem to use or release it.2

A driving force behind these efforts at maintaining confidentiality is to prevent actuarially desirable districts from acquiring the information and leaving the community-rated plan for less expensive individual coverage, thereby increasing the monthly costs for the Plan's remaining members (or, alternatively, forcing the Trust to expend additional resources from its reserve fund to pay down the resulting rate increases). Prior to the passage of L.D. 1326, the Trust was able to do exactly that, denying multiple school district requests for loss information over the years by virtue of the fact that the Trust, and not the individual school districts, was always the primary "policyholder." See 1995 Me. Laws, Ch. 71 (requiring disclosure of loss information concerning a group health insurance policy or contract only to "its policyholder").

Generally, group insurers in Maine require at least two years of aggregate loss information from an employer in order to provide a quote. Thus, without access to loss information, school districts cannot meaningfully explore insurance options outside those offered by the Trust. Following the enactment of L.D. 1326, at least one district has already submitted a renewed request to Anthem. In the absence of an injunction, it is apparent that school districts will continue to request access in order to obtain competing quotes based on that loss information. Depending on the quotes and the terms of each district's collective bargaining agreement, this process may or may not result in some districts withdrawing their employees from the Plan. In most cases, if not all of them, negotiating changes to the respective collective bargaining agreements is a lengthy process that may take months, if not years, to complete.

D. The Proceedings Below

The Trust filed suit in the district court shortly after L.D. 1326 was enacted. Its amended complaint claimed that the Act was preempted by federal law, and that it ran afoul of several constitutional impediments, including the Contracts Clause, the Takings Clause, and the Due Process Clause. The district court resolved the preemption question, as well as the impairment of contract and due process claims, in favor of the defendants-appellees, and it denied the Trust's subsequent request for a preliminary injunction, finding primarily that the Trust was not likely to succeed on the merits of its only remaining claim — that the Act works as an unconstitutional taking. This interlocutory appeal ensued, in which the Trust challenges only the denial of its request for a preliminary injunction on takings grounds.3 We have jurisdiction under 28 U.S.C. § 1292(a)(1), and on March 30, 2012, we entered a temporary injunction prohibiting the enforcement of L.D. 1326 pending appeal.

II.

In considering a request for a preliminary injunction, a trial court must weigh several factors: (1) the likelihood of success on the merits; (2) the potential for irreparable harm to the movant in the absence of an injunction; (3) the balance of the movant's hardship if relief is denied versus the nonmovant's hardship if the relief is granted; and (4) the effect, if any, of the decision on the public interest. Ross-Simons of Warwick, Inc. v. Baccarat, Inc., 102 F.3d 12, 15 (1st Cir.1996). Of these four factors, the movant's likelihood of success "is the touchstone of the preliminary injunction inquiry." Philip Morris, Inc. v. Harshbarger, 159 F.3d 670, 674 (1st Cir.1998) ("Philip Morris I"). "[I]f the moving party cannot demonstrate that he is likely to succeed in his quest, the remaining factors become matters of idle curiosity." New Comm Wireless Servs., Inc. v. SprintCom, Inc., 287 F.3d 1, 9 (1st Cir.2002).

We will set aside a district court's ruling on a request for a preliminary injunction only if the court clearly erred in assessing the facts, misapprehended the applicable legal principles, or otherwise is shown to have manifestly abused its discretion. Cohen v. Brown University, 991 F.2d 888, 902 (1st Cir.1993). This standard requires "a party who appeals from the issuance [or denial] of a preliminary injunction [to] bear[] the considerable burden of demonstrating that the trial court mishandled the fourpart framework." Philip Morris I, 159 F.3d at 674 (second and third alterations in original) (quoting Ross-Simons, 102 F.3d at 16) (internal quotation marks omitted). In this case, our analysis begins and ends with the evaluation of the often dispositive factor: whether the Trust has a reasonable likelihood of success on the merits of its takings claim.

A. Likelihood of Success on the Merits

The Takings Clause of the Fifth Amendment, which applies to the states through the Fourteenth Amendment, prohibits the taking of private property for public use without just compensation. Lingle v. Chevron U.S.A. Inc., 544 U.S. 528, 536, 125 S.Ct. 2074, 161 L.Ed.2d 876 (2005). This prohibition extends not only to the paradigmatic physical taking — i.e., where the government condemns or physically appropriates a person's property — but also to regulatory interferences, which transpire "when some significant restriction is placed upon an owner's use of his property for which `justice and fairness' require that compensation be given." Philip Morris, Inc. v. Reilly, 312 F.3d 24, 33 (1st Cir.2002) (en banc) ("Philip Morris II").4

The dichotomy between physical and regulatory takings is critical, for it often determines the level of scrutiny that a challenged government action will receive. In contrast to the law of physical takings, which involves, for the most part, the "straightforward application of per se rules," the Supreme Court's regulatory takings jurisprudence has eschewed any bright-line formulations. Tahoe-Sierra Pres. Council, Inc. v. Tahoe Reg'l Planning Agency, 535 U.S. 302, 322, 122 S.Ct. 1465, 152 L.Ed.2d 517 (2002). To assess the propriety of a regulatory takings claim, the Court has instead enumerated a more nuanced, three-pronged inquiry into (1) the extent to which the regulation interferes with the claimant's reasonable investment-backed expectations; (2) the regulation's economic impact on the property owner; and (3) the character of the government action. Penn Cent. Transp. Co. v. City of New York, 438 U.S. 104, 124, 98 S.Ct. 2646, 57 L.Ed.2d 631 (1978). Designed to facilitate a careful examination and weighing of all the relevant circumstances, the context-sensitive "Penn Central" factors operate not as a "checklist of items that can be ticked off as fulfilled or unfulfilled," but rather as "lenses through which a court can view and process the facts of a given case." Philip Morris I, 159 F.3d at 674.

The Supreme Court has recognized two circumstances under which a regulatory action may justify bypassing the Penn Central factors in favor of per se rules. First, where a regulation inflicts a permanent physical invasion of private property — however minor — the government must provide just compensation. Lingle, 544 U.S. at 538, 125 S.Ct. 2074; see, e.g., Loretto v. Teleprompter Manhattan CATV Corp., 458 U.S. 419, 102 S.Ct. 3164, 73 L.Ed.2d 868 (1982) (finding a taking where a state law required landlords to permit cable companies to install cable facilities in apartment buildings). A second categorical rule applies to "regulations that completely deprive an owner of all economically beneficial use of her property." Lingle, 544 U.S. at 538, 125 S.Ct. 2074 (internal quotation marks and alterations omitted); see, e.g., Lucas v. S.C. Coastal Council, 505 U.S. 1003, 1019, 112 S.Ct. 2886, 120 L.Ed.2d 798 (1992) (holding that a regulation prohibiting the erection of any permanent habitable structures constituted a taking). The allegation here — that L.D. 1326's disclosure requirement impinges upon the Trust's purported trade secret — encompasses neither of these scenarios. Accordingly, and because the challenged government action does not directly appropriate the Trust's property, we proceed to analyze L.D. 1326 under the law of regulatory takings.5

i. Reasonable Investment-Backed Expectations

Although generally recognized as the figurative ballast of the Penn Central framework, "reasonable investment-backed expectations" is a concept that can be difficult to define more concretely. See generally J. David Breemer & R.S. Radford, The (Less?) Murky Doctrine of Investment-Backed Expectations After Palazzolo, and the Lower Courts' Disturbing Insistence on Wallowing in the Pre-Palazzolo Muck, 34 S.W.U. L. Rev. 351 (2005). Nevertheless, one very general contour is clear: "Courts [will] only protect [a claimant's] reasonable expectations." Philip Morris II, 312 F.3d at 36. Thus, in considering whether the Trust possesses the requisite expectations to support a takings claim, the inquiry must acknowledge that "not every investment deserves protection and ... some investors inevitably will be disappointed." Id. The question reduces to whether the Trust, given all of the attendant facts and circumstances, has a probability of success of showing that it had a reasonable expectation that the Plan's loss information will be kept confidential. We think not.

As a baseline proposition, the Trust's expectations are substantially diminished by the highly regulated nature of the industry in which it operates. See Franklin Mem'l Hosp. v. Harvey, 575 F.3d 121, 128 (1st Cir.2009) (holding that a claimant's investment-backed expectations were "tempered by the fact that it operate[d] in the highly regulated hospital industry"); see also Ruckelshaus v. Monsanto Co., 467 U.S. 986, 1007, 104 S.Ct. 2862, 81 L.Ed.2d 815 (1984) (noting that expectations are necessarily adjusted in areas that "ha[ve] long been the source of public concern and the subject of government regulation"). Given the historically heavy and continuous regulation of insurance in Maine, see Lessard, 2001 WL 1712653, at *4, the Trust, in choosing how and where to allocate its resources, ought to at least be aware of the heightened possibility that new insurance regulations might hinder the use or value of its loss information, see Lucas, 505 U.S. at 1027-28, 112 S.Ct. 2886.

This is particularly true where, as here, the extensive regulatory framework in place prior to the passage of the challenged legislation has consistently regulated the type of property interest for which the claimant seeks constitutional protection. See, e.g., Connolly v. Pension Ben. Guar. Corp., 475 U.S. 211, 226, 106 S.Ct. 1018, 89 L.Ed.2d 166 (1986) (finding no reasonable investment-backed expectations where the property interest at issue had been the "object[] of legislative concern long before the passage of [the challenged legislation]"). Since 1989, holders of property and casualty insurance in Maine have been able to obtain their policies' loss information. See 1989 Me. Laws, Ch. 696. In 1995 — not long after the Trust was formed, and well before it began to insist on the inclusion of confidentiality provisions in its Group Agreements — this right of access to loss information was expanded to policyholders of group and blanket health insurance like the Plan at issue here. See 1995 Me. Laws, Ch. 71. Thus, for years the Trust was spared the obligation of disclosure solely because of the Plan's unique structure, whereby the Trust, and not the school district, was the technical policyholder. Throughout this extended period, several of the state's school districts were clamoring for loss information, and ultimately, for the legislature to close this perceived loophole. See Appolo Fuels, Inc. v. United States, 381 F.3d 1338, 1349 (Fed.Cir.2004) (suggesting that a key aspect of the investment-backed expectations inquiry is the claimant's awareness of "the problem that spawned the [challenged] regulation"). Under these circumstances, the prospect of the legislature's continued expansion of this right of access was reasonably foreseeable.6

The Trust, in challenging this conclusion, relies principally on the distinction between "policyholders" and "non-policyholders." It contends that because L.D. 1326 grants non-policyholders (the school districts) access to confidential policyholder information, it represents an "extraordinary and unprecedented" regulatory shift. We disagree.

As a preliminary matter, requiring the disclosure of otherwise private information to non-policyholders is not a novel concept in Maine insurance law. See, e.g., Me. Rev. Stat. Ann. tit. 24-A, § 2210 (requiring insurers to provide personal data to claimants and insureds over a policyholder's objection); Me. Code R. 90 590 243 §§ 2, 5 (2003) (requiring insurers to provide claims data to the government over a policyholder's objection).

Moreover, the significant role that school districts play in the Plan's overall scheme is important to the analysis. While it is true that the districts do not formally sign the insurance agreements on behalf of their employees, and therefore are not the formal sponsors of the Plan, they are otherwise integrally involved in just about every step of the process, from negotiating the collective bargaining agreements to organizing and participating in collaborative votes to determine whether their employees may enroll in the Plan in the first instance. In the event that loss information is to be disclosed for the purpose of exploring alternative district-specific coverage, the district itself is the obvious choice to assume "policyholder" status. Indeed, the Act forecloses the type of "unprecedented" access argued by the Trust, by restricting disclosure to loss information that concerns the inquiring district's own employees, retirees, and their dependents; it does not grant access to the Trust's policy-wide loss ratios. See Me. Rev. Stat. Ann. tit. 20-A, § 1001(14)(E). Thus, the districts are not, as the Trust suggests, mere "strangers to the insurance contract" in the ordinary sense, and L.D. 1326 is not at odds with Maine's existing body of insurance law.7

While this conclusion is largely dictated by precedent, it is also rational from a policy standpoint. The motivation behind the State's previous attempts to regulate loss information was clear: to prevent sellers of group health insurance from acquiring disproportionate leverage over buyers by ensuring that both parties have access to the relevant cost information. The construct created and advocated by the Trust, under which such information cannot be released to "non-policyholders," effectively precludes any member school district from opting out of the Plan, and thus is contrary to the regulation's intent by denying school districts the benefit of equal access.

We recognize that the Trust's purpose is to spread health risk among a larger population, thereby bringing insurance costs down for districts in which enrollees, on average, are less healthy and where health care costs are relatively high. Nonetheless, the fact remains that the legislation is highly consistent with legitimate policy objectives. L.D. 1326 simply continues what the 1989 and 1995 regulations started by addressing a unique scenario which, in all likelihood, was not contemplated by the original legislation. Cf. Connolly, 475 U.S. at 227, 106 S.Ct. 1018 ("Those who do business in the regulated field cannot object if the legislative scheme is buttressed by subsequent amendments to achieve the legislative end.") (quoting FHA v. The Darlington, Inc., 358 U.S. 84, 91, 79 S.Ct. 141, 3 L.Ed.2d 132 (1958)).

The Trust's remaining arguments on this issue require little discussion. It attempts to treat the aforementioned facts in isolation, asserting that none, standing alone, was sufficient to provide the Trust with constructive notice of L.D. 1326. It may be that neither the fact that the insurance industry is highly regulated nor the fact that school districts routinely sought access to the seemingly confidential loss information, standing alone, sufficiently undermines the reasonableness of the Trust's expectations. But taken together, and in light of the State's history of aggressively regulating the use and disclosure of loss information, L.D. 1326 is a relatively minor expansion which the Trust could, and should, have anticipated.

We do not discount the fact that, since its inception, the Trust has gone to lengths to preserve the confidentiality of its purported trade secret. Its Trust Agreements, as well as its post-2004 Group Agreements with Anthem, unequivocally reserve to the Trust all rights respecting the Plan's loss information, and the Trust has enforced those rights in the face of repeated school district inquiries. The Trust therefore has demonstrated an expectation of confidentiality. But unilateral expectations, no matter how adamantly pursued, are not enough. See Monsanto, 467 U.S. at 1005-06, 104 S.Ct. 2862. The expectation must be a reasonable one, and on the record before us, we cannot conclude that the plaintiffs have shown a probability of success on this issue.8

ii. Economic Impact

Evaluating the magnitude of the economic impact of regulatory action ordinarily requires an assessment of the extent to which the action "impairs the value or [typical] use" of the property. PruneYard Shopping Ctr. v. Robins, 447 U.S. 74, 83, 100 S.Ct. 2035, 64 L.Ed.2d 741 (1980). In this regard, a result of the law's requirement that loss information be disclosed to districts upon request is that the Trust may be less able to prevent districts with favorable claims experience from shopping for alternative coverage. Upon acquiring their respective loss information, some actuarially favorable school districts may leave the Plan for less expensive district-specific coverage, resulting in relatively higher premiums for the Plan's remaining members (or greater expenditures by the Trust to pay those rate increases down). We do not minimize this possible impact, and we acknowledge the Trust's defensible goal of making insurance more affordable for those in areas where health care costs are higher and teacher salaries are lower.

The likelihood and severity of the exodus of members, however, are completely unknown. The only concrete evidence on the issue includes the affidavits of the parties' dueling actuaries who debate, at a high level of generality, the extent to which L.D. 1326 will undermine the Trust's prospective viability. To wit, the Trust's actuarial expert, having been denied access to the district-specific loss ratios,9 was able to conclude only that "if individual school districts purchase insurance on their own, some groups with lower average age, more favorable utilization history, or a lower cost of geographic area might be able to take advantage of that more favorable profile and obtain insurance from Anthem or another carrier at lower cost than they currently pay to participate in the ... [P]lan." This evidence merely serves to highlight the conjectural nature of the Act's economic impact, and is "not sufficiently concrete to establish a taking." In re Jones Truck Lines, Inc., 57 F.3d 642, 651 (8th Cir.1995) (finding that economic impact was too speculative to support a takings claim); see also Tenn. Scrap Recyclers Ass'n v. Bredesen, 556 F.3d 442, 456 (6th Cir.2009) (finding that the economic impact of the challenged regulation did not support the movant's takings claim, in part because "the [economic] impact of [the regulation was] speculative").

Whether or not at trial, and with a more developed evidentiary foundation, the Trust may be able to prove the requisite economic impact, it has not yet done so sufficiently to show a probability of success on the merits.

iii. Character of the Government Action

The third and final consideration in the regulatory takings analysis — the character of the government action — also weighs against the Trust's takings claim. Under Penn Central, "[a] `taking' may more readily be found when the interference with property can be characterized as a physical invasion by government than when [the] interference arises from some public program adjusting the benefits and burdens of economic life to promote the common good." Id. (citation omitted). L.D. 1326 clearly falls on the latter end of the spectrum, reflecting the legislature's judgment that allowing school districts to access their employees' loss information will promote the common good by creating a wider array of competitively priced group health insurance options. See Me. Rev. Stat. Ann. tit. 20-A, § 1001(14)(E) (enabling school districts to request their loss information "[i]n order to facilitate the competitive bidding process in procuring [group] health insurance").

The Trust argues that even if the Act is intended to promote the common good, it does so improperly by reallocating to the Trust alone what is essentially a public burden. Claiming that it has been effectively "singl[ed] out" by the Maine legislature, the Trust directs us to five comments in the Act's legislative record which, more or less, appear to cast the Trust in a negative light.

The argument, however, is unavailing. L.D. 1326 does not apply solely to the Trust; it applies to every existing or future multi-employer group health insurance plan in which the State's public school districts choose to enroll. That the Trust attracted the attention of a handful of legislators, and currently bears the brunt of the Act's burden, is merely a byproduct of its holding the predominant share of the targeted market — a virtual monopoly, perpetuated by the very policy of non-disclosure which it seeks to protect. At the preliminary injunction stage, the Trust has not shown that the Act attempts to impose on the Trust an excessive burden that should in fairness be borne by other entities or by society as a whole. Cf. E. Enter. v. Apfel, 524 U.S. 498, 537, 118 S.Ct. 2131, 141 L.Ed.2d 451 (1998) (holding that the character of a state regulation supported the movant's takings claim only where it "single[d] out certain employers to bear a burden that is substantial in amount, based on the employers' conduct far in the past, and unrelated to any commitment that the employers made or to any injury that they caused"). Thus, the third Penn Central factor, like the two before it, counsels against finding that a taking has occurred.

We emphasize that because we hear this matter on appeal from the denial of a preliminary injunction, our likelihood-of-success determinations are to be understood only as probable outcomes. See Cohen, 991 F.2d at 902. At the upcoming trial, the Trust will have the opportunity to demonstrate more concretely and comprehensively the economic impact that it fears, namely, the withdrawal of school districts with better claims experience and the resulting increased cost of health insurance coverage for the Trust's remaining members. Based on the present state of the record, however, we cannot conclude that the Trust is likely to succeed on the merits.

III.

Because the plaintiffs have not established a likelihood of success, and such a showing is essential to the issuance of a preliminary injunction, see Philip Morris I, 159 F.3d at 674, it would serve no useful purpose to explore the remaining three facets of the preliminary injunction framework. We conclude that there was no abuse of discretion in the denial of the Trust's motion for a preliminary injunction, and leave to the trial court to determine whether it should stay enforcement of the Act pending hearing evidence and decision on the merits. The decision of the district court is affirmed. Each side shall bear its own costs of this appeal.