Following the financial collapse of the Studebaker Company in 1963, more than 11,000 autoworkers lost 85 percent of their vested pension interest when the company's retirement plan was terminated. The resulting political pressure culminated in passage of the Employee Retirement Income Security Act of 1974, 29 U.S.C. §§ 1001 - 1461 (ERISA), which regulates private-sector pension and health funds. In addition to setting up requirements for defined pension-benefit plans, as part of ERISA Congress also created the Pension Benefit Guaranty Corporation (PBGC), which insures uninterrupted payment of benefits under those plans upon their termination. The program is designed to be self-financed, funded primarily by insurance premiums paid by sponsoring companies and also from assets acquired from terminated plans and recovered from underfunded plan sponsors when bankruptcy occurs. To keep premiums as low as possible, ERISA provides that the sponsor of a terminated plan and the "trades or businesses" related to the sponsor through ties of common ownership (known as "control group members") are jointly and severally liable to PBGC for underfunded benefit liabilities.

It was against this background that PBGC sued to collect more than $30 million in underfunded pension liabilities from Findlay Industries following the shutdown of its operation in 2009, apparently a casualty of the worsening economy at the time. When Findlay could not meet its obligations, PBGC looked to hold liable a trust started by Findlay's founder, Philip D. Gardner (the Gardner Trust), treating it as a "trade or business" under common control by Findlay. PBGC also asked the court to apply the federal-common-law doctrine of successor liability to hold Michael J. Gardner, Philip's son, liable for some of Findlay's debt. Michael, a 45 percent shareholder of Findlay and its former-CEO, had purchased Findlay's assets and started his own companies using the same land, hiring many of the same employees, and selling to Findlay's largest customer. 1 The district court refused to hold either the trust or Michael and his companies liable.

In determining whether the Gardner Trust was a "trade or business" under Findlay's common control, the district court rejected the approach of our sister circuits that apply a "categorical test" to determine liability. The categorical test treats any entity leasing to a commonly controlled entity as a trade or business under ERISA. Instead of the categorical test, the district court applied a fact-intensive test cribbed from Commissioner v. Groetzinger , 480 U.S. 23 , 24, 107 S.Ct. 980 , 94 L.Ed.2d 25 (1987), a case interpreting the term "trade or business" as used in the tax code, 26 U.S.C. §§ 162 (a), 62(a)(1). The court held, under the so-called " Groetzinger test," that the trust was not liable. Next, after analyzing the requirements for creating and invoking federal common-law principles of successor liability, the district court declined to apply successor liability in this case. We conclude that the district court erred on both fronts.

First, an entity that owns land and leases it to an entity under common control should be considered, categorically, a "trade or business" under ERISA. As noted below, this interpretation recognizes the differences between ERISA and the tax code, satisfies the purposes of ERISA, and brings this court into agreement with its sister circuits. In addition, under the facts of this case, successor liability is necessary to implement the fundamental ERISA policy of protecting employees, in part by guaranteeing that employers who have promised pensions uphold their part of the deal. Refusing to apply successor liability here would allow Findlay to make promises to employees, fail to uphold those promises, and then engage in clever financial transactions that leave PBGC to pay millions in pension liabilities. Holding Findlay responsible, on the other hand, is a commonsense answer that fulfills ERISA's goals.

We therefore find it necessary to reverse the rulings below and remand the case to the district court.

BACKGROUND

Statutory Background

Private employers are not required to offer pension plans, but if they do, ERISA requires that the pension plans meet certain standards and retain certain protections. That way, "if a worker has been promised a defined pension benefit upon retirement-and if he has fulfilled whatever conditions are required to obtain a vested benefit-he actually will receive it." Nachman Corp. v. Pension Benefit Guar. Corp. , 446 U.S. 359 , 375, 100 S.Ct. 1723 , 64 L.Ed.2d 354 (1980). Before ERISA, lack of oversight and legal standards often left pension plans without enough money, and employees who counted on those funds with nothing for retirement. Id. at 374-75 , 100 S.Ct. 1723 .

As a "major part of Congress'[s] response to [that] problem," ERISA instituted a termination-insurance program, PBGC. Id. at 375 , 100 S.Ct. 1723 . Although ERISA's funding, disclosure, and other standards made it more likely that pension plans would have the money that they had promised their beneficiaries, Congress built in the extra protection of PBGC-operated insurance. Subchapter III of ERISA requires PBGC to charge participating companies premiums so that if a pension plan fails, PBGC can "provide for the timely and uninterrupted payment of pension benefits to participants and beneficiaries." 29 U.S.C. § 1302 (a).

Despite the significant increases in coverage ushered in by ERISA, a few years after its introduction, PBGC warned Congress "that ERISA did not adequately protect plans from the adverse consequences that resulted when individual employers terminate their participation in, or withdraw from, multiemployer plans" set up under collective bargaining agreements. Pension Benefit Guar. Corp. v. R.A. Gray & Co. , 467 U.S. 717 , 722, 104 S.Ct. 2709 , 81 L.Ed.2d 601 (1984). In other words, the statute provided the necessary protection for when a company ended its own pension plan, but when multiple companies pooled assets into a single pension plan, a withdrawing employer risked saddling the remaining companies with all of the plan's liabilities. In response, Congress passed the Multiemployer Pension Plan Amendments Act of 1980, 29 U.S.C. §§ 1381 - 1461 (MPPAA), amending ERISA to ensure that multiemployer plans also served the statute's goals.

Under the MPPAA, multiemployer plans are subject to many of the same standards as single-employer plans. For example, the MPPAA requires multiemployer plans to pay premiums for PBGC insurance, just as single-employer plans are required to do. 29 U.S.C. §§ 1321 - 1322a. And PBGC holds employers directly liable for underfunded-but promised-benefits, interest, and penalties, whether the liable employer is part of a single-employer pension plan or a multiple-employer pension plan. 29 U.S.C. §§ 1362 , 1381.

Factual and Procedural Background

Findlay Industries was a company that produced auto parts before going out of business in 2009. Since 1964 it had offered pension benefits to some of its employees, and by the time production was stopped, its pension obligation was underfunded by millions of dollars. To satisfy that liability, PBGC looked to assets that might be treated as Findlay's-specifically, a trust started by Findlay's founder and assets purchased from the company by the founder's son in 2009.

The Trust: At the end of 1986, Findlay transferred two pieces of property to the company's founder and owner, Philip D. Gardner. Less than a month later, Gardner transferred the property to an irrevocable trust. The trust was to provide for Gardner's sisters through the end of their lives, at which point the trust was to be distributed equally to Gardner's two sons-Philip J. and Michael Gardner. In addition, son Philip J. was the trustee and Michael was his successor.

From at least 1993 until 2009, when Findlay folded, the trust leased the two plots of land back to Findlay. Thus, for the majority of the time that the trust existed, it was leasing back to Findlay the very land that Findlay, through Gardner, had donated to the trust. Gardner's last sister died in early 2014, and a month later the entire trust was split between his sons, who ran and owned a majority of Findlay in its final years.

The Assets: In May 2009, after Findlay failed, a company named F I Asset Acquisition LLC purchased all of the equipment, inventory, and receivables from two of Findlay's plants. The two plants contained all of Findlay's equipment and machinery of value. The sale had a price tag of $2.2 million in cash and $1.2 million in assumed trade debt. It appears that Findlay's former assets then were transferred from F I Asset Acquisition to Michael Gardner and another company owned entirely by Michael. Shortly after, Michael Gardner transferred the assets again, this time to two other of his recently formed companies-Back in Black and September Ends.

Every step of the sale went through the hands of Michael Gardner. For the relevant time leading up to the sale, until just two months before the sale in May 2009, Michael was Findlay's CEO and a director. And at all times, he was an owner of almost 45 percent of Findlay's stock. At the end of 2008, an outside company offered to purchase Findlay's assets. Weeks later, Michael-who was still at Findlay-made Findlay an offer on behalf of F I Asset Acquisition, a company of which he was also a member (and, at some point, its managing member). As a Findlay director, Michael did recuse himself from considering other companies' bids for Findlay's assets. But he still had access to information that Findlay received about the potential sale, including a letter from PBGC and a request from a potential purchaser for indemnification for pension-plan liabilities. A month after the potential purchaser requested indemnification, F I Asset Acquisition made an offer that did not assume the underfunded-pension liabilities. That offer clinched the sale, and the assets were transferred from Findlay to F I Asset Acquisition, then to another company owned by Michael (MJG Inc.), and finally to September Ends and Back in Black-the two companies that ended up with the Findlay assets. September Ends and Back in Black were owned and controlled by Michael Gardner-he owned 52 percent of the stock and his minor children owned the other 48 percent.

More than mere ownership passed from father's company to son's companies, however. Michael Gardner's new businesses were duplicates of Findlay in many ways. The two businesses-September Ends and Back in Black-each established a plant on one of the old Findlay lots. One of those companies rehired substantially all of the former Findlay employees, and the other rehired six of nine salaried employees and 15 of 25 hourly employees. The two new companies also started selling to Findlay's largest customer.

According to PBGC, Michael's gambit paid off. When he purchased Findlay's assets, Michael knew or should have known that Findlay was responsible for over $18 million in pension liabilities. But without accepting any responsibility for those liabilities, Michael paid only $3.4 million for the company. Strikingly, between May 2009 and December 2013, the net income-or bottom line-of Back in Black and September Ends was $11.9 million, more than triple the amount Michael had paid. A cynic might observe that Michael was, indeed, "back in the black."

Although the former Findlay assets were being used to turn a profit, Findlay's pension remained drastically underfunded; taking into account interest and fines, PBGC claimed that Findlay's liability was more than $30 million. To collect on that liability, PBGC brought this suit and, in 15 counts, alleged that ten defendants, all connected to Findlay, engaged in a number of internal structures, set-ups, and sales to avoid liability for the pensions formerly promised to Findlay employees. This appeal addresses three of those counts, III, IX, and XV-each of which was dismissed by the district court on a motion to dismiss brought under Rule 12(b)(6) of the Federal Rules of Civil Procedure. Because Count IX depends completely on Count III, we will not address it separately; thus, only two of the three dismissed counts are at issue here.

First, the PBGC complaint alleged that the trust Philip D. Gardner started in 1987 was jointly and severally liable for Findlay's pension liabilities. Specifically, the complaint alleged that the trust was under the control of Philip D. Gardner's sons, Michael and Philip J., who also controlled Findlay. And under the control of the Gardners, the trust leased land to Findlay for at least 16 years. Because the trust shared a "substantial economic nexus" with Findlay, the complaint alleged that Findlay and the trust were under common control. The complaint also alleged that the trust was a trade or business for ERISA purposes. Thus, as a trade or business, commonly controlled, the trust was jointly and severally liable for Findlay's liabilities.

In its motion to dismiss, the Gardner Trust argued that PBGC had relied on the wrong legal standard to determine liability. Specifically, the trust argued that PBGC's "substantial economic nexus" theory had been rejected as a test to show that an entity was a trade or business for ERISA's purposes. Instead, the trust argued, the proper standard was the fact-intensive analysis of Groetzinger , 480 U.S. 23 , 107 S.Ct. 980 , the tax case. Therefore, because PBGC had not provided an analysis under Groetzinger , the trust contended that the complaint must be dismissed.

In response, PBGC argued that Groetzinger 's application was limited to the tax code and did not provide the correct standard in this circumstance. Instead, PBGC explained, the court should apply the "categorical test," concluding that an entity is categorically a trade or business when that entity leases to a commonly controlled entity

The district court agreed with the trust. Recognizing that neither this court nor the Supreme Court has defined "trade or business" under ERISA, the district court started with the dictionary definition of each word. The court explained that the dictionary "defines 'trade' as 'the business or work in which one engages regularly' and 'business' as 'a usually commercial or mercantile activity engaged in as a means of livelihood.' " 2 The court held that Groetzinger 's test-that a person must regularly engage in the activity in question primarily for profit or income-embodies the "ordinary, common-sense meaning of the words at issue." Because the trust was created "with the express purpose of providing for the care and eventual funeral expenses of [Gardner's] sisters," the court concluded that neither the plain meaning of the words nor the Groetzinger test supported a conclusion that the trust was a trade or business under ERISA.

Rejecting PBGC's argument that the categorical test was appropriate, the court reasoned that the case law from other circuits adopting the categorical test arose under the MPPAA, and not under single-employer pension plans. Aside from describing the MPPAA as "a separate statutory scheme with its own legislative history and purpose," the court did not explain why MPPAA case law should not apply to single-employer cases. In any event, the court concluded that because "the purpose of the [trust's] rental activity was not to dissipate Findlay's assets or to profit Gardner" and because "there is no possibility the rental activity was used to dissipate or fractionalize the employer's assets, there can be no controlled group liability."

Next, the court addressed PBGC's contention that Michael and his companies should be held liable under the federal common law of successor liability. Specifically, PBGC had alleged that these defendants had notice of Findlay's pension-plan liabilities, knew that Findlay was unable to pay its liabilities, and that Back in Black and September Ends had substantially continued Findlay's operations.

The district court disagreed. Applying the disjunctive three-part test outlined by this court in DiGeronimo Aggregates, LLC v. Zemla , 763 F.3d 506 , 511 (6th Cir. 2014), the court held that successor liability does not justify the rare exercise of creating common law under ERISA. The court held that PBGC failed the first part of the test because ERISA is not silent as to who can be responsible for successor liabilities of single-employer plans. Specifically, the district court held, because a portion of the statute discusses the effects of corporate reorganization, Congress did not intend for liability of entities beyond what is listed in the statute.

The court next concluded that the second part of DiGeronimo 's test was not satisfied because there is no "awkward gap" to fill in the statute. In doing so, the court rejected PBGC's argument relying on cases from other circuits that found successor liability under ERISA. The court distinguished those cases, pointing out that they all arose under the MPPAA and thus applied only to multiemployer plans. The court reasoned that because the MPPAA did not address corporate reorganizations, common law played a necessary gap-filling role in those other cases. But, the court said, because the statute does address corporate reorganization of employers in single-employer plans, there is no awkward gap.

Finally, the court held that successor liability is not essential to carrying out fundamental ERISA policies. Because the fundamental policy of ERISA is to make sure that employees get their pensions, and PBGC already has a list of who it can hold accountable, the court observed that "[a]dding more targets is not necessary to fulfill ERISA's policy of protecting plan participants."

DISCUSSION

Standard of Review

We review a district court's dismissal under Federal Rule of Civil Procedure 12(b)(6) under a de novo standard of review. United Food & Commercial Workers Union-Emp'r PensionFund v. Rubber Assocs., Inc. , 812 F.3d 521 , 524 (6th Cir. 2016). We accept all well-pleaded allegations as true and "determine whether they plausibly state a claim for relief." DiGeronimo , 763 F.3d at 509 (quotation marks and citation omitted). The complaint must address all material elements of the plaintiff's chosen legal theory. Id. Either direct or inferential allegations will suffice. Id.

Appropriate Test to Determine Trust's Liability

When an employer terminates its pension plan, ERISA liability does not end with the company that actually promised pension payments. Instead, a "trade or business" under "common control" of the employer is treated as part of the employer and so incurs joint-and-several liability under ERISA. See 29 U.S.C. §§ 1362 (a), 1301(a)(14)(B), 1301(b)(1). This standard applies to both single-employer and multiemployer plans. See, e.g. , 29 U.S.C. § 1301 (a)(3), (b)(1). The Gardner Trust assumes but does not concede that it and Findlay were under common control. The trust contends, however, that it is not a trade or business under ERISA.

ERISA does not define "trades or businesses," and neither the Supreme Court nor this court have defined the phrase in the context of ERISA. The Supreme Court, however, has defined those terms as used in the Internal Revenue Code. In Groetzinger , the Court applied a fact-intensive test to determine what constitutes a trade or business, examining (1) the primary purpose of the entity in question and (2) whether the entity's activity is continuous and regular. 480 U.S. at 35 , 107 S.Ct. 980 . Despite the Court's warning that its interpretation of "trade or business" was confined to "specific sections" of the tax code, id. at 27 n.8, 107 S.Ct. 980 , some courts have relied on Groetzinger to define the same terms under ERISA. See, e.g. , UFCW Local One Pension Fund v. Enivel Props., LLC , 791 F.3d 369 , 375 (2d Cir. 2015). But other courts, including some that have otherwise relied on Groetzinger , have eschewed the Groetzinger test when the entity-in-question's activity is leasing property to a company under common control. See, e.g. , Cent. States, Se. & Sw. Areas Pension Fund v. Nagy , 714 F.3d 545 , 551 (7th Cir. 2013). Those courts, instead, have concluded that the entity that leases property to its commonly controlled company is categorically a trade or business for ERISA purposes.

The first step in statutory construction is to "determine whether the language at issue has a plain and unambiguous meaning with regard to the particular dispute in the case." Fullenkamp v. Veneman , 383 F.3d 478 , 481 (6th Cir. 2004) (quoting Barnhart v. Sigmon Coal Co., Inc. , 534 U.S. 438 , 450, 122 S.Ct. 941 , 151 L.Ed.2d 908 (2002) ). The district court thus turned to a dictionary and reasoned that a "trade" is "the business or work in which one engages regularly" and that "business" is "a usually commercial or mercantile activity engaged in as a means of livelihood." Without any additional explanation, the court concluded that Groetzinger 's fact-intensive test "embodies this ordinary, common-sense meaning of the words at issue." The court then looked to the trust's "express purpose of providing for the care and eventual funeral expenses of [Gardner's] sisters" and concluded that there was "no possibility [that] the rental activity was used to dissipate or fractionalize the employer's assets." Thus, the court ruled, the trust was not a trade or business.

But, contrary to the district court's conclusion, the dictionary does not provide us the "plain and unambiguous meaning" that one might seek. See Fullenkamp , 383 F.3d at 481 . Both "trade" and "business" are broad terms, susceptible to a range of meanings. For example, "business" is defined as "a commercial or sometimes an industrial enterprise," MERRIAM-WEBSTER'S COLLEGIATE DICTIONARY (11th ed. 2006), "[c]ommercial, industrial, or professional dealings," AMERICAN HERITAGE COLLEGE DICTIONARY (4th ed. 2002), or "[c]ommercial transactions," BLACK'S LAW DICTIONARY (10th ed. 2014). "Trade" is defined as "[t]he business of buying and selling or bartering goods or services," BLACK'S LAW DICTIONARY (10th ed. 2014), and "[t]he business of buying and selling commodities; commerce," AMERICAN HERITAGE COLLEGE DICTIONARY (4th ed. 2002).

In light of the breadth of these definitions, Groetzinger 's test does not, as the district court held, embody the "ordinary, common-sense meaning" of "trade" or "business." Quite the opposite. By reasoning that "not every income-producing and profit-making endeavor constitutes a trade or business" and that "transactions entered into for profit" are not necessarily trade or business, Groetzinger highlights the fact that tax law's treatment of the terms "trade" and "business" does not, like many of the dictionary definitions, rely on a broad idea of "commerce" but, rather, is narrow and specific to tax law. 480 U.S. at 35 , 107 S.Ct. 980 .

What is more, in an application that the Supreme Court recognized was unique not only to tax law but also to one specific passage of the tax law, Groetzinger requires the court to determine the primary purpose of an activity. Id. at 27 n.8, 35, 107 S.Ct. 980 . The district court did not explain why the dictionary definitions it cited support a legal test that turns on the primary purpose of the entity in question. And moreover, reading a primary-purpose requirement into the statutory language would create dangerous incentives and would not serve ERISA's purposes.

Under the district court's decision, as long as the primary reason for dissipating one's assets was not to escape liability under ERISA, those assets would be shielded from a plan sponsor's liability. But companies can have more than one reason to dissipate assets. For example, if the owner of a construction business was personally stressed and put the majority of his company's assets into opening a bakery because baking was soothing to him, PBGC would have to pick up the tab when the construction company's pension was not funded because the primary purpose behind the bakery was stress relief. Under Groetzinger , it would not matter that the baker's secondary purpose could have been to shield his company's assets from ERISA liability for the underfunded pension. Or, as could have been the case here, one could want to stow company assets safely in trust and provide for the well-being of loved ones. There is nothing in the record that proves that avoiding ERISA liabilities was indeed Philip D. Gardner's motivating force. But if Groetzinger controls, entities certainly would be encouraged to try such reorganization and would not be held liable for it as long as they had a different primary purpose.

Not only would application of Groetzinger create dangerous incentives, it would not serve ERISA's purposes. Under ERISA, whether Gardner's primary motivation was to dissipate Findlay's assets is not important. What is important is determining whether those assets were effectively Findlay's and thus should be used to help pay what Findlay promised its employees. The commonsense conclusion is yes: when a business gives land to the business's sole owner, who then puts it in a trust-run by his sons-which then leases the land back to his business, that land never stopped being a part of the company's functional assets.

For all of these reasons, there is no plain and unambiguous reading of ERISA that supports adopting Groetzinger . But our analysis does not stop there. When, as here, the meanings of the words at issue are not plain and unambiguous, we turn to the purpose and the structure of a statute to determine the meaning of the terms at issue. See Fullenkamp , 383 F.3d at 483 .

Structurally, ERISA holds employers liable for the promises of pensions that they make to employees. After a PBGC determination that a pension plan has insufficient assets to meet its liabilities, 29 U.S.C. § 1341 (c), ERISA holds the plan sponsor liable, 29 U.S.C. § 1362 (a). The statute then guarantees that a liable sponsor cannot evade its responsibility through tactics such as corporate reorganization, see 29 U.S.C. § 1369 (b), or sales to avoid liability for an impending plan termination, see 29 U.S.C. § 1369 (a). And although PBGC exists to ensure that employees receive the pensions that they were promised, ERISA holds the employers primarily accountable and relies on PBGC to pay only as a last resort. To that end, ERISA enforces employers' promises by extending liability for those promises to commonly controlled entities. 29 U.S.C. § 1362 (a). Indeed, "the primary purpose of the common control provision is to ensure that employers will not circumvent their ERISA and MPPAA obligations by operating through separate entities." Mason & Dixon Tank Lines, Inc. v. Cent. States, Se. & Sw. Areas Pension Fund , 852 F.2d 156 , 159 (6th Cir. 1988). Put another way, ERISA generally seeks to hold employers liable for their promises to employees; the common-control rules stop employers from escaping that liability by spreading their assets.

In light of the purpose and the structure of the ERISA provisions at issue, we hold that the categorical test applies. That test concludes simply that any entity that leases property to a commonly controlled company is categorically a trade or business for ERISA purposes. Cent. States Se. & Sw. Areas Pension Fund v. Messina Prods., LLC , 706 F.3d 874 , 882 (7th Cir. 2013). In doing so, the categorical test stops leases between commonly controlled entities as a way of offering those entities protection from ERISA liability at very little risk. The facts here highlight how. By giving the land to a commonly controlled entity, Findlay guaranteed that it still had the benefit of use (and likely control) of the land, just the same as if it had never given the land away at all. But now, the land did not technically belong to Findlay, so it did not count among Findlay's assets. Thus, Findlay had all of the meaningful benefit of the land, but none of the risk or responsibility that came with outright ownership. And the Gardner Trust did not have to put in any of the effort or face any of the risk of an arms-length leasing arrangement with a lessee that was not under common control. This situation is precisely the type that the common-control rules exist to prevent.

Applying the categorical test also aligns us with other courts around the country. Indeed, the defendants here were unable to present a single case in which leasing between commonly controlled entities did not result in an entity being a trade or business for ERISA purposes. And we have found none. PBGC, on the other hand, cites cases of other circuits and of district courts in and outside of this circuit that support its conclusion. The Seventh Circuit, for example, adopted the categorical test, concluding that "the likelihood that a true purpose and effect of the 'lease' is to split up the withdrawing employer's assets is self-evident." Messina Prods. LLC , 706 F.3d at 882 . The Eighth Circuit noted that the business-or-trade inquiry is a factual inquiry but then upheld the district court's categorical conclusion that leasing between commonly controlled entities "established the existence of a trade or business for ERISA purposes." Vaughn v. Sexton , 975 F.2d 498 , 503 (8th Cir. 1992). The Ninth Circuit has gone even further, concluding that leasing for profit "is plainly sufficient" to be a trade or business-regardless of whether the investments are active or passive and regardless of whether the lease was to a commonly controlled entity. Bd. of Trustees of W. Conf. of Teamsters Pension Trust Fund v. Lafrenz , 837 F.2d 892 , 894-95 (9th Cir. 1988).

Rejecting this case law, the district court reasoned that the cases relied on by PBGC arose under the MPPAA and did not address single-employer plans. But the goal of stopping employers from splitting their assets to escape liability is equally as important for single-employer plans as it is for multiemployer plans. Neither the district court nor defendants provided any reason why multiemployer plans should be treated any differently; thus, neither provided any grounds for limiting the extensive case law outlined above to cases arising under the specific portions of ERISA that address multiemployer plans. And upon reflection, we cannot think of any. After all, the rules against dissipating assets are meant to protect both "ERISA and MPPAA obligations." Mason & Dixon Tank Lines, Inc. , 852 F.2d at 159 (emphasis added).

Defendants argue that we should follow three other circuits to adopt Groetzinger . 3 But only one of those cases involved a lease to a commonly controlled entity, and that case does not help defendants' argument: The Seventh Circuit reasoned that Groetzinger is the general standard to apply, but that the categorical test is appropriate for leases between commonly controlled entities. Messina Prods., LLC , 706 F.3d at 882-83 . Hence, we do not need to decide whether Groetzinger applies to leases made outside of a commonly controlled group-as the Seventh and Second Circuits have done-or whether any leasing activity is a trade or business-as the Ninth Circuit has done. It is sufficient here-and does not conflict with any sister circuits-to join the courts that have held that leasing to a commonly controlled entity is categorically a trade or business for ERISA purposes.

Defendants' remaining arguments fare no better. First, defendants make a single-paragraph argument that the ordinary meaning of trade or business does not include leasing, but they do not provide any support for that counterintuitive conclusion. Next, relying on ERISA's purpose of stopping employers from avoiding liability by dissipating assets, defendants contend that the facts prove that Philip D. Gardner did not intend to use the trust to avoid ERISA liabilities. But beyond asking this court to repeat the district court's error and draw factual conclusions in the defendants' favor, 4 that argument is an attempt to get Groetzinger 's fact-intensive analysis in through the back door. As explained above, such a fact-intensive test does not serve ERISA's purposes and, instead, would create significant problems with its administration.

On that note, Judge McKeague's concurring opinion, concerned that the categorical test will lead to unfair results, asks us to tread carefully and adopt a less-than-categorical version of the test. He urges us to imagine an alternative set of facts in which the trust vested in the heirs of the sisters, as opposed to the Gardners. How would it be fair, he aptly inquires, to take the trust assets from the sisters' heirs, who have nothing to do with Findlay? However, the legitimate concern raised in the concurrence is already contemplated by the common-control rules.

The rules for common control apply complex regulations to determine who has an actuarial interest in the trust, and thus how much of the trust the ultimate beneficiary is considered to own at any given time. See 26 C.F.R. §§ 1.414 (c)-4(b)(3), 1.414(c)-2. One look at PBGC's complaint shows that PBGC has applied those rules to allege properly that the Gardners were in common control of both Findlay and the trust. If instead of the Gardners, the sisters' heirs were the ultimate beneficiaries-as Judge McKeague hypothesizes-there would be no common control between Findlay and the trust. And because both entities being under common control is a prerequisite for the categorical test, the categorical test would not apply. Judge McKeague's concern over a possible inequitable result for "an 'innocent' third party," should be allayed by taking the common-control regulations into account.

Successor Liability

According to PBGC's complaint, Findlay owes more than $30 million in pension liability. Yet Michael Gardner purchased all of Findlay's valuable assets for only $3.4 million and within four-and-a-half years had turned a nearly $12 million profit using Findlay assets, employing former Findlay employees, making former Findlay products, and selling to Findlay's biggest customer. PBGC does not contend that the transfer of Findlay assets to Michael and his companies made Michael or his companies liable under 29 U.S.C. § 1369 (b), the section of ERISA that asserts liability for certain corporate reorganizations. Instead, PBGC asked the district court to rely on federal common law's treatment of successor liability to hold Michael and his companies accountable for Findlay's liability. The district court declined to do so.

The district court was correct to reason that the creation of common law under ERISA is something to be done in narrow circumstances. But because the federal-common-law doctrine of successor liability serves fundamental ERISA policies, we conclude that the creation and application of federal common law is appropriate in this case.

"At the time of ERISA's enactment, Congress in general encouraged the courts to develop a federal common law of employee benefits because many issues relating to employee benefits would arise where there would be no specific rule to govern the question." DiGeronimo , 763 F.3d at 510-11 . But "where Congress has established an extensive regulatory network ... courts do not lightly create additional rights under the rubric of federal common law." Id. at 511 .

To satisfy those competing interests, we have developed a three-part standard to determine whether and when it is appropriate to create federal common law under ERISA. We undertake such a step if (1) ERISA is silent or ambiguous on the issue before the court, (2) there is an awkward gap in the statutory scheme, or (3) "federal common law is essential to the promotion of fundamental ERISA policies."

Local 6-0682 Int'l Union of Paper v. Nat'l Indus. Grp. Pension Plan , 342 F.3d 606 , 609 (6th Cir. 2003) (quotation marks and citation omitted). The standard is phrased in the disjunctive so that if any one of the three circumstances is present, creation of federal common law is appropriate.

PBGC contends that all three circumstances are present here. The defendants, unsurprisingly, agree with the district court that none of the three are. We must resolve that precise dispute, because we conclude that the federal common law of successor liability is necessary to promote fundamental ERISA policies in this case. Hence, we need not address the other prongs of the standard.

ERISA's fundamental protections of employment benefits function in two ways: guaranteeing that employees receive the benefits they were promised and making sure that employers keep up their end of the deal. To that end, the official policy of ERISA is to protect "the interests of participants in employee-benefit plans and their beneficiaries" while "establishing standards of conduct, responsibility, and obligation for fiduciaries of employee benefit plans." 29 U.S.C. § 1001 (b).

Additionally, 29 U.S.C. § 1302 (a) creates PBGC and explains the purpose of Subchapter III of ERISA-Plan Termination Insurance. The purposes of Subchapter III-which gives PBGC the power to sue and lists the liabilities for which it can sue-include that PBGC "encourage the continuation and maintenance of voluntary private pension plans for the benefit of their participants" and "provide for timely and uninterrupted payment of pension benefits to participants and beneficiaries." 29 U.S.C. § 1302 (a). But most important, one of PBGC's purposes is to maintain the lowest possible PBGC insurance premiums. Id. The way that PBGC maintains low premiums is to pay out the lowest possible amount by holding employers liable for their promises to employees.

Thus, when 29 U.S.C. § 1001 (b) and § 1302(a) are viewed together, it is clear that PBGC enforcing employers' own promises to their employees is a fundamental premise of ERISA. The district court was correct to acknowledge that a fundamental policy of ERISA is to protect employees but wrong, however, to ignore the fundamental policy of PBGC's enforcement powers and instead treat successor liability as a desire to go after "more targets."

Successor liability promotes fundamental ERISA policies by guaranteeing that substance matters over form. Taking the complaint in this case as true, it appears that Michael Gardner had extensive information about Findlay's debts and pension funding. As Findlay's CEO, board member, and 45 percent shareholder, Michael offered to purchase Findlay's assets but refused to take on any pension liability. The assets that his company purchased for $3.4 million netted his two companies nearly $12 million in four-and-a-half years. And the companies operated from two former Findlay sites, with former Findlay employees, making the same products, and selling to Findlay's principal customer.

Because Michael purchased the assets-although he did so in a way that does not represent an arms-length sale-none of the provisions of § 1369(b) apply to him. But this result is certainly the kind of transaction that frustrates the fundamental policies of ERISA: Findlay did not keep its promises to its employees, and instead of using its assets to meet its obligations, it sold the assets to its CEO, who then left the government to pay millions of dollars in pension liabilities.

Not only does successor liability promote fundamental policies of ERISA, refusal to apply the principles of successor liability here would frustrate ERISA policies. If there is no successor liability here, this case will provide an incentive to find new, clever financial transactions to evade the technical requirements of ERISA and, thus, escape any liability-a result that flies in the face of § 1001(b). And if employers can so easily escape millions of dollars in liabilities, PBGC will be left to pay the underfunded pension benefits. That situation will force PBGC to raise its rates, which will strain still-existing plans further, and which risks forcing them to be underfunded and possibly fail. Such a result plainly would frustrate the purpose of Subchapter III. 5

In choosing the form of successor liability to apply in this case, we opt for the test developed under different provisions of federal labor and employment law. As PBGC points out, "ERISA's broad preemption provision makes it clear that Congress intended to establish employee benefit plan regulation as an exclusive federal concern, with federal law to apply exclusively, even where ERISA itself furnishes no answer." In re White Farm Equip. Co. , 788 F.2d 1186 , 1191 (6th Cir. 1986). In certain circumstances, such as in contract interpretation, "the federal court may take direction from the law of the state in which it sits" so long as "the rule used [is] the one that best comports with the interests served by ERISA's regulatory scheme." Regents of Univ. of Mich. v. Emps. of Agency Rent-A-Car Hosp. Ass'n , 122 F.3d 336 , 339 (6th Cir. 1997) (internal quotation marks and citation omitted). But, as a general matter, the court must look to the federal common law and should draw guidance from state common law only when federal common law does not provide an established standard. See Tinsley v. Gen. Motors Corp., 227 F.3d 700 , 704 (6th Cir. 2000).

Because there is a body of federal common law applying successor liability in employment and labor cases, it is appropriate to apply that law here, too. Successor liability is an equitable doctrine that requires the court to balance (1) the interests of the defendant, (2) the interests of the plaintiff, and (3) "the goals of federal policy, in light of the particular facts of a case and the particular legal obligation at issue." Cobb v. Contract Trans., Inc. , 452 F.3d 543 , 554 (6th Cir. 2006) (applying successor liability to the Family and Medical Leave Act).

Furthermore, adopting the federal common law of successor liability would best serve ERISA's purposes. "ERISA's goal, [the Supreme] Court has emphasized, is uniform national treatment of pension benefits." Raymond B. Yates, M.D., P.C. Profit Sharing Plan v. Hendon , 541 U.S. 1 , 17, 124 S.Ct. 1330 , 158 L.Ed.2d 40 (2004) (internal quotation marks and citation omitted). By applying the federal common law of successor liability, this court also will align itself with the Seventh and the Ninth Circuits, both of which have done so in MPPAA cases. See Resilient Floor Covering Pension Trust Fund Bd. of Trustees v. Michael's Floor Covering , 801 F.3d 1079 , 1095 (9th Cir. 2015) ; Upholsterers' Int'l Union Pension Fund v. Artistic Furniture , 920 F.2d 1323 , 1327 (7th Cir. 1990).

The defendants argue that if the court applies the common law of successor liability, it should apply Ohio common law and not federal common law. Instead of explaining why that rule would best serve ERISA's purposes, the defendants contend that the doctrine of federal successor liability is too broad and will "disrupt[ ] the settled commercial expectations of parties who purchase assets ... under state law." But this generic appeal to settled commercial expectations fails for two reasons. First, the defendants do not explain what their settled commercial expectations were and why this court should protect them. Perhaps for good reason: The complaint alleges that Michael Gardner underpaid for the profitable parts of Findlay-the company he ran-turned a hefty profit using those assets, and knowingly left the government to pay millions of dollars in Findlay's unkept pension promises. If true, those actions do not reflect commercial expectations that this court should ever protect, certainly not under ERISA. Second, the fear that applying successor liability will upset settled commercial expectations more generally is unfounded. Finding successor liability here does not mean that successor liability applies in every instance. All that we decide today is that when there is a sale that is not conducted at arm's length, successor liability can apply. And although we are reluctant to impose successor liability to reorganizations of failing businesses, that principle cannot be stretched so far as to demand judicial approval of deals that are not above board.

In a similar vein, instead of justifying the reliance on Ohio law, the defendants repeat their factual assertions, totally outside of the record, that Michael's companies would have failed had Michael not "taken the risk of purchasing these plants' assets ... and then succeeded in turning a profit and keeping employees in their jobs." The defendants then conclude-again, without any support in the record-that applying federal successor liability "would effectively bankrupt the Companies, put its employees out of work, and discourage the purchase and reorganization of failing businesses."

Although it is improper to consider the defendants' statement that the two companies will go out of business, the argument brings up a point that is worth acknowledging. It is true that this court is "reluctant to impose successor liability when it might inhibit the reorganization of failing businesses." Peters v. N.L.R.B. , 153 F.3d 289 , 301 (6th Cir. 1998). But as noted above, successor liability is an equitable doctrine. Cobb , 452 F.3d at 554 . As such, its application will balance the interests of both parties-protecting asset purchasers from being blindsided by massive liabilities, and guaranteeing that employers cannot easily avoid their ERISA obligations through clever financial transactions.

CONCLUSION

We conclude that the district court's decision is flawed in two respects. First, an entity that leases property to an entity under common control should be considered a "trade or business," categorically. This reading of the statute recognizes the differences between ERISA and the tax code, satisfies the purposes of ERISA, and brings this court in line with its sister circuits. Next, in this specific instance, successor liability is required to promote fundamental ERISA policies. Refusing to apply successor liability would allow employers to fail to uphold promises made to employees and then engage in clever financial transactions to leave PBGC paying out millions in pension liabilities. Holding the employers responsible, on the other hand, is a commonsense answer that fulfills ERISA's goals.

We therefore VACATE the district court's order of dismissal and REMAND the case for further proceedings.

DAVID W. McKEAGUE, Circuit Judge, concurring in part and dissenting in part.

CONCURRING IN PART AND DISSENTING IN PART

Findlay Industries ("Findlay") went out of business in 2009 with over $30 million in unfunded pension liabilities. The Pension Benefit Guaranty Corporation ("PBGC") picked up the tab and filed this lawsuit to recoup those losses. The issues raised by this interlocutory appeal involve two defendants: (1) A trust that obtained property from Findlay's founder but then leased it right back to Findlay ("the Trust"); and (2) the companies that eventually acquired all of Findlay's assets after it went under ("the Successors"). Neither the Trust nor the Successors expressly assumed-or believe they must assume-Findlay's pension liabilities. The district court agreed with the defendants and dismissed the PBGC's claims against them.

The majority concludes that the PBGC may sue both defendants. I agree with the majority that the claims against the Trust were improperly dismissed. However, since Congress deliberately chose not to impose liability on entities like the Successors in this case, I respectfully dissent from the majority's decision to revisit that policy judgment through the federal common law.

I

I agree that the Trust is a trade or business subject to common-control liability. However, I am hesitant to adopt the Categorical Test advocated by the PBGC as the rule for all future cases. Instead, I would follow a more circumscribed approach.

A

There are two kinds of ERISA liability relevant to this case. The first is "Termination Liability," which attaches when any plan terminates without enough funds to satisfy its obligations. 29 U.S.C. §§ 1307 (e)(2), 1362(a)(1). The second is "Withdrawal Liability," which attaches when one employer in a multiemployer pension plan leaves the group. Id. § 1381(a). Congress created Withdrawal Liability after the PBGC informed it that some employers were sending group pension plans into a death spiral by withdrawing their contributions. Mason & Dixon Tank Lines, Inc. v. Central States , 852 F.2d 156 , 158 (6th Cir. 1988) (discussing the Multiemployer Pension Plan Amendments Act, or "MPPAA"). Congress also empowered the PBGC-a government guarantor of pension payments-to sue any "employer" to which either form of liability attaches to recoup its losses. Id. at 158-60 .

Further, all "trades or businesses" under the common control of the ERISA plan sponsor are deemed to be one employer for liability purposes. Id. at 159 (quoting 29 U.S.C. § 1301 (b)(1) ("[A]ll employees of trades or businesses (whether or not incorporated) which are under common control shall be treated as employed by a single employer and all such trades or businesses as a single employer.")). This doctrine is colloquially known as "common-control liability." The point of common-control liability is to "ensure that employers will not circumvent their ERISA and MPPAA obligations by operating through separate entities." Id.

This statute and its accompanying regulations should not be applied in a wooden, formalistic manner. In re Challenge Stamping Co. , 719 F.2d 146 , 151 (6th Cir. 1983). Thus, we've previously held that when the bankruptcy laws deny the defendant actual control over the relevant entity, the PBGC cannot hold them liable, even if the defendant meets all the statutory and regulatory requirements for control. Id. We justified this rule by asserting that "Congress sought to determine the plain fact of control, rather than any subjective motives or reasons for control ... There is no support for a view that Congress's chief intent in employing [the common-control test] was to invade the deepest pocket in a business failure. ... The purpose of [the regulation] is obviously to find the party in control ." Id.

B

Not all entities under common control with the plan sponsor are subject to ERISA's common-control doctrine. Central States v. Messina Prods., LLC , 706 F.3d 874 , 880 (7th Cir. 2013). Only those entities that can fairly be said to be part of the common owner's trade or business (as opposed to mere investments) are governed by § 1301(b)(1), because ERISA does not abrogate the ordinary rule that shareholders are not personally liable for the obligations of a corporation. Id.

The issue here is whether a commonly controlled family trust qualifies under this rule. Specifically, the trust in this case obtained property from the plan sponsor-through its CEO-and then immediately leased it back to the plan sponsor in exchange for rent. Although the primary purpose of the trust was to provide for the well-being of the CEO's sisters during their life, the property reverted back to the CEO's sons upon the sisters' death. The sons, in turn, were the sole trustees and assumed control of the plan sponsor when their father retired. The parties have stipulated to the issue of common control for the purposes of this interlocutory appeal.

This is an issue of first impression for the Sixth Circuit. All the other circuits to have considered the issue hold that these leaseback arrangements are categorically a trade or business under ERISA, and the PBGC urges us to follow that rule ("the Categorical Test"). The Trust asks us to adopt a narrower, fact-intensive test originating from the Supreme Court's interpretation of the tax code ("the Groetzinger Test"). But I agree with the majority that the Groetzinger Test is a bad fit for these questions, and so I would apply a modified Categorical Test to cases like this one.

1

The Trust offers the Groetzinger Test as an alternative to the Categorical Test. But it makes no sense to apply Groetzinger here.

First, Groetzinger was about income taxes, and the Court expressly limited its holding to the sections of the Internal Revenue Code ("IRC") examined in that case. Commissioner v. Groetzinger , 480 U.S. 23 , 27 n.8, 107 S.Ct. 980 , 94 L.Ed.2d 25 (1987). As a matter of common sense, the Trust's invitation for us to disobey the Court's characterization of its own holding is ill-advised. The IRC uses the phrase "trade or business" about fifty times. If the Court was wary about defining the term throughout the IRC in one fell swoop, we should be even more skeptical when asked to export the meaning to a different statute entirely. And, of course, the same words can sometimes mean different things in different parts of the U.S. Code. See Nat'l Fed'n of Indep. Business v. Sebelius , 567 U.S. 519 , 544-45, 132 S.Ct. 2566 , 183 L.Ed.2d 450 (2012).

Second, the context of ERISA differs significantly from the sections interpreted by Groetzinger . That case involved the IRC's deduction for expenses "attributable to a trade or business carried on by the taxpayer." 26 U.S.C. §§ 62 (1), 162(a). These provisions are exculpatory; they reduce a person's tax liability. And the Court has held that tax exemptions and deductions must be construed strictly against the taxpayer. United States v. Burke , 504 U.S. 229 , 112 S.Ct. 1867 , 119 L.Ed.2d 34 (1992) (Souter, J., concurring in the judgment); United States v. Wells Fargo Bank , 485 U.S. 351 , 108 S.Ct. 1179 , 99 L.Ed.2d 368 (1988).

In contrast, the common-control provisions of ERISA are inculpatory and remedial. They exist to "fence in" employers who fragment their ownership to try and avoid contractual obligations. Mason & Dixon , 852 F.2d at 159 . As remedial sections, 1 they should therefore be construed broadly when their meaning is unclear. See A-T-O, Inc. v. PBGC , 634 F.2d 1013 , 1020 (6th Cir. 1980) ; Rettig v. PBGC , 744 F.2d 133 , 155 n.54 (D.C. Cir. 1984) ; see also Nachman Corp. v. PBGC , 446 U.S. 359 , 374, 100 S.Ct. 1723 , 64 L.Ed.2d 354 (1980) (cataloguing the remedial goal of ERISA).

The cases that the Trust cites do not alter the analysis. Indeed, in every published case where the Trust says a Circuit has applied the Groetzinger Test, the facts are starkly different. As explained in more detail below, the courts draw a sharp distinction between (1) cases where a common owner leases property back to the plan sponsor, and (2) cases where the common owner leases property to an unrelated third party. The Groetzinger Test has only been applied in the latter cases. See UFCW Local One Pension Fund v. Enivel Props., LLC , 791 F.3d 369 , 371 (2d Cir. 2015) ; Central States v. Fulkerson , 238 F.3d 891 , 895 (7th Cir. 2001) ; Central States v. White , 258 F.3d 636 , 644 (7th Cir. 2001) ; Connors v. Incoal, Inc. , 995 F.2d 245 , 246 (D.C. Cir. 1993). Since this case fits in the former category, the latter cases are not helpful. The Categorical Test, however, presents other problems.

2

The PBGC insists that all commonly controlled entities that lease property back to the plan sponsor are "categorically" trades or businesses under ERISA. It cites a litany of Circuit Court 2 and District Court cases in support of its conclusion. While the test's pedigree is impressive, I think we should apply it with care.

The most helpful case for the PBGC is Vaughn v. Sexton , 975 F.2d 498 , 502-03 (8th Cir. 1992). In Vaughn , the Eighth Circuit held that a family trust, by leasing to the plan sponsor, was categorically a "trade or business" under ERISA. Id. at 503 . However, the court offered little support for its conclusion, and it made no attempt to fend off the trust's argument that "its primary purpose was not to generate income or profit but instead to assist in ... estate planning arrangements." Id. Instead, the court cited a list of cases involving business leases and concluded that they were persuasive.

Every other Circuit Court case cited by the PBGC involved business leases where the lessor was either an individual or a commonly controlled, for-profit business. Messina Prods. , 706 F.3d at 882-83 (personal leases to the plan sponsor); Central States v. Nagy , 714 F.3d 545 , 546 (7th Cir. 2013) (same); Central States v. Slotky , 956 F.2d 1369 , 1371 (7th Cir. 1992) (individual leasing buildings to plan sponsor); Bd. of Trustees of the Western Conf. of Teamsters Pension Fund v. Lafrenz , 837 F.2d 892 , 893 (9th Cir. 1988) (single proprietorship leasing equipment to the plan sponsor). The categorical logic of these cases therefore does not fit perfectly with a case like this one, where the lessor is an irrevocable family trust for the primary benefit of the settlor's sisters. However, the test deserves serious consideration because of the universal acceptance it has received in other Circuits.

The PBGC and the Trust engage in a dizzying battle of citations and counter-citations, each attempting to prove that these other Circuits have adopted its preferred test and rejected the other side's. The PBGC wins this battle. After recognizing the tension in the case law, the Seventh Circuit recently drew a clear line between two types of leasing behaviors by commonly owned lessors: (1) cases where the lessee was the plan sponsor (where the Categorical Test applies); and (2) cases where the lessee was a business enterprise unrelated to the plan sponsor (where Groetzinger applies). Messina Prods. , 706 F.3d at 880-82 . Thus, both tests are valid; they simply apply to different leasing arrangements. The U.S. District Court in Chicago also recently explained why the Categorical Test is desirable in the latter case. Regardless of intent, "[i]t is the fact that the economic relationship could be used to dissipate or fractionalize assets that makes leasing property to a withdrawing employer a 'trade or business.' " Central States v. Sidney Truck & Storage , 182 F.Supp.3d 855 , 860 (N.D. Ill. 2016).

This logic fits with our precedent. It has long been the law in this Circuit that Congress created the common-control provisions to prevent employers from avoiding liability "by operating through separate entities." Mason & Dixon , 852 F.2d at 159 . Any test we adopt must adhere to that goal. For those reasons, I agree that we must reverse the district court on this count, but under the more circumscribed language offered below.

3

On the facts here, the Trust's argument collapses. Although set up as an irrevocable family estate-planning device, several facts show that it was part of the defendants' business enterprise:

• The settlor (Phillip D. Gardner) received the property as a gift from the plan sponsor, which he controlled and operated;

• The settlor donated the property to the trust himself;

• The overwhelming majority of the trust's corpus was the two plots of land on which the plan sponsor operates;

• The trustees (the settlor's sons and residual beneficiaries) immediately leased the property back to the plan sponsor;

• The benefits to the settlor's sisters were only for life, and on their death, the property reverted to the settlor 's sons, instead of vesting in the sisters' heirs; and

• The residual beneficiaries assumed control of the plan sponsor after the settlor's death.

The Trust complains that "the property that Phillip D. Gardner donated to the Trust could never revert to him or to Findlay." Appellee Br. at 28. True, the property would not return to Gardner himself, but Gardner knew the property would return to his two sons, whom he almost certainly intended to run Findlay after he died. This sort of a leasing arrangement is exactly the kind of "economic relationship [that] could be used to dissipate or fractionalize assets." Sidney Truck & Storage , 182 F.Supp.3d at 860 . The common-control rules prevent this sort of maneuvering.

Although we should follow the other Circuits in adopting the Categorical Test in most cases (including this one), I would leave some issues open for future litigation. Commercial leases from a common owner to the plan sponsor should categorically be considered "trades or businesses" within § 1301(b)(1)'s reach. Trusts, however, create problems depending on (1) the revenue sources of the trust, (2) who controls the trust assets, and (3) who ultimately benefits from the trust. Here, all three factors militate in favor of ERISA liability. First, the revenue of the trust was derived almost exclusively from rent paid by the plan sponsor, for land on which it operated, and on land which it had previously owned. If common control is assumed, the lease was never truly a liability or an asset on the balance sheet of the common owners: It simply shifted assets from one commonly controlled entity to another. Second, the same people who owned and controlled the plan sponsor also controlled the distribution of trust assets and ultimately received the land on the expiration of the sisters' life estate. Imposing liability on these facts fits with both the goals of common-control liability articulated by Mason & Dixon and the flexible, practical analysis used in Challenge Stamping .

However, where any of these factors are not met, I would leave the issue open for further litigation. Imagine one tiny change to the facts of this case. Instead of returning the trust property to the sons, suppose that the trust instrument dictated that title vested in the sisters' heirs (or someone else) upon the expiration of the life estate. I find it difficult to believe that Congress intended to wrench assets away from an "innocent" third party just to satisfy a company's pension obligations. Again, this logic flows from Challenge Stamping , where we stated that "[t]here is no support for a view that Congress's chief intent in employing [common-control liability] was to invade the deepest pocket" no matter who it belongs to. 719 F.2d at 151 . In such cases-like in cases where creditors seek to levy on a business venture with no connection to the plan sponsor-a more flexible, pragmatic inquiry may be appropriate from the start. Cf. Messina Prods. , 706 F.3d at 880-82 .

The majority suggests that I am proposing a "less-than-categorical version" of the Categorical Test. Maj. Op. at 610-11. I have done no such thing. Commercial leases from a common owner to the plan sponsor are categorically covered by § 1301(b)(1). That holding is coextensive with the judgments of our sister Circuits. I write separately only to note that our decision today should not be understood to go beyond the facts it presents, and that future panels should be free to consider creating narrow exceptions if a family trust arrangement presents it with more challenging facts and a potentially inequitable result.

II

However, I disagree completely with the majority on the next issue. The majority creates federal common law to hold the successors liable for Findlay's pension obligations. The Successors argue that 29 U.S.C. § 1369 enumerates the only circumstances where the PBGC can impose Termination Liability on the successor to a plan sponsor. The Successors are right.

ERISA is one of the few areas where the federal courts are empowered to create federal common law. Firestone Tire & Rubber Co. v. Bruch , 489 U.S. 101 , 110, 109 S.Ct. 948 , 103 L.Ed.2d 80 (1989) ; DiGeronimo Aggregates, LLC v. Zemla , 763 F.3d 506 , 510-11 (6th Cir. 2014). However, we exercise that power with care and only create ERISA common law "in a few and restricted instances." DiGeronimo , 763 F.3d at 511 . If "Congress has established an extensive regulatory network and has expressly announced its intention to occupy the field, courts do not lightly create additional rights under the rubric of federal common law." Id. Further, the Supreme Court's instructions in Firestone were directed primarily at the administration of ERISA plans and only have secondary application to other parts of the statute. See Firestone , 489 U.S. at 110 , 109 S.Ct. 948 ; Erwin Chemerinsky, Federal Jurisdiction § 6.3.2, p. 415 (7th ed 2016).

We have restricted our common-law authority under ERISA to circumstances in which: (1) ERISA is silent or ambiguous, (2) ERISA leaves an awkward gap in the statutory scheme, or (3) federal common law is "essential" to promote "fundamental ERISA policies." DiGeronimo , 763 F.3d at 511 . Using these principles, we have created common law to address restitution claims, some estoppel claims, and undue influence claims. Whitworth Bros. Storage v. Central States , 794 F.2d 221 , 233-36 (6th Cir. 1986) (restitution allowed when employer paid too much into the fund); Bloemker v. Laborers' Local 265 Pension Fund , 605 F.3d 436 , 440 (6th Cir. 2010) (beneficiaries can bring equitable estoppel claim against fund based on reasonable reliance on written benefit statement); Sprague v. Gen. Motors Corp. , 133 F.3d 388 , 403-04 (6th Cir. 1998) (en banc) (beneficiaries may bring promissory estoppel claim against plan sponsor when sponsor promised to provide coverage for life); Tinsley v. Gen. Motors Corp. , 227 F.3d 700 , 704-05 (6th Cir. 2000) (creating common-law standard for undue influence exerted by beneficiaries of life insurance policy). Notice, again, that each of these cases addressed ERISA plans-i.e., disputes between the plan and its beneficiaries-not disputes between the plan and third parties.

In the few cases where parties have asked us to create common-law doctrines in the latter circumstance, we have declined to do so. United Food & Commercial Workers Union v. Rubber Assocs., Inc. , 812 F.3d 521 (6th Cir. 2016) (refusing to create equitable reduction in withdrawal liability imposed by an arbitrator); Central States v. Mahoning Nat'l Bank , 112 F.3d 252 (6th Cir. 1997) (rejecting plaintiffs' common-law withdrawal liability claim because they could have-but did not-pursue their claim in a timely manner under ERISA's withdrawal-liability statute). This background creates an even stronger presumption against creating federal common law in this case, and the PBGC's arguments do not overcome it. With this background in place, I turn to the three factors identified by DiGeronimo : (1) whether ERISA is silent on the issue, (2) whether the statute leaves an awkward gap, or (3)

whether common law is necessary to promote fundamental ERISA policies.

A

If ERISA speaks to an issue, a party cannot complain because it does not like what the statute says. Girl Scouts of Middle Tenn. v. Girl Scouts of the U.S.A. , 770 F.3d 414 , 420-21 (6th Cir. 2015). Relevant to this case, "[w]here ERISA allows for recovery on an issue under some but not all circumstances, ERISA is not silent on that issue." Id. at 421 . Thus, in Girl Scouts , the court noted that "ERISA is far from silent on the contractual claims [plaintiff] alleges. ERISA simply fails to afford [plaintiff] an avenue for recovery in this context." Id. Here, the Successors point out that § 1369 addresses certain "transactions to evade liability" and "corporate reorganization[s]," indicating a Congressional intent to limit successor liability. The PBGC contends that Congress did not intend this section to be the sole vehicle for successor liability under ERISA. The legislative history refutes the PBGC's conclusion.

The Congress that passed ERISA considered two versions of the language that eventually became § 1369(b). The first version, drafted by the Labor Committees of both chambers, imposed liability on the employer "or any successor in interest to such employer ..." S. 4, 93rd Cong. § 405(a) (1973), reprinted in ERISA Legis. Hist. 143, 538, 1240-41 (1976); H.R. 2, 93rd Cong. § 405(a) (1973), reprinted in ERISA Legis. Hist. 2326. Both Committees included this language because they were concerned about "acquiring companies ... [that] failed to take over the liability for vested benefits owed to the employees of the predecessor company." The committees therefore felt that it was necessary for "successors in interest to be liable for [obligations] owed by predecessor companies." S. Rep. No. 93-127, at 26 (1973), reprinted in ERISA Legis. Hist. 612; H.R. Rep. 93-533 (1973), reprinted in ERISA Legis. Hist. 2363.

The second version, proposed by the Senate Finance Committee, contained the language now codified in § 1369(b). S. 1179, 93rd Cong. § 462(e) (1973), reprinted in ERISA Legis. Hist. 933-34. After a long, drawn-out fight between the Labor and Finance Committees, both chambers agreed to the Finance Committee's language, and it remained in that form until final passage. 119 Cong. Rec. 1579 (1973), reprinted in ERISA Legis. Hist. 1590-91; H.B. 2, 93rd Cong. § 462(e) ("Successor Liability") (as passed by the Senate), reprinted in ERISA Legis. Hist. 3727-28; Conference Bill on H.B. 2, 93rd Cong. § 4062(d) (recodified without substantive change by Conference Committee, "Successor Liability" header deleted), reprinted in ERISA Legis. Hist. 4509-10; Employee Retirement Income Security Act of 1974, Pub. L. No. 93-406 § 4062(d), 88 Stat. 829 (93rd Cong., Sept. 2, 1974) (same as Conference Bill). It was not until 1986 that Congress added the title "Effect of Corporate Reorganization," and even then, it did so in an Omnibus budget bill with no explanation whatsoever. Consolidated Omnibus Budget Reconciliation Act of 1985, Pub. L. No. 99-272, § 11013(a), 100 Stat. 82 , 260-61 (99th Cong., April 7, 1986); H.R. Rep. 99-300, at 318.

This history convinces me that Congress intentionally restricted successor liability to those circumstances indicated in § 1369(b). Where Congress considered and rejected "the very language that would have achieved the result [a party] urges," that fact "weighs heavily against" the party's interpretation. Hamdan v. Rumsfeld , 548 U.S. 557 , 578-80, 126 S.Ct. 2749 , 165 L.Ed.2d 723 (2006). This fact disposes of both the silence inquiry and the "awkward gap" question under DiGeronimo . It is true that several Circuits have imposed common-law successor liability in cases dealing with multiemployer plans-in contrast to the single-employer plan here-and the lack of a uniform rule would be somewhat awkward. See Resilient Floor Covering Pension Tr. Fund v. Michael's Floor Covering, Inc. , 801 F.3d 1079 , 1093-95 (9th Cir. 2015) ; Tsareff v. ManWeb Servs., Inc. , 794 F.3d 841 , 844-47 (7th Cir. 2015). But Congress created this awkward situation; it should be the one to fix it. Girl Scouts , 770 F.3d at 420-21 . Thus, the PBGC's only viable remaining argument is that federal common law is essential to further ERISA's fundamental purposes.

B

Not every ERISA policy justifies creating common law. Similarly, the fact that a plaintiff's claim is based on a "fundamental" ERISA policy does not itself mandate the creation of common law. DiGeronimo , 763 F.3d at 511 . Only when the interest is "fundamental" and the creation of common law is "essential" to protect that interest should the courts exercise their lawmaking authority. Although this factor is independent of the silence and awkward-gap question, it does not ignore the facts discovered in those inquiries. See Local 6-0682 Int'l Union of Paper v. Nat'l Indus. Grp. Pension Plan , 342 F.3d 606 , 610 (6th Cir. 2003) ; Tassinare v. Am. Nat'l Ins. Co. , 32 F.3d 220 , 225 (6th Cir. 1994) ).

Here, we must ask whether the asserted policy is "fundamental." If it is, then we must examine whether the creation of federal common law -not merely the creation of a new federal remedy -is "essential" to accomplish that policy. Previous panels of this Court have suggested that ERISA established a fundamental policy of "ensuring that ... participants and beneficiaries obtain the benefits to which they are entitled." Tassinare , 32 F.3d at 225 (quoting Diduck v. Kaszycki & Sons Contractors, Inc. , 974 F.2d 270 , 281 (2d Cir. 1992) ). But the court has not expressly created federal common law on this ground, or anything close to it. The only other cases relying on the fundamental-ERISA-policy rationale asserted interests in protecting the integrity of written ERISA plans, Bloemker , 605 F.3d at 440-41 ; Sprague , 133 F.3d at 403-04 , or in fulfilling Congress's desire to fully preempt state law, Whitworth Bros. , 794 F.2d at 235-36 . Further, the interest in ensuring payment to beneficiaries is distinct from the PBGC's interest in recouping those payments, which is the real issue here. Admittedly, the PBGC's recoupment claims are an important cog in ERISA's enforcement mechanism. But even if this policy is fundamental, common law is not essential to protect it here.

Something is essential if it is "basic and necessary." Black's Law Dictionary, Essential (10th ed. 2014). In our federal system, common law is only necessary as a last resort-if no one else has done something about the problem, and if it's unlikely that anyone else will. Individual beneficiaries who find themselves left out in the cold by a gap in ERISA will usually not be able to muster the political clout to work clarifications in the law. Thus, at least when it comes to rules governing specific ERISA plans, the equitable powers of a court are crucial to filling these gaps. See Firestone , 489 U.S. at 110 , 109 S.Ct. 948 ; Bloemker , 605 F.3d at 440-41 .

The PBGC, however, is not so powerless to instigate legislative change. Indeed, the MPPAA was enacted, in part, because the PBGC notified Congress of a loophole in ERISA's original language. See 29 U.S.C. § 1001a ; PBGC v. R.A. Gray & Co. , 467 U.S. 717 , 722, 104 S.Ct. 2709 , 81 L.Ed.2d 601 (1984). The problem identified here is similar: someone exploited a loophole in § 1369's reorganization language. However, the legislative history indicates that the narrow scope of this section was at least somewhat intentional. It is possible, even likely, that the Senate Finance Committee was "reluctant to impose successor liability when it might inhibit the reorganization of failing businesses." Peters v. NLRB , 153 F.3d 289 , 301 (6th Cir. 1998). But it compromised-acknowledging the Labor Committee's concerns about evasive corporate transactions. Thus, if the problem at bar is truly so serious that it upsets the legislative balance set in 1974, the PBGC is certainly capable of convincing Congress to right the ship.

The majority is absolutely correct that, ultimately, the PBGC might not be able to persuade Congress to change the law, even though it has more clout than the ordinary citizen. Maj. Op. at 611 n.5. But that is precisely the point. Congress deliberately selected the narrow kind of successor liability we have before us, and it may decide (over the PBGC's objection) to choose that narrow road again. The outcome of this policy battle is not our concern-Congress's current intent is clear, and we are obligated to honor it. Miller v. French , 530 U.S. 327 , 336, 120 S.Ct. 2246 , 147 L.Ed.2d 326 (2000). The majority, however, veers around Congress's intent without even discussing it by declaring that the outcome advocated by the Successors "plainly would frustrate the purpose of Subchapter III." Maj. Op. at 611. I find it unlikely that Congress "shot itself in the foot," so to speak, by deliberately adopting language in such fundamental conflict with the purpose of the statute it was enacting. And I find it even more difficult to believe that we should find such "frustration" when the legislative history resolves any tension between the statute's text and its purpose.

Our power to create federal common law is the authority to fill gaps created or neglected by Congress. DiGeronimo , 763 F.3d at 511 . But it is emphatically not the authority to sit as a "superlegislature" to rewrite laws we think are unfair or to alter policy judgments we think are unwise. Cf. Exxon Corp. v. Maryland , 437 U.S. 117 , 124, 98 S.Ct. 2207 , 57 L.Ed.2d 91 (1978) ; Hodel v. Indiana , 452 U.S. 314 , 333, 101 S.Ct. 2376 , 69 L.Ed.2d 40 (1981). I would welcome a discussion of common-law successor liability if Congress had forgotten about it when passing ERISA. But Congress didn't forget about successor liability-it deliberately adopted a narrow form of the concept. I therefore cannot agree that expanding successor liability is essential to the promotion of fundamental ERISA policies when the policymaker has already considered and rejected that argument. I therefore respectfully dissent from the opinion of the court on this issue.