Unquestionably, tobacco is subject to heavy regulation by federal and state governments. This case concerns one attempt, by Massachusetts, to further regulate tobacco products by requiring tobacco companies to submit to Massachusetts the ingredient lists for all cigarettes, snuffs, and chewing tobaccos sold in the state. For each brand, the manufacturer must list, by relative amount, all ingredients besides tobacco, water, or reconstituted tobacco sheet. Mass. Gen. Laws ch. 94, § 307B (2002). Currently, the appellees, a group of tobacco companies, treat these ingredient lists as trade secrets and either do not disclose brand-specific information at all or do not disclose it without some guarantee of confidentiality.

The tobacco companies brought suit claiming that the Massachusetts statute, which allows the public disclosure of these ingredient lists whenever such disclosure "could reduce risks to public health," Mass. Gen. Laws ch. 94, § 307B, creates an unconstitutional taking. Appellees also argued that the Massachusetts statute violates their Due Process rights by effecting a taking of their property without first providing a meaningful opportunity to be heard. The district court concurred and granted summary judgment in favor of the tobacco companies. A divided panel of this Court rejected appellees' arguments and reversed the district court's judgment. After en banc review, however, Judge Selya and I agree with the district court and, therefore, affirm its grant of summary judgment and award of injunctive and declaratory relief in favor of plaintiffs-appellees.

Appellees are various manufacturers of cigarettes and smokeless tobacco products.1 They all currently sell their products in Massachusetts and are potentially subject to the requirements of Mass. Gen. Laws ch. 94, § 307B ("Disclosure Act").

Defendants-appellants are the Attorney General of Massachusetts and the Massachusetts Commissioner of Public Health.

All of the tobacco products manufactured by appellees include a variety of additives (in addition to tobacco, water, and reconstituted tobacco sheet). For example, common ingredients include sugars, glycerin, propylene glycol, cocoa, and licorice. These various additives are used as solvents, processing aids, pH modifiers, formulation aids for reconstituted tobacco, preservatives, humectants, tobacco protection aids, "plasticizing" agents, and, perhaps most importantly, flavorings. It is undisputed that appellees have spent millions of dollars developing formulas for their different brands, and when successful, those brands are worth billions of dollars. A major factor of each brand's success is its distinctive flavor, taste, and aroma.

While appellants argue that the added ingredients are neither pre-approved by regulators nor tested for safety, it is undisputed that most of the added ingredients are approved for consumption in food or "Generally Recognized As Safe" by the Food and Drug Administration. The one additive not found on either list is denatured alcohol, and this has been approved by the Bureau of Alcohol, Tobacco, and Firearms for use in the manufacture of tobacco products.

Each of the appellees closely guards its valuable ingredient lists. For example, within each company, only a few individuals are privy to the entire formula for any one brand. Suppliers are subject to confidentiality agreements and ship their products in packages which disguise their contents.

It is true that some ingredients of particular brands are known, and all ingredients used in any tobacco product are publicly available. However, this does not mean that complete brand-specific ingredient information can be obtained. In fact, various appellees have tried to "reverse engineer" the formulas of their competitors, but these attempts have been unsuccessful. Apparently, they have been able to determine the chemical composition of the various brands, but this information does not translate into a formula to recreate the product. Appellees assert, however, that if they were able to combine the chemical composition derived from this "reverse engineering" with a list of specific ingredients, arranged by relative amount, it would be much easier to discover a competitor's formula. Therefore, the tobacco companies argue that publication of their ingredient lists, organized by relative amount, on a brand-by-brand basis would likely destroy the secrecy of their formulas. This contention is not disputed by appellants.

B. Current Federal and State Disclosure Requirements

Tobacco companies currently have to disclose their ingredient lists to both the federal government and at least two state governments.

The federal government requires only that an aggregate list of all ingredients used in cigarettes and smokeless tobacco products be provided to the Department of Health and Human Services. 15 U.S.C. § 1335a(a). These lists, each of which contains hundreds of ingredients, neither identify the ingredients in any particular brand nor reveal which ingredients are used by which manufacturer. Id. The Department of Health and Human Services can study and report to Congress on the health effects of tobacco additives, including information on specific ingredients which may pose a health risk to consumers. Id. at § 1335a(b)(1)(A)-(B). However, without further legislation and disclosure, the federal government has no ability to warn consumers of the use of harmful additives in specific brands.

Two states, besides Massachusetts, require some disclosure of additives to tobacco products. Minnesota mandates that tobacco companies report only the use of several targeted additives in their products. Minn.Stat. § 461.17 (Supp.1997). Texas requires that the tobacco companies report brand-specific ingredient information, in descending quantities. Tex. Health & Safety Code Ann. §§ 161.351-55 (West Supp.2001). While this scheme superficially looks like the challenged Massachusetts legislation, Texas protects the ingredient lists by prohibiting public disclosure when those lists would be considered trade secrets under either federal or state law. Id. at § 161.254(c). The tobacco companies have complied and continue to comply with these disclosure requirements and have never challenged their validity.

In 1996, Massachusetts enacted the Disclosure Act, ostensibly to promote public health. Citing the fact that various tobacco product additives may have adverse health effects when burned, either alone or in combination with other additives, Massachusetts expressed an interest in being able to study more accurately the health effects of tobacco products on consumers. Massachusetts was also concerned that certain additives may increase nicotine delivery and that those additives might be used in cigarettes advertised as having a lower nicotine content.

In Massachusetts' view, previous disclosure requirements did not allow it to investigate adequately these public health concerns. For example, the publicly available ingredient lists do not identify additives according to brand or manufacturer. Therefore, Massachusetts could not study the interaction of additives and know whether those additives are actually combined. Nor could Massachusetts study the additives used in more popular brands and those brands targeted to younger consumers. No one disputes that these suggested studies are laudable and within the health and safety realm of the state's traditional police powers.

Massachusetts, however, has an additional goal to be realized through the Disclosure Act: it hopes to publicize the ingredient lists of various brands. This information, Massachusetts believes, will help consumers make more informed choices about the tobacco products they choose to consume. The envisioned effect is greater public awareness about the potential health effects of tobacco additives.

With these considerations in mind, Massachusetts enacted the Disclosure Act, which reads, in relevant part:

For the purpose of protecting the public health, any manufacturer of cigarettes, snuff or chewing tobacco sold in the commonwealth shall provide the department of public health with an annual report, in a form and at a time specified by that department, which lists for each brand of such product sold the following information:

(a) The identity of any added constituent other than tobacco, water or reconstituted tobacco sheet made wholly from tobacco, to be listed in descending order according to weight, measure, or numerical count; and

... [Any] information in the annual reports with respect to which the department determines that there is a reasonable scientific basis for concluding that the availability of such information could reduce risks to public health, shall be public records; provided, however, that before any public disclosure of such information the department shall request the advice of the attorney general whether such disclosure would constitute an unconstitutional taking of property, and shall not disclose such information unless and until the attorney general advises that such disclosure would not constitute an unconstitutional taking.2

Mass. Gen. Laws § 307B (emphasis added). Therefore, the Disclosure Act establishes two threshold requirements before an ingredient list "shall" be made public: (1) there must be a finding that publication "could reduce risks to public health;" and (2) the Massachusetts Attorney General must find that disclosure would not be an unconstitutional taking. Id.

Some further requirements have been established by regulations enacted under the Disclosure Act. Mass. Regs.Code tit. 105, § 660.200. These regulations require the Massachusetts Department of Public Health ("DPH") to provide sixty days' notice to the manufacturer before the proposed disclosure. Id. at § 660.200(E). To prevent disclosure, the manufacturer may remove its product from Massachusetts or reformulate it. Id. at § 660.200(F). An amendment to the regulations also allows the manufacturer to suspend disclosure by filing a lawsuit. Id. at § 660.200(G). Finally, until all requirements of the Disclosure Act and its enabling regulations have been met, the regulations provide that the DPH will keep the tobacco companies' ingredient lists confidential. Id. at § 660.200(G)(2).

The various tobacco companies filed this action in 1996, shortly after the Disclosure Act was enacted.3 Their complaint alleges that the Disclosure Act violates various provisions of the United States Constitution: the Commerce, Takings, and Due Process Clauses. U.S. Const. art. I, § 8, cl. 3, amend. V; amend. XIV.

A threshold issue arose as to whether the Disclosure Act is preempted by either the Federal Cigarette Labeling and Advertising Act, 15 U.S.C. § 1331-41, or the Comprehensive Smokeless Tobacco Health Education Act of 1986, 15 U.S.C. § 4401-08. The district court held that there was no preemption, and we affirmed. Philip Morris, Inc. v. Harshbarger, 122 F.3d 58 (1st Cir.1997) [hereinafter Philip Morris I].

Thereafter, the tobacco companies moved for a preliminary injunction based on their constitutional claims. On December 10, 1997, the district court entered an order that preliminarily enjoined appellants from enforcing the ingredient-reporting provisions of the Disclosure Act until further order of the court. On an interlocutory appeal, we again affirmed. Philip Morris, Inc. v. Harshbarger, 159 F.3d 670 (1st Cir.1998) [hereinafter Philip Morris II]. We found that the district court had neither "committed a clear error of law [nor] an abuse of discretion" in finding that the tobacco companies showed a reasonable likelihood of success on their claim that the Disclosure Act violates the Takings Clause. Id. at 680.

Following this affirmance, all parties filed motions for summary judgment. On September 7, 2000, the district court entered a Memorandum and Order granting the tobacco companies' motions and denying appellants' motion for summary judgment. Phillip Morris, Inc. v. Reilly, 113 F. Supp. 2d 129 (D.Mass.2000) [hereinafter Phillip Morris III]. The court found that the Disclosure Act violates the Takings, Due Process, and Commerce Clauses and issued a permanent injunction forbidding Massachusetts from requiring the disclosure of brand-specific ingredient information from the tobacco companies. Id. at 151.

Appellants filed timely appeals in which they challenged the district court's findings on the constitutional claims. Additionally, they argued that the tobacco companies' claims are not ripe because the Disclosure Act does not mandate publication of the submitted ingredient lists. In an opinion which has subsequently been withdrawn, a divided panel of this Court reversed. It found that the tobacco companies' claims are ripe, but agreed with appellants that the Disclosure Act does not violate the Takings, Due Process, or Commerce Clauses. The dissent agreed that the claims are ripe and the Disclosure Act does not contravene the Commerce Clause. However, it found violations of both the Takings and Due Process Clauses.

After a timely petition, we granted en banc review as to whether the Disclosure Act violates either the Takings or Due Process Clauses. Our review does not include revisiting the issues of whether the tobacco companies' claims are ripe or whether the Disclosure Act violates the Commerce Clause.

Because this case reaches us on appeal from a grant of summary judgment, I review the district court's judgment de novo. Euromotion, Inc. v. BMW of N. Am., Inc., 136 F.3d 866, 869 (1st Cir.1998).

The tobacco companies allege, and the district court found, that the Disclosure Act unconstitutionally takes the tobacco companies' property when it requires the tobacco companies to disclose their ingredient lists to Massachusetts, which may, in turn, publish those lists. To support this claim, the tobacco companies first argue that their ingredient lists are trade secrets and, as such, are property protected by the Takings Clause. Second, they argue that the public disclosure of these trade secrets destroys their value, thereby effecting a taking. Appellants counter with two separate arguments. First, they claim that the tobacco companies' interest in keeping their ingredient lists secret does not defeat the state's ability to require public disclosure where, as here, the requirement is "rationally related to a legitimate governmental interest." Ruckelshaus v. Monsanto Co., 467 U.S. 986, 1007, 104 S. Ct. 2862, 81 L. Ed. 2d 815 (1984). The asserted legitimate governmental interest is the health and safety of its citizens. Second, appellants dispute that Massachusetts law creates a property interest in trade secrets that are required by law to be disclosed to public agencies. I begin with analysis of the question of whether Massachusetts law protects the tobacco companies' ingredient lists as trade secrets.

In most states, trade secrets are property protected by the Takings Clause, see Monsanto, 467 U.S. at 1003-04, 104 S. Ct. 2862 (holding that Missouri law, which follows the Restatement of Torts, creates cognizable property right in trade secrets), and neither side disputes that Massachusetts has long recognized and protected trade secrets. See Jet Spray Cooler, Inc. v. Crampton, 377 Mass. 159, 385 N.E.2d 1349, 1354 (1972) (noting that Massachusetts has protected trade secrets based on public policy principles since at least 1868). Also, neither side suggests that Massachusetts treats trade secrets differently from other states or argues that the district court's application of the Restatement (First) of Torts was incorrect. See Phillip Morris III, 113 F.Supp.2d at 135-36. Finally, appellants do not contest that the tobacco companies' ingredient lists are trade secrets.4

Rather, appellants make a more subtle, but nonetheless ultimately ineffective, argument. Despite recognizing that Massachusetts' laws provide a remedy for misappropriation of trade secrets by private actors, see Junker v. Plummer, 320 Mass. 76, 67 N.E.2d 667, 669-70 (1946), appellants argue that Massachusetts has long established that it can require public disclosure of trade secrets to advance public health and safety.

In support of this argument, appellants first point to the Restatement (First) of Torts which says that the law may require the disclosure of a trade secret to "promote some public interest." § 757, cmt. d (1939). Certainly, courts have long recognized that trade secrets generally can be subject to disclosure under certain limited circumstances. See, e.g., Corn Prods. Ref. Co. v. Eddy, 249 U.S. 427, 431-32, 39 S. Ct. 325, 63 L. Ed. 689 (1919) (upholding required disclosure of ingredient lists to prevent consumer fraud) [hereinafter Corn Prods. II]. However, the fact that the public interest can sometimes override private property interests does not establish that the tobacco companies have no cognizable property interest when a state decides that publication of their trade secrets will further public health. In fact, Massachusetts continues to protect the integrity of many trade secrets despite the potentially valuable impact on the public interest if those trade secrets were to be placed in the public sphere. See, e.g., Gen. Chem. Corp. v. Dep't of Env't Quality Eng'g, 19 Mass.App.Ct. 287, 474 N.E.2d 183, 185 (1985) (discussing Massachusetts statute which specifically guarantees confidentiality of trade secrets belonging to hazardous waste industries and submitted pursuant to regulations). Instead, the potential for mandated disclosure in the public interest forms part of the inquiry as to whether a particular disclosure requirement is constitutional. See Penn Cent. Transp. Co. v. City of New York, 438 U.S. 104, 124, 98 S. Ct. 2646, 57 L. Ed. 2d 631 (1978) (establishing reasonable investment-backed expectations as one prong of the regulatory takings inquiry) [hereinafter Penn Central]. Finally, the Supreme Court specifically found that jurisdictions which follow the Restatement create a cognizable property interest in trade secrets. Monsanto, 467 U.S. at 1003-04, 104 S. Ct. 2862.

Second, appellants argue that General Chemical Corp. establishes that the state may generally seize trade secrets in the public interest. That case established no such proposition. Rather, the court only assumed, arguendo, that the state legislature could deprive hazardous waste industries of certain trade secrets in the context of regulating those industries. Id. at 185. Therefore, the case provides no notice that trade secrets are subject to disclosure.

Third, appellants point to the Massachusetts public records law which establishes that when a law requires trade secret information to be filed with a state agency, nothing requires those trade secrets to be treated as confidential. In fact, the public records law makes such information publicly available. See Mass. Gen. Laws ch. 4, § 7, cl. 26 (providing trade secret protection only when the information is submitted voluntarily, to further public policy development, and with a guarantee of confidentiality). Therefore, appellants argue that the tobacco companies have no property interest in their ingredient lists once a law requires them to submit that information to the state. Whether Massachusetts guarantees the confidentiality of trade secrets once they have been submitted to a state agency has no bearing on whether Massachusetts creates a property interest in trade secrets that is protected by the Takings Clause. Holders of trade secrets can always voluntarily submit their information to a state, consequently losing their property right. See, e.g., Monsanto, 467 U.S. at 1006-07, 104 S. Ct. 2862 (noting that Monsanto had voluntarily submitted its trade secrets information, knowing it was subject to public disclosure, as part of a regulatory scheme). The question is not whether trade secrets can be lost but whether trade secrets are a protected property interest in Massachusetts.

And the answer to that question is clear. Massachusetts protects trade secrets, Gen. Chem. Corp., 474 N.E.2d at 185 ("The words `trade secret' are commonly thought to carry a connotation of a property interest."), and appellants fail to identify any background principles of state law that successfully obviate appellees' property interest in their trade secrets. The fact that trade secrets are potentially subject to disclosure does not destroy the tobacco companies' interest because trade secrets still enjoy general protection. See Palazzolo v. Rhode Island, 533 U.S. 606, 630, 121 S. Ct. 2448, 150 L. Ed. 2d 592 (2001) ("A regulation or common-law rule cannot be a background principle for some owners but not for others. The determination whether an existing, general law can limit all economic use of property must turn on objective factors ..."). Specific laws simply cannot destroy property interests. In fact, this is precisely what the Takings Clause is designed to prevent: "a State, by ipse dixit, may not transform private property into public property without compensation.... This is the very kind of thing that the Taking Clause of the Fifth Amendment was meant to prevent. That clause stands as a shield against the arbitrary use of governmental power." Webb's Fabulous Pharmacies, Inc. v. Beckwith, 449 U.S. 155, 164, 101 S. Ct. 446, 66 L. Ed. 2d 358 (1980). Massachusetts cannot provide trade secret protection to some parties and then refuse others the same protections. Therefore, it is clear that the tobacco companies have a property interest in their trade secrets that is implicated by the Disclosure Act. In light of this, I turn to the question of whether the Disclosure Act violates the Takings Clause.

The Supreme Court has distinguished between two branches of Takings Clause cases: physical takings and regulatory takings. See Tahoe-Sierra Pres. Council, Inc. v. Tahoe Reg'l Planning Agency, 535 U.S. 302, 122 S. Ct. 1465, 1479, 152 L. Ed. 2d 517 (2002) (distinguishing "between acquisitions of property for public uses ... and regulations prohibiting private uses") [hereinafter Tahoe-Sierra]; see also Yee v. City of Escondido, 503 U.S. 519, 522, 112 S. Ct. 1522, 118 L. Ed. 2d 153 (1992) (delineating between claims of physical occupation and mere regulation). A physical taking occurs either when there is a condemnation or a physical appropriation of property. Tahoe-Sierra, 122 S.Ct. at 1478. Generally, courts apply "straightforward" per se rules when addressing physical takings. Id. A regulatory taking transpires when some significant restriction is placed upon an owner's use of his property for which "justice and fairness" require that compensation be given. Gold-blatt v. Hempstead, 369 U.S. 590, 594, 82 S. Ct. 987, 8 L. Ed. 2d 130 (1962); accord Penn. Coal Co. v. Mahon, 260 U.S. 393, 415, 43 S. Ct. 158, 67 L. Ed. 322 (1922) ("The general rule at least is that while property may be regulated to a certain extent, if that regulation goes too far it will be recognized as a taking."). For the most part, courts apply a three-part "ad hoc, factual inquiry" to evaluate whether a regulatory taking has occurred: (1) what is the economic impact of the regulation; (2) whether the government action interferes with reasonable investment-backed expectations; and (3) what is the character of the government action. Penn Central, 438 U.S. at 124, 98 S. Ct. 2646. However, the Supreme Court has developed at least one per se rule in the regulatory takings sphere. See Tahoe-Sierra, 122 S.Ct. at 1480. When a regulation denies all economically beneficial or productive uses of land, it is a taking. Lucas v. S.C. Coastal Council, 505 U.S. 1003, 1015, 112 S. Ct. 2886, 120 L. Ed. 2d 798 (1992).

Here, there is an alleged taking of intellectual property — trade secrets. The Supreme Court has addressed an alleged taking of trade secrets only once, in Monsanto. There, the Court simply applied the multi-factored regulatory takings analysis enunciated in Penn Central. Monsanto, 467 U.S. at 1004-06, 104 S. Ct. 2862.5 It failed to address any physical takings cases, id.,6 and therefore failed to resolve whether trade secrets can be the subjects of physical takings. Since the Supreme Court has previously limited its analysis to the regulatory takings sphere, I choose to begin there.7

Before proceeding to the Penn Central analysis, I note that the tobacco companies argue that the Lucas per se rule governs this case.8 The decision in and reasoning behind Lucas certainly raise some interesting questions about the constitutionality of the Disclosure Act. However, I am uncomfortable with the suggestion that we simply import that per se rule into this case. Lucas dealt with real, not personal, property, and the Court cautioned that the value of personal property could be wiped out without triggering the strictures of the Takings Clause. 505 U.S. at 1027-28, 112 S. Ct. 2886. This is not to say that Lucas is not relevant to the disposition of this case. Cf. Nixon v. United States, 978 F.2d 1269, 1284-85 (D.C.Cir.1992) (arguing that there is no compelling distinction between real and personal property as to make the application of per se rules inappropriate in regard to personal property). Rather, I simply choose to address these arguments while also considering the Penn Central factors.

Furthermore, I note that applying the Penn Central regulatory takings framework is not practically different from utilizing per se rules. Functionally, these per se rules are simply shortcuts. See Tahoe-Sierra, 122 S.Ct. at 1478 n. 17 (explaining that the same premise underlies both regulatory and physical takings cases but that the analysis is simply more complex for regulatory takings). An example of this principle is Loretto, a case which announced a per se rule in a physical takings context. See Tahoe-Sierra, 122 S.Ct. at 1478-79 (identifying the situation in Loretto as one which categorically requires compensation). There, the Court recited the Penn Central factors but then held that "when a physical intrusion reaches the extreme form of a permanent physical occupation, a taking has occurred." Loretto, 458 U.S. at 426, 102 S. Ct. 3164. The character of the government action was the dispositive factor, and the Court bypassed the remaining Penn Central factors. Id. at 435, 102 S. Ct. 3164. Similarly, in some regulatory takings cases, one factor is frequently dispositive. See Hodel, 481 U.S. at 717, 107 S. Ct. 2076 (focusing on the character of the government action); see also Monsanto, 467 U.S. at 1005, 104 S. Ct. 2862 (finding the interference with reasonable investment-backed expectations to be dispositive). There is, of course, one stark difference: once a per se rule has been announced, future courts do not have the luxury to consider the public interest, reasonable investment-backed expectations, or economic impact. Were I to import a per se rule into this case, either in a physical or regulatory takings context, I would ignore those Penn Central factors. However, whether I apply a regulatory takings analysis or a per se rule should not impact the ultimate decision. If the Disclosure Act's provisions are so "extraordinary," Hodel, 481 U.S. at 716, 107 S. Ct. 2076, as to make it properly subject to a per se rule, the considerations that led to adoption of that rule will also counsel me to find a taking under the Penn Central framework. Therefore, the concerns of Lucas will continue to inform my analysis.

As a final point before considering the Penn Central factors as they apply to this case, I would like to address the heavy charge leveled by the concurrence: that application of the Penn Central framework to this case ignores principles of stare decisis. See infra p. 47. I emphatically disagree with this characterization and am of the view that such a conclusion is only possible by the use of a self-serving definition of the term stare decisis.

As noted by the concurrence, the jurisprudence in this area is convoluted and subject to various interpretations. The fact that the concurrence and I understand Monsanto differently is not surprising. What is surprising is that the concurrence takes that understandable difference in opinion and translates it into an accusation that I am ignoring stare decisis. The heart of our disagreement lies with our conflicting interpretations of Monsanto, particularly as to the Court's discussion of the second scheme. See infra pp. 36-38, 40-44, 45-46. The concurrence finds that this discussion disposes of our current case and I simply do not agree, for reasons explained elsewhere. I do not think that Monsanto established a per se rule that once a trade secret holder establishes a reasonable investment-backed expectation the government may not require disclosure without triggering the protections of the Takings Clause. Since I interpret Monsanto to require courts to apply the Penn Central framework in cases like ours, I now proceed with that analysis.

i. Reasonable Investment-Backed Expectations

Despite the importance of reasonable investment-backed expectations under the Penn Central framework, the courts have struggled to adequately define this term. See generally R.S. Radford & J. David Breemer, Great Expectations: Will Palazzolo v. Rhode Island Clarify the Murky Doctrine of Investment-Backed Expectations in Regulatory Takings Law?, 99 N.Y.U. Envtl. L.J. 449, 449-50 (2001). Some very general contours are clear. Courts protect only reasonable expectations. Ideally, the relevant inquiry should recognize that not every investment deserves protection and that some investors inevitably will be disappointed. See Frank I. Michelman, Property, Utility, and Fairness: Comments on the Ethical Foundations of "Just Compensation" Law, 80 Harv. L.Rev. 1165, 1213 (1967). However, beyond the general landscape, there is a paucity of clear landmarks that can be used to navigate the terrain. Some recent decisions have added specific details, see, e.g., Palazzolo, 533 U.S. at 627, 121 S. Ct. 2448 (holding that whether property is acquired before or after a regulation is enacted does not completely determine the owner's reasonable investment-backed expectations), but many areas are still uncharted. As I proceed into this quagmire, the first guidepost is Monsanto.

Monsanto answered a challenge to disclosures by the EPA of data which had been submitted under the Federal Insecticide, Fungicide, and Rodenticide Act ("FIFRA"). 7 U.S.C. §§ 136-136y. FIFRA was enacted in 1947, amended in 1972, and amended again in 1978, each time offering different protections to submitted data. Monsanto challenged disclosures of data which had been submitted under each of these schemes, and the Court looked at the protections provided by each scheme to determine whether applicants had a reasonable investment-backed expectation that their trade secrets would remain secret. Following the Supreme Court's lead, I recount the provisions of each scheme and the Court's corresponding concerns.

In the period between 1947 and 1972, "FIFRA was primarily a licensing and labeling statute," Monsanto, 467 U.S. at 991, 104 S. Ct. 2862, and it failed to specify the government's ability to use and disclose data submitted by pesticide manufacturers. Id. at 1008, 104 S. Ct. 2862. Therefore, manufacturers like Monsanto had no guarantee that their data would be treated confidentially, nor did the government have specific authority to disclose such data. Id. The Court concluded that without a guarantee of confidentiality, Monsanto had no reasonable investment-backed expectation that its submitted data would remain secret. Id. Therefore, any disclosures of this data by the government did not constitute an unconstitutional taking of property.

The 1972 amendments transformed FIFRA from a labeling statute to a "comprehensive regulatory statute." Id. at 991, 104 S. Ct. 2862. Additional requirements were imposed on pesticides submitted for registration, and the EPA, as the administrative agency in charge of such regulations, gained additional enforcement powers. Id. at 991-92, 104 S. Ct. 2862. Congress also amended FIFRA to provide for certain public disclosures of data, but it explicitly prohibited the EPA from disclosing information which was deemed to be a trade secret. Id. at 992, 104 S. Ct. 2862. Another addition was a "mandatory data-licensing scheme." Id. This allowed the EPA to use data submitted by one registrant when considering subsequent applications as long as those later applicants agreed to compensate the original registrant. Id.

It was this second scheme that raised possible constitutional problems. Id. at 1010-14, 104 S. Ct. 2862. The difference arose because there was an "explicit governmental guarantee [which] formed the basis of a reasonable investment-backed expectation" that submitted data, designated as trade secrets, would be kept confidential. Id. at 1011, 104 S. Ct. 2862. A trade secret's value lies in the "right to exclude others." Id. If others are given the trade secret, the "holder of the trade secret has lost his property interest." Id. Therefore, if the government discloses the data that Monsanto submitted during this second period, a taking potentially occurs because the disclosure destroys the value of Monsanto's trade secrets. Id. at 1013-14, 104 S. Ct. 2862. Whether such a taking is unconstitutional hinges on whether Monsanto received adequate compensation, a question not before the Court. Id. The final amendments relevant to Monsanto occurred in 1978. They provided that any data submitted could be cited and considered by subsequent applications for fifteen years, so long as the original submitter is compensated. Id. at 994, 104 S. Ct. 2862. Finally, any qualified person could request that all health, safety, and environmental data be disclosed, regardless of whether such information had been designated a trade secret.9 Here, it was dispositive that Monsanto knew that the government might disclose any confidential data:

If, despite the data-consideration and data-disclosure provisions in the statute, Monsanto chose to submit the requisite data in order to receive a registration, it can hardly argue that its reasonable investment-backed expectations are disturbed when EPA acts to use or disclose the data in a manner that was authorized by law at the time of the submission.

Id. at 1006-07, 104 S. Ct. 2862. This notice negated any reasonable investment-backed expectations and, consequently, Monsanto's argument that a taking had occurred.

Despite appellants' arguments to the contrary, neither the first nor the third regime presented in Monsanto is directly analogous to the Disclosure Act. One stark difference sets them both apart and undermines their usefulness in this case. Monsanto complained about current and future disclosures of already submitted data. Monsanto, 467 U.S. at 1004, 104 S. Ct. 2862 ("Having determined that Monsanto has a property interest in the data it has submitted to EPA, we confront the difficult question whether a `taking' will occur when EPA discloses those data...." (emphasis added)). In answering that challenge, the Court measured Monsanto's reasonable investment-backed expectations at the time it submitted the data. Because there was no promise by the government under these two schemes to keep the data confidential, Monsanto had no basis on which to expect that its data would remain secret. In essence, Monsanto had "constructive notice" that its trade secrets might later be made public. Here, the tobacco companies challenge the ability of Massachusetts to compel future submissions of data which would be subject to disclosure. The tobacco companies have not voluntarily provided their ingredient lists. Therefore, their situation is fundamentally different from two of the scenarios that confronted Monsanto.10

The second scheme addressed by the Monsanto Court does shed some light on the current case, but it is not entirely dispositive. There, FIFRA provided Monsanto with an explicit guarantee of confidentiality. This guarantee established a reasonable investment-backed expectation that Monsanto's trade secrets would remain protected. When the Court decided that the government could not disclose submitted data which had been guaranteed confidentiality, the Court simply enforced the terms of the statute. Here, Massachusetts only generally protects trade secrets, establishing a right to recovery when a third party discloses or uses a trade secret without permission, Jet Spray Cooler, 385 N.E.2d at 1354, and the Disclosure Act only provides for publication of submitted data. It explicitly disclaims any long-term confidentiality.11 This distinction is important because a trade secret is lost if its holder gives the trade secret to another without extracting a guarantee of confidentiality. Monsanto, 467 U.S. at 1002, 104 S. Ct. 2862. Monsanto preserved its trade secrets because there was a promise of confidentiality. The tobacco companies will lose their trade secrets because there is no similar promise here. Therefore, a slightly different question is posed by the current case. In Monsanto, the question was whether the government could disclose trade secrets it had previously agreed to keep secret. Here, the question is whether Massachusetts can force the tobacco companies to cede their trade secrets.

To answer that question I must look at the tobacco companies' reasonable investment-backed expectations that they can maintain the integrity of their trade secrets. The fact that the Disclosure Act has been enacted is not dispositive because, as discussed above, Massachusetts cannot simply redefine property rights without regard to previously existing protections. See Webb's Fabulous Pharmacies, Inc., 449 U.S. at 164, 101 S. Ct. 446; cf. Palazzolo, 533 U.S. at 627, 121 S. Ct. 2448 (holding that enactment of a regulation inhibiting development before a purchaser acquires his property does not alone negate the purchaser's reasonable investment-backed expectations because otherwise "[a] State would be allowed, in effect, to put an expiration date on the Takings Clause"). I must examine the tobacco companies' reasonable investment-backed expectations "in light of the whole of our legal tradition," Lucas, 505 U.S. at 1035, 112 S. Ct. 2886 (Kennedy, J., concurring), not just in light of the provisions of the Disclosure Act.

To understand that legal tradition, I begin with a Supreme Court case from the early twentieth century which arguably provides constructive notice that ingredient lists are not inviolable. In Corn Products II, the Court considered whether it was a taking to require a manufacturer to disclose its ingredient list. In a tersely worded decision, the Court simply said:

And it is too plain for argument that a manufacturer or vendor has no constitutional right to sell goods without giving to the purchaser fair information of what it is that is being sold. The right of a manufacturer to maintain secrecy as to his compounds and processes must be held subject to the right of the state, in the exercise of its police power and in promotion of fair dealing, to require that the nature of the product be fairly set forth.

249 U.S. at 431-32, 39 S. Ct. 325 (emphasis added). While this language can be read to suggest that ingredient lists are subject to full disclosure, it refers only to "fair information." Such "fair information" could be something short of complete disclosure of all additives. For example, if Massachusetts found that the addition of one or more ingredients to tobacco products presented a health risk, disclosing when those specific ingredients are used might constitute "fair information." Cf. Minn.Stat. § 461.17 (requiring manufacturers to report the use of ammonia, arsenic, cadmium, formaldehyde, and lead in tobacco products and making such information public record).

This second interpretation gains credence from a closer reading of Corn Products II. The Court was addressing not a public health statute but a statute to prevent consumer deception. Corn Prods. II, 249 U.S. at 431, 39 S. Ct. 325 ("Evidently the purpose of the [labeling] requirement is to secure freedom from adulteration and misbranding...."). To prevent deception, it might make sense to require a complete list of ingredients. Only requiring a partial list could, in fact, increase consumer deception. In contrast, there is no evidence in the record that publication of only those ingredients which create health risks undermines the goal of promoting public health. In fact, such partial lists seem closest to the "fair information" referred to in the Corn Products II decision.12

More recent regulation, of both tobacco and other products, supports the idea that "fair information" is not always a complete ingredient list. While the federal government and other states worry about the health effects of tobacco additives, none of their regimes requires the publication of brand-specific ingredient lists. They either do not require brand-specific disclosures, or they grant the tobacco companies protections against public disclosure of ingredient lists submitted to the states. Furthermore, regulations governing other products recognize the difference between requiring accurate labeling and protecting secret formulas. For example, the Food, Drug, and Cosmetic Act allows additives to be grouped as "spices, flavoring, and coloring" without specifically identifying the individual ingredients. 21 U.S.C. § 343(i)(2). This allows many manufacturers to maintain their secret formulas.

Given this complex background and the fact that Massachusetts has long protected trade secrets, see Jet Spray Cooler, 385 N.E.2d at 1354, I cannot hold that the tobacco companies have no reasonable investment-backed expectation that their ingredient lists will remain secret. Therefore, I proceed to the other elements of the Penn Central inquiry.

ii. Economic Impact

In contrast to reasonable investment-backed expectations, the law regarding economic impact is fairly straightforward. The inquiry is whether the regulation "impair[s] the value or use of [the] property" according to the owners' general use of their property. PruneYard Shopping Ctr., 447 U.S. at 83, 100 S. Ct. 2035. Not only is the use to which the property owner puts her property important, but the economic impact needs to be considered in the context of other laws and regulatory schemes. See Connolly v. Pension Benefit Guar. Corp., 475 U.S. 211, 225-26, 106 S. Ct. 1018, 89 L. Ed. 2d 166 (1986) (evaluating economic impact of imposing withdrawal fees on employers who leave pension funds within context of entire ERISA scheme).

The evidence presented here is similarly straightforward. The appellees' have spent millions of dollars developing the formulas for different brands. The evidence shows that public disclosure of the appellees' ingredient lists, even in part, will make it much easier to reverse engineer those formulas. If competitors can obtain these formulas, they can replicate appellees' products, undermining the value of appellees' brands. Some of those brands, such as Marlboro, are worth billions of dollars. While it is impossible to predict the exact economic impact that the Disclosure Act will have, it is potentially tremendous.

iii. Character of the Government Action

In this last section, I delve into how the Disclosure Act regulates and what that regulation does to the tobacco companies' trade secrets. See Hodel, 481 U.S. at 716, 107 S. Ct. 2076 (examining the effect of the escheat provisions of the Indian Land Consolidation Act of 1983). As mentioned above, the tobacco companies believe that the Disclosure Act regulations are so egregious that they rise to the level of a per se taking. They ground this claim on the fact that the Disclosure Act gives Massachusetts the right to publish the ingredient lists. The Act, in essence, prevents the tobacco companies from excluding others from their trade secrets, destroying their essential attribute. It also, allegedly, destroys the entire value of the trade secrets. I now address those arguments in full. However, I will also balance the effects of the Disclosure Act against Massachusetts' interests. See Keystone Bituminous Coal Ass'n v. DeBenedictis, 480 U.S. 470, 488, 107 S. Ct. 1232, 94 L. Ed. 2d 472 (1987) (considering the state's asserted interests in "health, the environment, and the fiscal integrity of the area") [hereinafter Keystone]. Here, the asserted state interest is the promotion of public health.

I begin with the tobacco companies' argument that they will lose the right to exclude others from their trade secrets and, consequently, their trade secrets will lose all value. It appears paradigmatic that these assertions are true. In Monsanto, the Supreme Court recognized that, "[i]f an individual discloses his trade secret to others who are under no obligation to protect the confidentiality of the information, or otherwise publicly discloses the secret, his property right is extinguished." 467 U.S. at 1002, 104 S. Ct. 2862. That is exactly what happens here. The Disclosure Act requires the tobacco companies to share their trade secrets with Massachusetts, which is under no obligation to keep the information secret.13 In fact, the Disclosure Act spells out the terms under which Massachusetts will publish those trade secrets. It, thus, provides specific notice to the tobacco companies that Massachusetts need not respect their property rights. Therefore, if the tobacco companies comply with the requirements of the Disclosure Act, their property right will be extinguished. In the future, should a competitor use published data, the tobacco companies will have no ability to enforce their rights. See Jet Spray Cooler, 385 N.E.2d at 1354 ("The essence of an action for the wrongful use of trade secrets is the breach of the duty not to disclose or to use without permission confidential information acquired from another."). Similarly, the value of the trade secrets will be lost because their value lies in the ability of the tobacco companies to exclude others. See Monsanto, 467 U.S. at 1012, 104 S. Ct. 2862. The Disclosure Act essentially destroys the tobacco companies' trade secrets.

This fact may very well prove to be dispositive in this case. In Armstrong v. United States, 364 U.S. 40, 80 S. Ct. 1563, 4 L. Ed. 2d 1554 (1960), the Supreme Court considered the implications of a government action which, as a secondary effect, destroyed a private party's lien. The Court held that this was a taking and "not a mere `consequential incidence' of a valid regulatory measure." Id. at 48, 80 S. Ct. 1563. The Court then continued: "Before the liens were destroyed, the lienholders admittedly had compensable property. Immediately afterwards, they had none. This was not because their property vanished into thin air. It was because the Government for its own advantage destroyed the value of the liens." Id. The Disclosure Act creates a similar situation. The tobacco companies have a protected property interest which the Disclosure Act will completely destroy.14

On the other hand, in Andrus v. Allard, 444 U.S. 51, 100 S. Ct. 318, 62 L. Ed. 2d 210 (1979), the Supreme Court made it clear that regulation can severely undermine the economic value of personal property and not rise to the level of a taking. Id. at 66, 100 S. Ct. 318. There, the federal government had banned sales of all items containing eagle parts. Id. at 56, 100 S. Ct. 318. The end result was that some people who had artifacts made of lawfully acquired eagle parts could not sell their products. Id. at 62-63, 100 S. Ct. 318. Consequently, the artifacts lost essentially all of their economic value. Id. at 66, 100 S. Ct. 318 (positing that some value could be extracted by displaying the artifacts for an admissions charge). While this was a "significant restriction," the Court held that this "destruction of one `strand' of the bundle" of property rights did not constitute a taking. Id. Rather, the substantial state interest in preserving eagles justified the regulation. Id. at 58, 100 S. Ct. 318 (discussing policy rationale behind the regulation) and 66-68 (upholding the regulation despite the burden it places on lawful property owners).

The end result reached in Andrus, however, must be compared with the result in Hodel.15 In Hodel, the Supreme Court addressed whether the Indian Land Consolidation Act of 1983 created an unconstitutional taking when it destroyed the rights of descent and devise which had previously attached to undivided fractional interests in land. 481 U.S. at 706-10, 107 S. Ct. 2076. Congress had enacted this legislation to attempt to revise an "administratively unworkable and economically wasteful" system of administering Indian lands. Id. at 707, 107 S. Ct. 2076. To further that goal, the statute destroyed the rights of descent and devise for small fractional interests of land and, instead, had those interests escheat to the tribe. Id. at 709, 107 S. Ct. 2076. This, in fact, was such an "extraordinary" government action as to make it a taking, despite the indeterminancy of the other Penn Central factors and the "serious public problem" which the regulation addressed. Id. at 714-18, 107 S. Ct. 2076.

The question then arises as to which line of cases governs here. The simple loss of economic value, alone, is probably not enough. See Lucas, 505 U.S. at 1027-28, 112 S. Ct. 2886 (noting that regulations can constitutionally render personal property "economically worthless"). "[G]overnment regulation — by definition — involves the adjustment of rights for the public good. Often this adjustment curtails some potential for the use or economic exploitation of private property. To require compensation in all such circumstances would effectively compel the government to regulate by purchase." Andrus, 444 U.S. at 65, 100 S. Ct. 318. There is a point, however, at which compensation is due, see Penn. Coal Co., 260 U.S. at 415, 43 S. Ct. 158, and this is not simply a case where the tobacco companies' property has been rendered worthless. Their property right has been "extinguished." Monsanto, 467 U.S. at 1002, 104 S. Ct. 2862. Consequently, it appears unconstitutional.

Appellants urge us, however, to consider the asserted state interest, promoting public health, as a counterbalance.16 I recognize that appellants have asserted a significant, perhaps compelling, state interest: a right for Massachusetts to protect and promote the health of its citizens. If I was convinced that this regulation was tailored to promote health and was the best strategy to do so, I might reconsider our analysis. Numerous cases show that a crucial part of the regulatory takings equation is the government interest. See, e.g., Keystone, 480 U.S. at 488, 107 S. Ct. 1232 ("[T]he nature of the State's interest is a critical factor in determining whether a taking has occurred."). However, the cases also show that the means should bear some reasonable relationship to the ends. See id. at 487 n. 16, 107 S. Ct. 1232 (noting that Pennsylvania Coal Co. rejected the legislature's proffered goal in enacting the regulation when it found an unconstitutional taking).

I simply am not convinced that the Disclosure Act, particularly the provisions about which the tobacco companies complain, really helps to promote public health. The Disclosure Act allows for full disclosure of the ingredient lists when doing so "could" further public health. This places an extremely low burden on Massachusetts. Frankly, for a state to be able to completely destroy valuable trade secrets, it should be required to show more than a possible beneficial effect. Cf. Keystone, 480 U.S. at 485-93, 107 S. Ct. 1232 (explaining that courts should balance the public and private interests when evaluating regulatory takings claims). The tremendous individual loss is simply not justified by such a speculative public gain. Furthermore, it is not at all clear that protecting the overall integrity of the tobacco companies' ingredient lists will interfere with Massachusetts' goal of promoting public health. See Hodel, 481 U.S. at 717-18, 107 S. Ct. 2076 ("The difference in this case is the fact that both descent and devise are completely abolished; indeed they are abolished even in circumstances when the governmental purpose sought to be advanced, consolidation of ownership of Indian lands, does not conflict with the further descent of the property."). I note that the tobacco companies comply, without complaint, with regimes which require them to make confidential disclosure of brand-specific, ingredient information, see Tex. Health & Safety Code Ann. §§ 161.351-55, or which require public disclosures when specific ingredients are used, see Minn.Stat. § 461.17. There is no evidence that suggests that regimes similar to those adopted by Texas and Minnesota, or some combination thereof, would not achieve the goals which appellants claim underlie the requirements of the Disclosure Act.17

Because this is a crucial point, I wish to further clarify what I have just concluded above. I am not requiring Massachusetts to adopt the narrowest regulation possible to address its laudable goals. Rather, the tobacco companies complain about a specific element of the Disclosure Act, namely, that Massachusetts can publish their entire ingredient lists if doing so "could" further public health. A prior holding, which is not currently before us, decided that under this standard, Massachusetts will publish the tobacco companies' ingredient lists.18 Phillip Morris III, 113 F. Supp. 2d 129. I simply find that this actual publication, or right to publish, under the minimal standard set forth, has not been shown to further the stated goal of promoting public health in such a way as to counterbalance the tremendous private loss involved. Therefore, it is clear that the character of the government action weighs heavily against sustaining the Disclosure Act.

iv. Conclusion — Regulatory takings analysis

As I conclude my analysis of the Penn Central factors, I first note that there is no formula as to how to weigh the importance of the various factors. As has been clear from the preceding discussion, different factors can be dispositive. See, e.g., Hodel, 481 U.S. at 716, 107 S. Ct. 2076 (resting on the character of the government regulation which the Court found to be "extraordinary"); Monsanto, 467 U.S. at 1005, 104 S. Ct. 2862 (finding the lack of reasonable investment-backed expectations to be dispositive).

Here, the tobacco companies have at least some reasonable investment-backed expectation that their trade secrets will remain secret and the economic impact of revelation is likely to be great. These factors, alone, may not be sufficient to raise this case to the level of an unconstitutional taking. However, the character of the government action determines the case. The Disclosure Act causes the tobacco companies to lose their trade secrets, entirely, and appellants advance no convincing public policy rationale to justify the taking itself. Instead, they point to a general, laudable goal which cannot justify the specific action of which the tobacco companies complain. Therefore, I find that the Disclosure Act violates the Takings Clause by taking appellees' property without just compensation.

This, unfortunately, does not completely end the inquiry. I must turn briefly to the doctrine of unconstitutional conditions.

The Disclosure Act is unlike some other challenged government actions because the tobacco companies do not need to cede their ingredient lists to Massachusetts. They can opt out entirely, simply by not selling their products in Massachusetts.19 Therefore, their claim is really that Massachusetts has placed an unconstitutional condition on their right to sell their products in Massachusetts. If the Disclosure Act simply required the tobacco companies to submit their ingredient lists for possible publication, it would be unconstitutional. The question then is whether Massachusetts can constitutionally condition the right to sell tobacco products in Massachusetts on submission to this scheme.

The doctrine of unconstitutional conditions is fairly well-developed. "[T]he government may not require a person to give up a constitutional right — here the right to receive just compensation when property is taken for public use — in exchange for discretionary benefit conferred by the government where the benefit sought has little or no relationship to the property." Dolan v. City of Tigard, 512 U.S. 374, 385, 114 S. Ct. 2309, 129 L. Ed. 2d 304 (1994). Beyond these general contours, different inquiries have developed which apply to different types of property. In the case of intellectual property, Monsanto provides the relevant standard.20

In the final element of Monsanto, the Court addressed whether the government could require pesticide manufacturers to submit trade secrets which could then be disclosed to other parties. Monsanto claimed that these data disclosure provisions created an unconstitutional condition. Monsanto, 467 U.S. at 1007, 104 S. Ct. 2862. In respect to this claim, the Court said "as long as Monsanto is aware of the conditions under which the data are submitted, and the conditions are rationally related to a legitimate Government interest, a voluntary submission of data by an applicant in exchange for the economic advantages of a registration can hardly be called a taking." Id. This holding depended on the fact that Monsanto submitted its data in exchange for a valuable government benefit: a registration. See Nollan v. Cal. Coastal Comm'n, 483 U.S. 825, 833 n. 2, 107 S. Ct. 3141, 97 L. Ed. 2d 677 (1987) ("[W]e found merely that the Takings Clause was not violated by giving effect to the Government's announcement that application for `the right to [the] valuable Government benefit,' of obtaining registration of an insecticide would confer upon the Government a license to use and disclose the trade secrets contained in the application." (citations omitted)).21 Therefore, as part of a regulatory scheme which confers some government benefit upon a manufacturer, Monsanto establishes that the government may require that manufacturer to relinquish its rights to a trade secret.

Appellants argue that this holding governs here. I disagree. They claim that Massachusetts has a "legitimate Government interest" in protecting the health and safety of its citizens. I agree that this is indeed a legitimate state interest. My disagreement lies, rather, with the other side of the equation. The state must offer a valuable government benefit. Id. The right offered here is the right to sell tobacco products in Massachusetts. In Nollan, the Supreme Court considered what constitutes such a benefit with regard to land. The Court held that the right to build upon one's land is not such a benefit that would allow a state to require a landowner to grant a public easement across his property. Id. at 833 n. 2, 107 S. Ct. 3141. In contrast, in Monsanto, the government granted a license, created a de jure data-licensing scheme, and established a period of exclusive use for new ingredients in exchange for the right to disclose some trade secrets. 467 U.S. at 994, 104 S. Ct. 2862. Applying these two precedents to our case, I conclude that allowing a manufacturer to simply sell its legal product is more similar to building on one's land than to the complex regulatory scheme in Monsanto. Therefore, Massachusetts cannot condition the right to sell tobacco on the forfeiture of any constitutional protections the appellees have to their trade secrets. As a result, the Disclosure Act is invalid because it creates an unconstitutional taking of the tobacco companies' products.

Because I find that the Disclosure Act is invalid under the Takings Clause, I will not address the question of whether it also violates the Due Process Clause.

For the reasons discussed above, I find that the Disclosure Act violates the Takings Clause. Therefore, I affirm the district court's judgment.22

Affirmed.