BORMAN, LLC, Plaintiff-Appellant, v. 18718 BORMAN, LLC; Joseph Schwebel, Defendants-Appellees, 777 F.3d 816

OPINION COOK, Circuit Judge. This appeal arises from Defendant-Appellee 18718 Borman, LLC’s (“Borrower”) default on a nonrecourse loan secured by property located on Detroit’s Borman Avenue, the street after which Borrower takes its name. The lender foreclosed, and Borman, LLC (“Purchaser”) — an unrelated entity also named after Borman Avenue— purchased the property at auction with a winning bid of $756,000. Then Purchaser, standing in the lender’s shoes, sued Borrower and its guarantor to collect a roughly $6 million deficiency. The district court granted

summary judgment in favor of Borrower, and we AFFIRM. I. The issues on appeal require familiarity with the structure of Commercial Mortgage-Backed Securities (“CMBS”) loans and their recent history in Michigan. We review this background information before recounting the particular facts of this case. A. CMBS Loans In a nutshell, CMBS loans provide commercial real estate developers with project financing through capital markets. See Wells Fargo Bank, NA v. Cherryland Mall Ltd. P’ship (Cherryland I), 295 Mich.App. 99, 812 N.W.2d 799, 802 (2011) (per curiam) (citing Am. Br. of Commercial Mortg. Sec. Assoc. & Mortg.

Bankers Assoc, at 4-14, In re Gen. Growth Props., Inc., 409 B.R. 43 (S.D.N.Y.2009)). Commercial real estate developers obtain CMBS loans to finance individual projects. The developer mortgages the property being developed as security for the CMBS loan. Lenders issue a number of CMBS loans and then pool the resulting mortgages into a trust. Finally, the trust issues securities backed by the cash flow from the developers’ loan payments.

CMBS financing attracted new sources of capital to the commercial real estate market and “played a major role in leading the country out of the nationwide real estate depression caused by the savings and loan crisis of the late 1980s.” Id. (punctuation omitted). To attract these non-traditional investors, such as pension funds and individual bond investors, lenders structured CMBS loans to limit the risk that one developer’s failed project might “drag the entire securitized mortgage pool into bankruptcy.” (R. 51, Summ. J.

Order at 3.) And bankruptcy proceedings could hamper foreclosure by the trust and jeopardize timely payment to bondholders. CMBS loans contain two “bedrock elements” to limit the risks posed by bankruptcy. Cherryland I, 812 N.W.2d at 802 (citation omitted). First, CMBS loans are nonrecourse. When a borrower defaults on a nonrecourse loan, the lender may foreclose on the asset and any other prearranged security, but the borrower and its guarantors do not become personally liable for any deficiency.

In return for nonrecourse liability, the borrower promises to refrain from certain “bad boy” acts: financial behavior likely to increase the risk of default and bankruptcy. These covenants against “bad boy” acts constitute the second bedrock principle of CMBS loans. In general, the covenants require the borrower to maintain single-purpose-entity status. To achieve this, the developer “ring fences” the mortgaged property’s operations from his other projects, usually by creating a separate legal entity concerned solely with managing the mortgaged property to act as the borrower.

This way, a developer’s personal financial difficulty resulting from his other, more risky projects will not affect the trust’s ability to foreclose on the mortgaged property and — more importantly — its ability to make scheduled payments to bondholders. Relevant to this ease, the CMBS loan covenants commonly include a so-called solvency covenant, a promise that the borrower “shall not ... fail to remain solvent or pay its own liabilities (including, without limitation, salaries of its own employees) only from its own funds.” Cherryland I, 812 N.W.2d at 803 (citation omitted).

Many in the CMBS loan industry intended solvency covenants to serve as a promise against filing for bankruptcy or colluding in an involuntary bankruptcy. (See R. 40-14, Mins. Mich. Senate'Econ. Dev. Comm, at 22-23.) For if “insolvent” meant simply the loss of cash flow resulting in default, then defaulting on a CMBS loan due to adverse market forces would expose a borrower — and its developer-principal — to personal liability, effectively obviating the first bedrock principle of nonrecourse liability. B.

The Cherryland Decisions and the NMLA In Cherryland I, the Michigan Court of Appeals considered how to interpret the solvency covenant quoted in the previous paragraph. In that case, the borrower failed to make its scheduled payments, and the trustee foreclosed on the collateral. Cherryland I, 812 N.W.2d at 804. The trustee then sued for a deficiency judgment against the borrower and its principal-guarantor, arguing that the default constituted insolvency, which destroyed the borrower’s single-purpose-entity status and permitted recourse liability. Id.

The borrower and its principal-guarantor countered that the promise of a nonrecourse loan would prove illusory if the court interpreted the solvency covenant to include simple default. Id. at 806, 814-15. The court of appeals sided with the trustee, holding that the CMBS loan documents unambiguously required the borrower to remain solvent in order to maintain its single-purpose-entity status. The court rejected the borrower’s

interpretation because it relied on extrinsic evidence of industry practice, and the borrower failed to demonstrate any textual ambiguity with respect to the terms “single-purpose-entity status” or “solvent.” Id. at 809-10, 814-15. The court also rejected the stance that the trustee’s

interpretation created an illusory promise, demurring on the ground that the loan’s allegedly illusory nature was a matter of Michigan legislative policy. Id. at 815-16. The court remarked in closing: “We recognize that our

interpretation seems incongruent with the perceived nature of a nonrecourse debt and are cognizant of the amici curiae’s arguments and calculations that, if accurate, indicate economic disaster for the business community in Michigan....” Id. at 815. While the borrower’s appeal of Cherry-land I pended before the Michigan Supreme Court, the Michigan Legislature accepted the court of appeals’s invitation to address the propriety of solvency covenants. See Wells Fargo Bank, NA v. Cherryland Mall Ltd. P’ship (Cherryland II), 300 Mich.App. 361, 835 N.W.2d 593, 597 (2012) (per curiam).

In late March 2012, the legislature passed the Nonrecourse Mortgage Loan Act (“NMLA”) after conducting a senate hearing during which witnesses uniformly criticized Cherryland I and augured statewide economic disaster based on the number of CMBS loans in default — and therefore subject to a deficiency- — after the 2008 recession. (See generally R. 40-14, Mins. Mich. Senate Econ. Dev.

Comm.) The NMLA applies retroactively to render solvency covenants in nonrecourse loans unenforceable, declaring them “inconsistent with ... the nature of a nonrecourse loan[ and] ... an unfair and deceptive business practice ... against public policy [that] should not be enforced.” 2012 Mich. Legis. Serv. P.A. 67, enacting § 1 (S.B.992); see also Mich. Comp. Laws (“MCL”) §§ 445.1591-95. In light of the NMLA, the Michigan Supreme Court declined to review the Cherryland I decision and remanded the case to the court of appeals for reconsideration. See Cherryland II, 835 N.W.2d at 596.

The court of appeals held that the NMLA rendered the solvency covenant unenforceable and barred the trustee’s deficiency action. Id. The court went on to reject the trustee’s challenge to the statute’s validity under the United States and Michigan Constitutions. Id. at 598-606. With this background, we turn to the facts of the case before us. C. Case-Specific Facts In June 2005, Borrower obtained an $8.7 million CMBS loan from Morgan Stanley Mortgage Capital, Inc., secured by the commercial property located at 18718 Borman Avenue in Detroit.

Defendant-Appellee Joseph Schwebel, Borrower’s principal, guaranteed all obligations on the loan for which Borrower might become personally liable. Borrower used the loan to purchase the property from and lease it back to a subsidiary of the Great Atlantic & Pacific Tea Company for use as a grocery distribution center. In December 2010, the grocery chain’s subsidiary filed for bankruptcy. The bankruptcy court eventually permitted the subsidiary to terminate the lease and abandon the property. Borrower tried and failed to locate a replacement tenant or sell the property for its pre-recession value.

In October 2010, Morgan Stanley’s loan servicer sent Borrower a formal notice of default, and Borrower turned the property over to a receiver. A year later, the servicer foreclosed on the property and purchased it with a $2.1 million credit bid. After the auction, either Morgan Stanley or its loan servicer took possession of approximately $1.76 million in escrow and a $500,000 letter of credit from Schwebel, both deposited as additional collateral under Borrower’s loan. But neither Morgan Stanley nor the servicer sought a deficiency judgment.

In 2012, the loan servicer marketed the property onAuction.com, advertising that Borrower held the property subject to a nonrecourse loan before foreclosure. Purchaser’s principals obtained the property, then appraised at $4.6 million, and an assignment of Morgan Stanley’s rights under the loan agreement with a high bid of $756,000. Purchaser’s principals stated in their depositions that they never read the underlying loan documents before executing the sale and that they purchased the property because of the low asking price.

Several months later- — and two months after the Michigan Court of Appeals decided Cherryland II — Purchaser filed the instant action seeking a deficiency judgment. Purchaser took the position, as it does now, that it stands in the shoes of the lender. Pointing to the solvency covenant in Borrower’s loan documents, Purchaser argued that Borrower lost single-purpose-entity status upon default and — along with Schwebel — became personally hable for a deficiency of $6 million plus interest. Each party moved for

summary judgment, and the district court granted it to Borrower and Schwebel. The .court found that the NMLA: (1) rendered the solvency covenant in Borrower’s CMBS loan unenforceable; (2) violated neither the Contract nor Due Process Clauses of the United States and Michigan Constitutions; and (3) comported with Michigan’s constitutional provision mandating the separation of governmental powers. Purchaser challenges each conclusion. II. We review the district court’s grant of

summary judgment de novo. Chapman v. United Auto Workers Local 1005, 670 F.3d 677, 680 (6th Cir.2012). We construe all reasonable inferences in favor of the non-moving party, Matsushita Elec. Indus. Co. v. Zenith Radio Corp., 475 U.S. 574, 587, 106 S.Ct. 1348, 89 L.Ed.2d 538 (1986), and uphold the grant of

summary judgment if “there is no genuine dispute as to any material fact and the movant is entitled to judgment as a matter of law,” Fed. R.Civ.P. 56(a). III. Purchaser first challenges the district court’s ruling that the NMLA bars its deficiency action, arguing that the terms of Borrower’s loan take it outside the NMLA’s ambit.

Specifically, Purchaser argues that Borrower’s CMBS loan neither fits within the NMLA’s definition of “non-recourse loan,” nor includes a prohibited “post closing solvency covenant.” The district court rightly began by explaining that Michigan law governs construction of the NMLA’s provisions. When interpreting Michigan statutes, we “[f]irst and foremost ... give effect to the Legislature’s intent.” Tryc v. Mich.

Veterans’ Facility, 451 Mich. 129, 545 N.W.2d 642, 646 (1996) (citation omitted). “If the language of a statute is clear and unambiguous, the plain meaning of the statute reflects the legislative intent and judicial construction is not permitted.” Id. (citation omitted); see also Driver v. Naini, 490 Mich. 239, 802 N.W.2d 311, 316 (2011) (“Statutory provisions must be read in the context of the entire act, giving every word its plain and ordinary meaning.”). “If reasonable minds can differ regarding [the statute’s] meaning, then judicial construction is appropriate.” Yaldo v. N. Pointe Ins.

Co., 457 Mich. 341, 578 N.W.2d 274, 276 (1998) (citation omitted); see also Girard v. Wagenmaker, 437 Mich. 231, 470 N.W.2d 372, 375 (1991) (“[T]his Court must still give the statute a valid and reasonable construction that will reconcile any inconsistencies and give effect to all its parts.”). A.

The Loan’s “Nonrecourse” Status The district court correctly found that Borrower’s loan qualifies as a “nonrecourse loan” under the NMLA, which defines that term as any “commercial loan secured by a mortgage on real property located in this state and evidenced by loan documents” containing one or more enumerated nonrecourse provisions. MCL § 445.1592(b)-(c).

Purchaser concedes that Borrower’s loan satisfied the NMLA’s requirements for a nonrecourse loan at the time of its execution but disputes the NMLA’s applicability on the ground that the loan lost its nonrecourse status before the act’s March 29, 2012 effective date. To make this argument, Purchaser relies on a

section of the NMLA separate from the definition of “nonrecourse loan.” See MCL § 445.1595 (stating that the NMLA “applies to the enforcement and

interpretation of all non-recourse loan documents in existence on, or entered into on or after, the effective .date of this act”). Borrower’s promissory note provides: Notwithstanding the foregoing, the agreement of Lender not to pursue recourse liability as set forth in subsection (

a) above SHALL BECOME NULL AND VOID and shall be of no further force and effect and the Debt shall be fully recourse to Borrower in the event that: ... Borrower ... fails to comply with any provision of

Section 4.2 of the Security Instrument [which includes a solvency covenant].... (R. 1-3, Note art. XI, § (c).) Purchaser contends that this language rendered the loan’s nonrecourse provision “null and void” upon Borrower’s default in 2010, transforming it into a recourse loan before the NMLA’s effective date. According to Purchaser, not only must the loan documents in existence on March 29, 2012, include one of the enumerated nonrecourse provisions, but also the nonrecourse provision must remain enforceable through that date. Compare MCL § 445.1592, with § 445.1595.

As the district court recognized, however, Purchaser’s reading of the NMLA presents neither the only possible construction nor the one that most closely hews to the legislature’s intent. The statute defines nonrecourse loans as including one or more enumerated provisions in the loan documents without reference to whether those provisions remain enforceable. See MCL § 445.1592(b)-(c).

Under the district court’s alternative construction, one first asks whether the loan documents contain the necessary provision and, second, whether those documents were “in existence on, or entered into on or after, [March 29, 2012].” If one answers both questions in the affirmative, the NMLA applies regardless of whether the nonrecourse provision remained enforceable beyond March 29, 2012. Here, as specified by MCL § 445.1595, Borrower’s loan documents continued to exist past March 29, 2012, and contain one of the enumerated triggering provisions. (R. 1-3, Note art.

XI, § (a).) So the NMLA bars Purchaser’s deficiency action. The district court correctly determined that this construction best embodies the intent behind the NMLA. See Yaldo, 578 N.W.2d at 276. Purchaser’s suggestion would permit lenders to seek deficiency judgments on many, if not all, CMBS loans in default before March 29, 2012. As the district court recognized, such a result would contradict the legislature’s prohibition on the enforcement of solvency covenants as violative of public policy. See 2012 Mich. Legis. Serv.

P.A. 67, enacting § 1 (S.B.992) (“It is the intent of the legislature that this act applies to any claim made or action taken to enforce a post closing solvency covenant on or after the effective date of this act....”). Permitting deficiency judgments on CMBS loans that went into default during the 2008 recession, moreover, would do nothing to prevent the potential crisis in commercial real estate financing that the legislation looked to forestall. The district court identified another contradiction of the legislature’s intent in Purchaser’s construction.

Purchaser acknowledges that the legislature enacted the NMLA primarily to counteract the effect of Cherryland I. But if the NMLA applied only to CMBS loans not yet in default on March 29, 2012, it could not have applied to the loan at issue in the Cherryland decisions. Although Purchaser makes much of the fact that Borrower’s promissory note expressly renders the nonrecourse provision “null and void,” Purchaser fails to adequately distinguish this wording from the contractual language invalidating the nonrecourse provision in Cherryland.

Recognizing as much, the district court found that the Michigan Court of Appeals implicitly rejected Purchaser’s construction in Cherryland II, when it held that the NMLA barred the deficiency judgment it had previously upheld in Cherryland I. And where a state’s highest court has yet to construe a statute, intermediate appellate court decisions offer persuasive authority on how the highest court would rule. Kingsley Assocs., Inc. v. Moll PlastiCrafters, Inc., 65 F.3d 498, 507 (6th Cir.1995).

Alternatively, Purchaser argues that Borrower’s loan ceased to qualify as non-recourse under the NMLA after Morgan Stanley’s 2011 foreclosure because the loan was no longer “secured by a mortgage on real property” as of March 29, 2012. See MCL § 445.1592(b). The same logic applies here. Although the NMLA defines nonrecourse loans as “secured by a mortgage on real property,” the statute includes no caveat that this state of affairs must obtain beyond the effective date. MCL § 445.1595. Purchaser again attempts to borrow from a separate

section of the NMLA and read a temporal requirement into the definition of nonrecourse loan. And once again, we read the statutory sections independently because Purchaser’s construction narrows the NMLA’s scope, in contravention of the legislative intent. B. NMLA’s Definition of “Post Closing Solvency Covenant” Purchaser next argues that the covenant it seeks to enforce is not one of the “post closing solvency covenants” prohibited by the NMLA. See MCL § 445.1593.

The statute defines a “post closing solvency covenant” as: [A]ny provision of the loan documents for a nonrecourse loan, whether expressed as a covenant, representation, warranty, or default, that relates solely to the solvency of the borrower, including, without limitation, a provision requiring that the borrower maintain adequate capital or have the ability to pay its debts, with respect to any period of time after the date the loan is initially funded. The term does not include a covenant not to file a voluntary bankruptcy or other voluntary insolvency proceeding or not to collude in an involuntary proceeding.

MCL § 445.1592(d) (emphasis added). Borrower’s solvency covenant promises that Borrower “has not and shall not ... become insolvent or fail to pay its debts and liabilities from its assets as the same shall become due.” (R. 1-2, Mortg. § 4.2(j).) Purchaser maintains that this language falls outside the NMLA’s definition of “post closing solvency covenant” because “Borrower’s personal liability is not based ‘solely’ on ... insolvency”; Borrower’s willful or negligent failure to pay its debts would also appear to trigger recourse liability. (Purchaser’s Br. at 23.) But Purchaser’s argument ignores the facts.

As the district court noted, the Michigan Legislature phrased the NMLA’s definition of “post closing solvency covenant” inclusively. Thus, even though Borrower’s, solvency covenant employs the disjunctive “or” to indicate that it independently prohibits insolvency or the failure to pay debts for any other reason, this language easily fits within the definition of “post closing solvency covenant” to the extent it makes the loan subject to recourse upon insolvency. And Purchaser concedes that insolvency, rather than negligence or willfulness, caused Borrower’s default. (See R. 29-12, Pl.’s Resp. Defs.’ Interrogs.

No. 2.) Thus, we need not consider whether the NMLA would bar Purchaser’s deficiency action in the eounterfactual scenario of Borrower’s willful default. The district court committed no error in concluding that the NMLA bars Purchaser’s deficiency action. Thus, Purchaser must persuade us that the act violates either the Michigan or United States Constitution in order to pursue a deficiency judgment. IV. Purchaser first challenges the NMLA’s constitutionality under the Contract Clause. The United States Constitution forbids states from “passing] any ... Law impairing the Obligation of Contracts.” U.S. Const, art.

I, § 10, cl. 1. The Michigan Constitution includes a similar prohibition, Mich. Const, art. I, § 10, which courts interpret “no[ ] ... more expansively than its federal counterpart,” Attorney Gen. v. Mich. Pub. Serv. Comm’n, 249 Mich.App. 424, 642 N.W.2d 691, 698 (2002); see also Blue Cross & Blue Shield of Mich. v. Milliken, 422 Mich. 1, 367 N.W.2d 1, 14 (1985) (“The fedéral balancing approach has been adopted by our Court for purposes of adjudicating state Contract Clause claims as well as federal Contract Clause claims.”). In Energy Reserves Group, Inc. v.

Kansas Power & Light Co., the Supreme Court articulated a three-part inquiry for Contract Clause challenges. 459 U.S. 400, 103 S.Ct. 697, 74 L.Ed.2d 569 (1983). “The threshold inquiry is ‘whether the state law has, in fact, operated as a substantial impairment of a contractual relationship.’ ” Id. at 411, 103 S.Ct. 697 (quoting Allied Structural Steel Co. v.

Spannaus, 438 U.S. 234, 244, 98 S.Ct. 2716, 57 L.Ed.2d 727 (1978)). “If the state regulation constitutes a substantial impairment, the State, in justification, must have a significant and legitimate public purpose behind the regulation.Id. at 411, 103 S.Ct. 697 (internal citation omitted). “[T]he [final] inquiry is whether the adjustment of the rights and responsibilities of contracting parties is based upon reasonable conditions and is of a character appropriate to the public purpose justifying the legislation’s adoption.” Id. at 412, 103 S.Ct. 697 (quoting U.S. Trust Co. of N.Y. v.

New Jersey, 431 U.S. 1, 22, 97 S.Ct. 1505, 52 L.Ed.2d 92 (1977)) (internal punctuation omitted). A. Applicability of Energy Reserves' As a threshold matter, Purchaser contests the applicability of Energy Reserves to this case. Relying, on nineteenth-century Supreme Court precedent, Purchaser argues that debt contracts are special, characterizing the NMLA as a populist debt-relief law like those the Founders intended the Contract Clause to proscribe. See, e.g., Walker v.

Whitehead, 83 U.S. 314, 318, 16 Wall. 314, 21 L.Ed. 357 (1872) (invalidating a post-Civil-War Georgia law that would have allowed former slave owners to deduct the value of freed men and women from their debt obligations); Sturges v. Crowninshield, 17 U.S. 122, 197-98, 4 Wheat. 122, 4 L.Ed. 529 (1819) (articulating a narrow reading of the Contract Clause in striking down a New York law that absolved debtors from personal liability—in addition to releasing them from prison—upon the surrender of all their property).

But as the district court observed, the Supreme Court’s nineteenth-century treatment of debt contracts is not a principled basis on which to depart from the clearly established modern approach that balances the contractual right impaired against the public interest served. See Keystone Bituminous Coal Ass’n v. DeBenedictis, 480 U.S. 470, 503, 107 S.Ct. 1232, 94 L.Ed.2d 472 (1987) (“[O]ur cases construing the Clause, indicate that its primary focus was upon legislation that was designed to repudiate or adjust pre-existing debtor-creditor relationships that obligors were unable to satisfy.

Even in such cases, the Court has refused to give the Clause a literal reading.” (internal citations omitted)); Exxon Corp. v. Eagerton, 462 U.S. 176, 190, 103 S.Ct. 2296, 76 L.Ed.2d 497 (1983) (“This Court has long recognized that a statute does not violate the Contract Clause simply because it has the effect of restricting, or even barring altogether, the performance of duties created by contracts entered into prior to its enactment.”); Faitoute Iron & Steel Co. v. City of Asbury Park, 316 U.S. 502, 513, 62 S.Ct. 1129, 86 L.Ed. 1629 (1942) (“The dictum from Sturges v.

Crowninshield is one of those inaccurate generalizations that has gained momentum from uncritical repetition.”). Purchaser alternatively contends that the Energy Reserves framework applies only to laws enacted in response to crises, pointing out that the shift to modern Contract Clause jurisprudence coincided with the Court’s opinions upholding state laws enacted to combat the effects of the Great Depression. The Supreme Court flatly rejects this argument. E.g., Energy Reserves, 459 U.S. at 412, 103 S.Ct. 697 (“Furthermore, since [Home Bldg. & Loan Ass’n v.

Blaisdell, 290 U.S. 398, 54 S.Ct. 231, 78 L.Ed. 413 (1934) ], the Court has indicated that the public purpose need not be addressed to an emergency or temporary situation.”). In short,' Purchaser offers no persuasive justification for extending additional constitutional protection to debt contracts. We follow the Energy Reserves framework in assessing whether the NMLA comports with the Contract Clause. B.

Substantial Impairment The district court ended its Contract Clause inquiry at the first stage of the Energy Reserves framework, finding that the NMLA does not substantially impair Purchaser’s rights under the loan agreement. Although the Supreme Court has yet to develop a formal test for assessing the substantiality of a contractual impairment, the dividing line falls somewhere between the “[t]otal destruction of contractual expectations,” Energy Reserves, 459 U.S. at 411, 103 S.Ct. 697, and “[l]aws which restrict a party to those gains reasonably to be expected from the contract,” City of El Paso v.

Simmons, 379 U.S. 497, 515, 85 S.Ct. 577, 13 L.Ed.2d 446 (1965). In other words, an impairment takes on constitutional dimensions only when it interferes with reasonably expected contractual benefits. See U.S.

Trust, 431 U.S. at 21, 31, 97 S.Ct. 1505 (explaining that the Contract Clause does not protect “windfall benefits” resulting from the “unforeseen and unintended” effects of legislative schemes); Simmons, 379 U.S. at 514-15, 85 S.Ct. 577 (holding that no constitutional violation exists where the state abridges contractual rights that, if exercised, would render a promise illusory); Faitoute, 316 U.S. at 514, 62 S.Ct. 1129 (rejecting a challenge based on the impairment of “paper rights” with no practical value); City of Charleston v. Pub. Serv. Comm’n of W.

Va., 57 F.3d 385, 392 (4th Cir.1995) (“It is established, however, that a finding of ‘technical impairment’ is insufficient....”). In addition to the parties’ reasonable expectations, the Supreme Court considers objective evidence of reliance on the impaired term. See City of Charleston, 57 F.3d at 392-93 (listing examples). For example, the impaired term might induce a plaintiff to contract in the first place. See U.S.

Trust, 431 U.S. at 18-19, 97 S.Ct. 1505 (finding, based on the “chronology” of events, that New Jersey used a statutory covenant to induce individuals to purchase Port Authority bonds); Simmons, 379 U.S. at 514, 85 S.Ct. 577 (“We do not believe it can seriously be contended that the buyer was substantially induced to enter into these contracts on the basis of a defeasible right to reinstatement....”). Or a plaintiff might rely on the impaired term when carrying out his contractual obligations.

See Allied Structural Steel, 438 U.S. at 247, 98 S.Ct. 2716 (concluding that statutory pension-eligibility requirements created a substantial impairment where they would have left the plaintiff employer $185,000 short on its obligations as a result of the plaintiffs decade-long reliance on contractual terms when depositing funds into its employees’ pension trust).

In keeping with this precedent, the district court considered evidence of Morgan Stanley and Purchaser’s reliance and reasonable expectations, concluding that neither entity relied on the solvency covenant or reasonably expected the right to obtain a deficiency judgment under the circumstances presented here. (R. 51, Summ. J. Order at 29-30.) Several facts buttressed the court’s assessment:

(1) Morgan Stanley foreclosed on the Property and sold it at auction without seeking a deficiency;

(2) Morgan Stanley advertised Borrower’s loan as “non-recourse”;

(3) Morgan Stanley sold the property “for $756,000-more than $5 million less than it could have sought ... if the Loan allowed it to pursue the deficiency”; and

(4) Purchaser’s principals purchased the property without reading the loan documents. (Id. at 28-30.) Purchaser makes no effort to contest the district court’s findings. Instead, Purchaser argues that the district court erred in looking outside “the four corners of the parties’ contract” to assess the parties’ reliance and reasonable expectations. In support, Purchaser cites Michigan case law prohibiting courts from considering extrinsic evidence of the parties’ reasonable expectations in ascertaining their contractual obligations.

As a threshold matter, federal constitutional law — not Michigan contract law — governs this court’s substantiality analysis. Although we might have applied Michigan contract law to determine whether a contract existed or whether the NMLA impaired a contractual term, see Gen. Motors Corp. v. Romein, 503 U.S. 181, 187, 112 S.Ct. 1105, 117 L.Ed.2d 328 (1992); Irving Trust Co. v. Day, 314 U.S. 556, 561, 62 S.Ct. 398, 86 L.Ed. 452 (1942), these questions are not before us.

The inquiry into the substantiality of an impairment presumes interference with a contractual right and asks whether that interference deserves further constitutional scrutiny. See Simmons, 379 U.S. at 506—07, 85 S.Ct. 577 (“For it is not every modification of a contractual promise that impairs the obligation of contract under federal law, any more than it is every alteration of existing remedies that violates the Contract Clause.”); Faitoute, 316 U.S. at 514, 62 S.Ct. 1129 (rejecting a challenge based on the impairment of “paper-rights”).

Purchaser offers no federal authority prohibiting courts from looking beyond the “four corners” of the contract when determining whether an impairment is substantial enough to implicate the Constitution. Contrary to Purchaser’s suggestion, the Supreme Court consistently discusses extrinsic evidence when ascertaining substantiality. See, e.g., Energy Reserves, 459 U.S. at 413-16, 103 S.Ct. 697 (considering the history of federal and state energy regulation); Allied Structural Steel, 438 U.S. at 246-47, 98 S.Ct. 2716 (reviewing the economics of pension plans); U.S.

Trust, 431 U.S. at 18-19, 97 S.Ct. 1505 (examining the “chronology” of events surrounding a bond issue); Simmons, 379 U.S. at 510-14, 85 S.Ct. 577 (recounting the history and purpose of a Texas land-purchase program). This circuit likewise considers extrinsic evidence. See Toledo Area AFL-CIO Council v. Pizza, 154 F.3d 307, 323-24 (6th Cir.1998) (considering financial statistics showing the importance of wage checkoffs to union activities); cf. United States v. Cnty. of Muskegon, 298 F.3d 569, 585 (6th Cir.2002) (rejecting substantiality claim due to an absence of extrinsic evidence showing reliance).

Accordingly, the district court properly considered evidence of general practices in the CMBS loan industry when ascertaining the reasonableness of Purchaser’s expectations as well as evidence of Morgan Stanley and Purchaser’s behavior in determining their reliance on the impaired term. In light of this evidence, we agree with the district court that the NMLA, at' most, technically impaired Purchaser’s contractual rights. The nonrecourse promise performs a central function in CMBS loans by discouraging borrowers from committing “bad boy” acts.

If parties expected solvency covenants to permit recourse liability in the event of default, borrowers and their principals would no longer have an incentive to maintain single-purpose-entity status in difficult economic circumstances. Removing these incentives effectively eliminates the CMBS loan’s primary benefit to the lender: lowering the risk of a bankruptcy involving the collateral that upsets distributions to bondholders. Cf. Faitoute, 316 U.S. at 515, 62 S.Ct. 1129 (quoting W.B. Worthen Co. ex rel. Bd. of Comm’rs of Street Improvement Dist. No. 513 v.

Kavanaugh, 295 U.S. 56, 60, 55 S.Ct. 555, 79 L.Ed. 1298 (1935)) (explaining that courts may consider “the quality of an acceptable investment for a rational investor” in assessing whether an impairment is substantial). The NMLA ensures the continuing viability of this incentive while leaving lenders the right to foreclose on mortgaged properties: the original bargain. See U.S.

Trust, 431 U.S. at 31-32, 97 S.Ct. 1505 (rejecting the substantiality of lost windfall benefits); Simmons, 379 U.S. at 514-15, 85 S.Ct. 577 (rejecting a challenge where the impaired right effectively rendered illusory the purchaser’s promise to pay). The NMLA does not violate the Contract Clause. The district court correctly found no substantial impairment because Purchaser and its predecessor in interest neither reasonably expected nor relied upon the right to obtain a deficiency upon Borrower’s default. V. Purchaser next challenges the NMLA’s constitutionality under the Due Process Clause. U.S. Const, amend.

XIV, § 1; see also Mich. Const. art. I, § 17; Kelley ex rel. Mich. v. E.I. duPont de Nemours & Co., 786 F.Supp. 1268, 1282 (E.D.Mich.1992) (noting that Michigan follows federal due-process law). “[T]he retroactive aspects of economic legislation ... must meet the test of due process: a legitimate legislative purpose furthered by rational means.” Romein, 503 U.S. at 191, 112 S.Ct. 1105 (citation and internal punctuation omitted).

This test is not a demanding one: As the Supreme Court often has reiterated, the party challenging a legislative enactment subject to rational basis review must negative every conceivable basis which might support it. Under rational basis review, it is constitutionally irrelevant what reasoning in fact underlay the legislative decision. We will be satisfied with the government’s rational speculation linking the regulation to a legitimate purpose, even unsupported by evidence or empirical data.

Thus, if a statute can be upheld under any plausible justification offered by the state, or even hypothesized by the court, it survives rational-basis scrutiny. Am. Express Travel Related Servs. Co. v. Kentucky, 641 F.3d 685, 690 (6th Cir.2011) (internal citations and punctuation omitted). Applying this standard, the district court determined that “the [Michigan] Legislature was presented with substantial evidence that CMBS financing in Michigan — a multi-billion dollar industry— would collapse, which could be catastrophic to future development in the state.” (R. 51, Summ. J.

Order at 32-33.) The court then concluded that “[t]he means chosen to address these concerns-declaring that one form of [single-purpose-entity] covenant, which never was intended to trigger recourse liability, was invalid and unenforceable — were not arbitrary or irrational.” (Id. at 33.) Purchaser disputes the logic of this assessment.

It posits that “the concerns presented to the legislature involved the lending community not being willing to make future loans,” and argues that the “retroactive application of the NMLA does nothing to address those concerns.” (Purchaser’s Reply 13.) But Purchaser misapprehends the rationale behind the NMLA. Witnesses at the state senate hearing testified unanimously that solvency covenants — as construed in Cherryland I — undermined the bedrock principle that CMBS loans are nonrecourse. (R. 40-14, Mins. Mich. Senate Econ. Dev.

Comm, at 22.) Allowing third-party assignees standing in the lenders’ shoes to seek deficiency judgments from defaulting borrowers created enormous, unexpected contingent liabilities for the developers. (Id. at 18-19, 23.) As a result of these contingent liabilities, the legislature worried that banks would discontinue lending to developers with outstanding CMBS loans, or only do so at prohibitive interest rates. Less liquidity in the commercial real estate market would stall development and reverse Michigan’s fragile recovery from the 2008 recession.

By declaring solvency covenants unenforceable, therefore, the legislature not only protected developers from third-party assignees seeking windfall benefits, but also secured liquidity in the commercial real estate market to maintain an environment that supports development. Importantly, the Michigan Legislature gave the NMLA the necessary retroactive effect, thereby protecting developers with CMBS loans on the books from contingent liabilities.

The NMLA’s retroactive effect thus rationally addresses the fear that allowing deficiency judgments against defaulting borrowers would stifle Michigan’s multi-billion-dollar commercial real estate market. Purchaser also asks us to reject the state’s proffered rational basis on the ground that it is “based solely upon statements made by proponents of the law and ... totally unsupported by any competent ‘evidence.’ ” (Purchaser’s Br. at 47.) But supported by the evidence or not, the state offers a plausible, internally consistent rationale for enacting the NMLA. This is all the law requires. Am.

Express Travel Related Servs., 641 F.3d at 690. VI. With its last challenge, Purchaser contends that the NMLA violates the Michigan Constitution’s separation-of-powers provision because the statute requires courts to interpret loan contracts in a manner predetermined by the legislature. See Mich. Const, art. 3, § 2. Although this argument relies entirely on state law, Purchaser misses the fact that the Michigan Court of Appeals rejected an identical separation-of-powers argument in Cherryland II: However, the [NMLA] does not “interpret” the contract or direct this Court or any court to do anything.

It declares that the postclosing solvency covenant is invalid, unenforceable, and against public policy. This may have the effect of invalidating plaintiffs entitlements based on the contract, but if so it will be because the courts apply the new law, not because the Legislature has directly dictated the outcome in this case. 835 N.W.2d at 606. The court of appeals’s decision constitutes persuasive authority on this matter of state law, and Purchaser offers no reason to think that the Michigan Supreme Court would reach a different conclusion. See Kingsley Assocs., 65 F.3d at 507.

We therefore follow the district court in adopting Cherryland IPs rationale. The NMLA comports with Michigan’s separation-of-powers provision. VII. We AFFIRM the district court’s judgment.