The Employee Retirement Income Security Act of 1974 ("ERISA") permits an employer seeking reorganization in Chapter 11 bankruptcy to terminate a pension plan if the employer satisfies certain notice requirements and demonstrates to a bankruptcy court that it will be unable to pay its debts and continue in business outside of Chapter 11 unless the pension plan is terminated. ERISA § 4041(c)(2)(B)(ii)(IV), 29 U.S.C. § 1341(c)(2)(B)(ii)(IV) (2000). Courts typically refer to this requirement for a plan termination as the "reorganization test." The instant case raises a question of first impression among the courts of appeals: when a Chapter 11 debtor seeks to terminate multiple pension plans simultaneously under the reorganization test, should a court apply the test to each plan independently, or to all of the plans in the aggregate?

Kaiser Aluminum Corporation and twenty-five of its affiliates ("Kaiser") are debtors in a Chapter 11 bankruptcy. As part of their reorganization, they requested that the Bankruptcy Court approve the termination of six pension plans under the reorganization test. The Bankruptcy Court applied the test to all six plans in the aggregate and concluded that their termination was required for Kaiser to emerge from Chapter 11. The Pension Benefit Guaranty Corporation ("PBGC"), which is responsible under ERISA to provide benefits to participants in terminated plans, appealed the Bankruptcy Court's decision, arguing that it should have applied the reorganization test on a plan-by-plan basis to each of Kaiser's pension plans. Under this approach, the PBGC contends that some of Kaiser's plans would not fulfill the reorganization test, and therefore could not be terminated. The District Court upheld the Bankruptcy Court's decision and the PBGC appealed to our Court.

We conclude that the Bankruptcy Court correctly applied the reorganization test in the aggregate to all of the plans Kaiser sought to terminate. Congress has not provided any guidance as to how to apply the reorganization test given the fact pattern before us, and the plan-by-plan approach appears unworkable. By contrast, applying the reorganization test to multiple plans in the aggregate is straightforward. A basic principle of statutory construction is that we should avoid a statutory interpretation that leads to absurd results. See Griffin v. Oceanic Contractors, Inc., 458 U.S. 564, 575, 102 S. Ct. 3245, 73 L. Ed. 2d 973 (1982). Because it would be anomalous for Congress to mandate an unworkable approach to the reorganization test, we read ERISA as requiring an aggregated analysis.

We are also persuaded that applying the reorganization test on a plan-by-plan basis would result in unfair and inequitable consequences in that it would require bankruptcy courts to give preference to some similarly situated constituents over others. The bankruptcy courts are courts of equity that are guided by equitable principles. Absent a clear congressional mandate to the contrary, we will not impose upon them an approach to the reorganization test that would conflict with their tradition of preventing unfairness in bankruptcy proceedings. Congress must speak more clearly than it has in ERISA if it wishes the bankruptcy courts to take a plan-by-plan approach to the reorganization test.

Finally, we consider, and reject, the PBGC's arguments based on legislative history, deference to its administrative interpretation, and public policy. We will therefore affirm the decision of the District Court upholding the Bankruptcy Court.

Kaiser is involved in all aspects of the aluminum industry, including mining raw materials, refining them, and manufacturing aluminum products. As of January 1, 2003, Kaiser employed approximately 3,000 workers domestically. In addition, it was responsible for the retiree benefits (primarily medical) of more than 15,300 retirees and dependent spouses and the pension benefits of over 11,000 retirees and beneficiaries. In late 2001 and early 2002, weak industry conditions, imminent debt maturities, burdensome asbestos litigation, and growing legacy obligations for future retiree medical and pension costs took its toll on Kaiser. Unable to restructure their obligations outside of bankruptcy, Kaiser and its related corporate entities filed for relief under Chapter 11 of the Bankruptcy Code between February 2002 and January 2003.

Congress established the PBGC in 1974 as part of ERISA. Its purpose is to encourage the continuation and maintenance of private-sector defined benefit pension plans, provide timely and uninterrupted payment of pension benefits, and keep pension insurance premiums at a minimum. 29 U.S.C. § 1302(a). To this end, the PBGC provides a minimum level of pension benefits to participants in qualified pension plans in the event that the plans cannot pay benefits. As of September 30, 2005, the PBGC insured 44.1 million American workers participating in 30,330 private-sector defined benefit pension plans. PBGC, PBGC Performance and Accountability Report for Fiscal Year 2005 1 (2005), available at http:// www.PBGC.gov/docs/2005par.pdf ("PBGC Performance Report").

The PBGC is not funded by general tax revenues. Rather, it collects insurance premiums from employers that sponsor insured pension plans, earns money from investments, and receives funds from pension plans it takes over. Id. at 1. The PBGC pays monthly retirement benefits, up to a guaranteed maximum, to about 683,000 retirees in 3,595 pension plans that have terminated.1 Including those who have not yet retired and participants in multiemployer plans receiving financial assistance, the PBGC is directly responsible for the current and future benefits of 1.3 million active and retired workers whose plans have failed. Id. at 2. The benefits guaranteed by the PBGC are often substantially lower than the fully vested pensions due to plan participants. See Mertens v. Hewitt Assocs., 508 U.S. 248, 250, 113 S. Ct. 2063, 124 L. Ed. 2d 161 (1993). Since 1987, a plan's sponsor is liable to the PBGC for the total amount of unfunded benefit liabilities to all participants and beneficiaries under the plan. 29 U.S.C. § 1362(b).

As part of its reorganization, Kaiser originally sought to replace seven pension plans that had been established in connection with collective bargaining agreements ("CBAs") then in effect with various unions.2 The plans and associated CBAs are listed in the table below:

------------------------------------------------------------
PENSION PLAN                 UNION(S) COVERED
------------------------------------------------------------
Kaiser Aluminum Pension      United Steelworkers of
Plan ("KAP Plan")            America ("USWA")
------------------------------------------------------------
Kaiser Aluminum Tulsa        USWA
Pension Plan ("Tulsa Plan")
------------------------------------------------------------
Kaiser Aluminum Bellwood     USWA; International Association
Pension Plan ("Bellwood      of Machinists &
Plan")                       Aerospace Workers
                             ("IAM")
------------------------------------------------------------
Kaiser Aluminum Sherman      IAM
Pension Plan ("Sherman
Plan")
------------------------------------------------------------
Kaiser Aluminum Inactive     USWA; IAM; United Automobile,
Pension Plan ("Inactive      Aerospace, and
Plan")                 Agricultural Implement
                             Workers of America
                             ("UAW")
------------------------------------------------------------
Kaiser Aluminum Los Angeles  International Brotherhood
Extrusion Pension            of Teamsters ("Teamsters")
Plan ("LA Extrusion Plan")
------------------------------------------------------------
Kaiser Center Garage         Teamsters Automotive
Pension Plan ("Garage        Employees, Local 78
Plan")
------------------------------------------------------------

These plans covered nearly 13,500 active hourly workers, participants on leave or layoff, individuals who were terminated from employment, retirees, and beneficiaries. On January 11, 2004, Kaiser filed a motion seeking the Bankruptcy Court's approval to terminate all seven plans in a voluntary "distress termination" under Title IV of ERISA, 29 U.S.C. § 1341(c)(2)(B)(ii). Kaiser asserted in its motion that it owed nearly $48 million in unfunded minimum contributions for the 2003 plan year and would be required to make $230 million in minimum contributions to the plans between 2004 and 2009. In separate motions, Kaiser also requested that the Bankruptcy Court (1) use its authority under 11 U.S.C. § 1113 to reject its CBAs with USWA and IAM, under which several of the pension plans had been established, and (2) authorize the modification of retiree benefits pursuant to 11 U.S.C. § 1114.

Prior to the February 2, 2004 hearing at which the Bankruptcy Court was to consider these motions, Kaiser reached agreements with USWA, IAM, and the official committee of salaried retirees ("1114 Committee") providing for consensual termination of the KAP Plan, Tulsa Plan, Bellwood Plan, Sherman Plan, and Inactive Plan, and for the institution of replacements for these plans. At the hearing, Kaiser asked the Bankruptcy Court to approve these agreements. The company had not yet reached agreements with UAW or the Teamsters to terminate the Inactive Plan and the LA Extrusion Plan, respectively.

At the hearing, Kaiser also withdrew its motion for approval to terminate the Garage Plan on the grounds that it was not underfunded. Thus, the Bankruptcy Court actually considered only whether to terminate six of Kaiser's seven active plans.

The PBGC opposed Kaiser's motion to terminate the six pension plans. In its briefs and at the February 2, 2004 hearing, the PBGC argued that the Bankruptcy Court should make separate determinations of whether each pension plan that Kaiser sought to terminate satisfied the reorganization test. Thus, rather than considering whether Kaiser could afford to fund all six plans in the aggregate, the PBGC urged the Bankruptcy Court to determine whether the contributions required for each individual plan, considered independently and without regard to the obligations under the other plans, jeopardized Kaiser's ability to reorganize successfully. The PBGC contended that the text and legislative history of ERISA required such a "plan-by-plan" approach.

The PBGC acknowledged at the hearing that two plans — the KAP Plan and the Inactive Plan — satisfied the reorganization test for distress termination even if considered under a plan-by-plan analysis. These plans were much larger than the others and would clearly impose an unsustainable burden on Kaiser. The PBGC contested only Kaiser's request to terminate the Tulsa Plan, Bellwood Plan, Sherman Plan, and LA Extrusion Plan. The combined minimum funding contributions for these four plans were projected to be roughly $12.8 million between 2004 and 2009, less than six percent of the estimated $230 million required to fund all of Kaiser's pension plans during that time frame. When these smaller plans were considered on a plan-by-plan basis, rather than in the aggregate with the KAP and Inactive Plans, the PBGC argued that Kaiser could continue funding some or all of them and still emerge successfully from Chapter 11 reorganization.

The Bankruptcy Court concluded that the PBGC's plan-by-plan approach would violate the Bankruptcy Code's requirement that debtors bargain fairly and equitably with unions. See 11 U.S.C. § 1113(b). The Bankruptcy Court believed that considering the plans piecemeal would give creditors "the kind of leverage that would force the debtor to [initiate] bargaining ... with one union and not with another." (Hr'g Tr., Feb. 2, 2004, at App. 445.) Likewise, debtors could use a plan-by-plan approach to gain leverage against creditors that Congress did not intend. The Court acknowledged that Kaiser could maintain up to three of its smaller plans under the PBGC's plan-by-plan approach, but held that, in order to be fair to all employees covered under the pensions, it would apply the reorganization test by considering the company's pension plans in the aggregate.

Under this standard, the Bankruptcy Court found that the reorganization test was satisfied with respect to all six plans that Kaiser sought to terminate. In a February 5, 2004 order, the Bankruptcy Court approved termination of the KAP Plan, the Sherman Plan, the Tulsa Plan, and the Bellwood Plan, effective upon the Court's filing of a contemporaneous order approving Kaiser's agreements with USWA, IAM, and the 1114 Committee for consensual termination of these plans. Though the Inactive Plan and the LA Extrusion Plan also satisfied the reorganization test, the Bankruptcy Court refused to approve their termination at that time because the CBAs that Kaiser had with UAW and the Teamsters presented a contractual bar to their termination. Kaiser has since reached agreements with both UAW and the Teamsters to remove the contractual bar in each of these CBAs, and the Bankruptcy Court has now formally approved the termination of the Inactive Plan and the LA Extrusion Plan.

The PBGC appealed the Bankruptcy Court's decision to the District Court, which upheld the Bankruptcy Court's aggregate analysis of the plans under the reorganization test. The District Court concluded that ERISA did not mandate a plan-by-plan analysis as urged by the PBGC. Furthermore, the Court believed that ERISA's reorganization test must be read in light of § 1113 of the Bankruptcy Code, which requires debtors to engage in fair and equitable bargaining with unions before unilaterally modifying CBAs. In the District Court's view, fairness and equity required the Bankruptcy Court to consider the plans in the aggregate. In re Kaiser Aluminum Corp., Civ.A. 04-145-JJF, 2005 WL 735551, at *3 (D.Del. Mar.30, 2005). The PBGC timely appealed the District Court's decision to our Court.

The Bankruptcy Court had jurisdiction under 28 U.S.C. § 157 and 28 U.S.C. § 1334(a). The District Court had jurisdiction over the PBGC's appeal under 28 U.S.C. § 158(a)(1). We have jurisdiction pursuant to 28 U.S.C. § 158(d).

We exercise plenary review of an order issued by a district court sitting as an appellate court in review of a bankruptcy court. In re CellNet Data Sys., Inc., 327 F.3d 242, 244 (3d Cir.2003). We will set aside the Bankruptcy Court's findings of fact if they are clearly erroneous and review its conclusions of law de novo. Id.

The question before us is whether the Bankruptcy Court should have made separate determinations as to whether each of the six plans Kaiser sought to terminate satisfied the reorganization test, or whether it properly applied the reorganization test to all six plans in the aggregate. "As in any case of statutory construction, our analysis begins with the language of the statute." Hughes Aircraft Co. v. Jacobson, 525 U.S. 432, 438, 119 S. Ct. 755, 142 L. Ed. 2d 881 (1999) (internal quotation omitted).

Title IV of ERISA establishes the exclusive means of terminating single-employer pension plans. 29 U.S.C. § 1341(a)(1); Hughes Aircraft, 525 U.S. at 446, 119 S. Ct. 755. Plans may be terminated voluntarily by a plan sponsor or involuntarily by the PBGC. A plan sponsor may voluntarily terminate a pension plan in one of two ways. First, it may proceed with a "standard termination" if it has sufficient assets to pay all benefit commitments. 29 U.S.C. § 1341(b)(1)(D). Such a situation does not implicate the PBGC's insurance responsibilities. Alternatively, if a plan's assets are not sufficient to satisfy all benefit liabilities, a plan sponsor may initiate a "distress termination" under 29 U.S.C. § 1341(c).4

A single-employer plan may terminate in a distress termination only if the plan administrator provides affected parties with at least sixty days of advance written notice of its intent to terminate, 29 U.S.C. § 1341(b)(2)(B); 29 C.F.R. § 4041.43, the administrator provides the PBGC with certain information about the termination no more than 120 days after the proposed termination date, 29 U.S.C. § 1341(b)(2)(A); 29 C.F.R. § 4041.45, and the PBGC determines that the plan sponsor meets one of four "distress tests" under 29 U.S.C. § 1341(c)(2)(B). The tests are known as (1) the liquidation test, (2) the reorganization test, (3) the inability to continue in business test, and (4) the unreasonably burdensome pension cost test. 29 U.S.C. §§ 1341(c)(2)(B)(i)-(iii); 29 C.F.R. § 4041.41(c)(4). In the instant case, Kaiser sought a distress termination of its plans under the reorganization test only.

Four requirements must be satisfied for a distress termination under the reorganization test. First, the plan sponsor must have filed a petition seeking reorganization in bankruptcy. Second, the bankruptcy case must not have been dismissed as of the proposed termination date. Third, the plan sponsor must submit to the PBGC a request for bankruptcy court approval of the plan termination. 29 U.S.C. §§ 1341(c)(2)(B)(ii)(I)-(III). Finally, the bankruptcy court must "determine[ ] that, unless the plan is terminated, [the plan sponsor] will be unable to pay all its debts pursuant to a plan of reorganization and will be unable to continue in business outside the chapter 11 reorganization process and approve[ ] the termination." Id. § 1341(c)(2)(B)(ii)(IV).

A plan sponsor may not voluntarily terminate a plan "if the termination would violate the terms and conditions of an existing collective bargaining agreement." 29 U.S.C. § 1341(a)(3).5 However, a plan sponsor seeking a distress termination while in bankruptcy may remove a contractual bar to a plan's termination by receiving the bankruptcy court's approval to unilaterally reject or modify the CBA under 11 U.S.C. § 1113. Section 1113 requires, among other things, that the debtor

make a proposal to the authorized representative of the employees covered by such agreement ... which provides for those necessary modifications in the employees benefits and protections that are necessary to permit the reorganization of the debtor and assures that all creditors, the debtor and all affected parties are treated fairly and equitably.

11 U.S.C. § 1113(b)(1)(A). A bankruptcy court must refuse to reject or modify a CBA if the debtor's proposal to the union was not fair and equitable. Wheeling-Pittsburgh Steel Corp. v. United Steelworkers, 791 F.2d 1074, 1093 (3d Cir.1986).

ERISA does not explicitly state how the reorganization test applies when an employer seeks to terminate several pension plans at once. The reorganization test is satisfied when a bankruptcy court determines that a plan sponsor will be unable to continue business outside of Chapter 11 "unless the plan is terminated." 29 U.S.C. § 1341(c)(2)(B)(ii)(IV) (emphasis added). The statute lacks any parallel provision for cases in which multiple plans are at issue. Other provisions of § 1341 likewise set forth requirements for the voluntary termination of "a single-employer plan," but do not specify how the requirements should apply in the context of multiplan terminations. See, e.g., id. § 1341(a)(1) (noting that ERISA provides the exclusive means under which "a single-employer plan may be terminated"); id. § 1341(a)(3) (barring voluntary termination of a "plan" that would violate the terms of an existing CBA); id. § 1341(b)(1) (listing the general requirements for a plan administrator to terminate a "single-employer plan" under a standard termination); id. § 1341(c)(2)(B) (requiring the PBGC to determine whether the sponsor of a "plan" satisfies one of the distress criteria).

Absent any express statutory instruction about how the reorganization test applies in the multiplan context, we must examine ERISA's text for indicia of congressional intent on the issue. See Lamie v. United States Trustee, 540 U.S. 526, 534, 124 S. Ct. 1023, 157 L. Ed. 2d 1024 (2004) ("The starting point for discerning congressional intent is the existing statutory text...."). The parties have not cited, and we have not found, any case in which a court has performed such an analysis. In every case that we have identified in which a debtor sought to terminate multiple pension plans under the reorganization test, bankruptcy courts have applied an aggregate analysis, apparently without protest from the PBGC. See In re Aloha Airgroup, Inc., No. 04-3063, 2005 WL 3487724, at *1 (Bankr.D.Haw. Dec. 13, 2005), vacated as moot, No. 05-00777, 2006 WL 695054, at * 3 (D.Haw. Mar.14, 2006); In re Philip Servs. Corp., 310 B.R. 802, 808 (Bankr.S.D.Tex.2004); In re Wire Rope Corp. of Am., 287 B.R. 771, 777-78 (Bankr.W.D.Mo.2002). However, these courts provided no rationale as to why they employed the aggregate approach and did not discuss whether they considered an alternate approach. Thus, these cases provide us with very little guidance or authority as to how to interpret ERISA's text.

The PBGC argues that Congress's use of the singular terms "single-employer plan" and "plan" mandates a plan-by-plan approach to terminations under § 1341 generally, and under the reorganization test in particular. In its view, ERISA defines the reorganization test in terms of a singular "plan" because Congress intended that bankruptcy courts would consider independently each plan that an employer seeks to terminate. Had Congress meant for courts to apply the aggregate approach to the reorganization test, the PBGC urges that § 1341(c)(2)(B)(ii)(IV) would instruct bankruptcy courts to determine whether an employer can continue in business outside Chapter 11 "unless the plans are terminated."

To support its textual interpretation, the PBGC notes that Congress chose to use the singular terms "single-employer plan" or "plan" throughout Title IV in a manner that it contends created a plan-specific statutory scheme to govern the single-employer plan termination insurance program. See, e.g., 29 U.S.C. § 1321 (detailing when a "plan" is covered by the termination insurance program); id. § 1322 (identifying the benefits guaranteed under a "single-employer plan"); id. § 1342 (granting the PBGC authority to involuntarily terminate a "single-employer plan" and establishing the criteria on which to evaluate a "plan"); id. § 1344(a) (establishing asset allocation scheme for a "single-employer plan" that is terminated); id. § 1347 (setting forth requirements for restoration of a terminated "plan"); id. § 1348(a) ("For purposes of this subchapter the termination date of a single-employer plan is ...." (emphasis added)); id. § 1362 (imposing liability on the sponsor of a "single-employer plan" that is terminated). Furthermore, the fact that Congress used the plural "plans" in certain Title IV provisions shows that it was cognizant of the different contextual uses of the singular "plan" and plural "plans." See id. § 1302(a) (stating that the purposes of Title IV are "to encourage the continuation and maintenance of voluntary private pension plans" and "to provide for the timely and uninterrupted payment of pension benefits ... under plans" (emphasis added)); id. § 1303(a) (mandating that the PBGC audit annually a "statistically significant number of plans" terminating in standard terminations); id. § 1310(a) (requiring plan sponsors to provide "information ... necessary to determine the liabilities and assets of plans" covered by ERISA). Given that ERISA "is a comprehensive and reticulated statute" that Congress drafted with care, Nachman Corp. v. Pension Benefit Guar. Corp., 446 U.S. 359, 361, 100 S. Ct. 1723, 64 L. Ed. 2d 354 (1980), the PBGC urges us to conclude that Congress intentionally defined the reorganization test in terms of a singular "plan" and that this choice reflects Congress's intent that the test should be applied on a plan-specific basis.

We disagree with the PBGC's textual analysis. Its "linguistic argument makes too much out of too little." United States v. Fior D'Italia, Inc., 536 U.S. 238, 244, 122 S. Ct. 2117, 153 L. Ed. 2d 280 (2002). The use of the singular form of "plan" in § 1341 does not constitute a congressional mandate to the bankruptcy courts to apply a plan-by-plan approach to the reorganization test. As a general matter of statutory construction, the singular form of a word "include[s] and appl[ies] to several persons, parties, or things" "unless the context indicates otherwise." 1 U.S.C. § 1. Here, nothing about the use or context of the singular terms "plan" or "single-employment plan" in ERISA suggests that "application of the typical rule of statutory construction set forth in 1 U.S.C. § 1 would be inappropriate." Toy Mfrs. of Am., Inc. v. Consumer Prods. Safety Comm'n, 630 F.2d 70, 74 (2d Cir.1980).

Furthermore, we do not think that Congress intended that its use of the singular "plan" would require a plan-by-plan approach to the reorganization test because, as the statute is currently written, such an approach is essentially unworkable. This is because the reorganization test cannot be rationally applied on a plan-by-plan basis unless a court makes basic assumptions about the order in which the plans should be considered and the status of the other plans that the employer is seeking to terminate. ERISA conspicuously fails to provide any ground rules whatsoever about how courts could employ a plan-by-plan analysis. If Congress had intended the bankruptcy courts to employ the reorganization test on a plan-by-plan basis, it would have done more than simply employ the singular form of "plan" in § 1341; it would also have provided some details about how courts are to apply such an approach. ERISA as it is currently drafted leaves open too many questions about how to engage in a plan-by-plan analysis for us to conclude that Congress envisioned such an approach in the multiplan context.

A brief hypothetical illustrates the problems inherent in applying the reorganization test on a plan-by-plan basis under the current statutory scheme. Assume that a debtor has three pension plans, each of which will cost $20 million annually. The evidence shows that the debtor can devote no more than $40 million annually to its pension liabilities and continue in business outside of Chapter 11. Under the PBGC's plan-by-plan approach, a bankruptcy court would approve the termination of just one plan because the debtor could afford to fund two of the three. But which plan should be terminated? Based simply on their cost, any of the plans could conceivably be eliminated; which one depends wholly on the mechanics of how the bankruptcy court applies the reorganization test. For example, the order in which the bankruptcy court examines the plans is potentially decisive (i.e., the first two considered will be deemed affordable, and the third one will not). Likewise, when examining one plan, the bankruptcy court must make a critical assumption about the status of the other two plans. A court would deem any one plan to be affordable if it assumed that one or both of the other plans had been terminated, and it would conclude that any plan is unaffordable if it assumed that the other two plans still existed. Thus, the outcome of a plan-by-plan analysis changes dramatically based on the ground rules that a court employs.6

These are fundamental problems with applying the reorganization test on a plan-by-plan basis, yet Congress did nothing to address, let alone resolve, them. Nor are these problems merely theoretical. In the instant case, the Bankruptcy Court noted that, if the KAP and Inactive Plans were terminated first, Kaiser could afford to fund as many as three of the four smaller plans (i.e. the Tulsa, Sherman, LA Extrusion, and Bellwood Plans).7 However, nothing in ERISA suggests that the KAP and Inactive Plans should be terminated first under a plan-by-plan analysis. A court could also conclude that it should apply the reorganization test to the smallest plans first, rather than in the manner that would require it to terminate the fewest number of plans. If a court were to apply the reorganization test to Kaiser's plans in order of smallest to largest, it would determine that Kaiser could not afford even the smaller plans due to the cost of the KAP and Inactive Plans.

Similarly, if the Bankruptcy Court assumed while applying the reorganization test to one plan that all the other plans remained active, it would conclude that the reorganization test was satisfied as to any of the plans considered independently. The cost of the KAP and Inactive Plans would make any other plan prohibitively expensive. On the other hand, if one assumed that the KAP and Inactive Plans would be eliminated, several of the smaller plans could be funded outside of Chapter 11. We see no principled textual basis in ERISA to adopt one set of assumptions over another.

A basic tenet of statutory construction is that courts should interpret a law to avoid absurd or bizarre results. See Demarest v. Manspeaker, 498 U.S. 184, 191, 111 S. Ct. 599, 112 L. Ed. 2d 608 (1991) (applying statute's terms where the result was not "so bizarre that Congress could not have intended it" (internal quotation omitted)); Griffin, 458 U.S. at 575, 102 S. Ct. 3245 ("It is true that interpretations of a statute which would produce absurd results are to be avoided if alternative interpretations consistent with the legislative purpose are available."). If we adopted the PBGC's interpretation of § 1341, we would be concluding that Congress required courts to apply the reorganization test on a plan-by-plan basis, but provided no guidance on the mechanics of this approach, making it essentially unworkable. We will not adopt a statutory construction that leads to such an anomalous result, especially where the aggregate approach represents an alternative that is "neither irrational nor arbitrary." DiGiacomo v. Teamsters Pension Trust Fund of Philadelphia and Vicinity, 420 F.3d 220, 228 (3d Cir.2005).

Nor would we make the plan-by-plan approach workable by "filling in the gaps" left by Congress ourselves. See Griggs v. E.I. DuPont de Nemours & Co., 385 F.3d 440, 453 n. 7 (4th Cir.2004). In the context of such an "enormously complex and detailed statute," Mertens, 508 U.S. at 262, 113 S. Ct. 2063, whose text we should be cautious about supplementing, Hughes Aircraft, 525 U.S. at 447, 119 S. Ct. 755, it is not appropriate for us to invent for the bankruptcy courts the fundamental baseline assumptions required to apply the reorganization test workably on a plan-by-plan basis. To do so would require us to weigh sensitive policy issues that have potentially important consequences for employers, American workers, and the PBGC without any meaningful guidance from Congress. "The authority of courts to develop a `federal common law' under ERISA is not the authority to revise the text of the statute." Mertens, 508 U.S. at 259, 113 S. Ct. 2063 (internal citation omitted). We may develop federal common law under ERISA only when it is "`necessary to fill in interstitially or otherwise effectuate the statutory pattern enacted in the large by Congress.'" Fotta v. Trs. of the UMW, Health & Ret. Fund of 1974, 165 F.3d 209, 211-212 (3d Cir.1998) (quoting Bollman Hat Co. v. Root, 112 F.3d 113, 118 (3d Cir.1997)); see also Bill Gray Enters., Inc. Employee Health and Welfare Plan v. Gourley, 248 F.3d 206, 220 n. 13 (3d Cir.2001) ("[C]ourts have held that importing federal common law doctrines to ERISA plan interpretation is generally inappropriate...."). The significant gap-fillers required to apply a plan-by-plan approach workably are better developed by a policy-making or legislative body than by this Court. See DiGiacomo, 420 F.3d at 228.

The PBGC contends that "[i]t is difficult to imagine how Congress could have spoken to the `precise question' of whether the distress termination requirements ... apply to a sponsor on a plan-by-plan basis more clearly than it did." (PBGC Br. at 18.) To the contrary, we have no trouble envisioning how it could have done so. It could have used the phrase "plan-by-plan" in articulating the reorganization test, explicitly instructed courts to apply the test to each plan independently, or given even a modicum of guidance about how such an approach should be applied and what assumptions should be used in the multiplan context. Absent any such textual indicators of congressional intent, we will not read them into the statute ourselves. Instead, we adopt a construction of ERISA's reorganization test that does not lead to these problems, namely, the aggregate approach.

We find additional support for our textual analysis in the fact that a plan-specific approach to the reorganization test would disrupt the bankruptcy courts in their traditional role as agents of equity. The PBGC would have the Bankruptcy Court terminate some of Kaiser's plans while leaving the others in place, seemingly without a principled basis on which it could make the determination of which workers to prefer over others. We will not impose this result, which we believe would treat Kaiser's workers unfairly and inequitably, without a clear congressional mandate.

The Supreme Court has long recognized that bankruptcy courts are courts of equity that apply equitable principles in the administration of bankruptcy proceedings. See Local Loan Co. v. Hunt, 292 U.S. 234, 240, 54 S. Ct. 695, 78 L. Ed. 1230 (1934) ("[C]ourts of bankruptcy are essentially courts of equity, and their proceedings inherently proceedings in equity."). Though the enactment of the Bankruptcy Code in 1978 "increased the degree of regulation Congress imposed upon bankruptcy proceedings," it did not alter their "fundamental nature" as courts of equity. Official Comm. of Unsecured Creditors of Cybergenics Corp. ex rel. Cybergenics Corp. v. Chinery, 330 F.3d 548, 567 (3d Cir.2003) (en banc); see also Young v. United States, 535 U.S. 43, 50, 122 S. Ct. 1036, 152 L. Ed. 2d 79 (2002) ("[B]ankruptcy courts ... are courts of equity and `apply the principles and rules of equity jurisprudence.'" (quoting Pepper v. Litton, 308 U.S. 295, 304, 60 S. Ct. 238, 84 L. Ed. 281 (1939))).

The "`great principles of equity'" are aimed at "`securing complete justice'" for the parties before a court. Porter v. Warner Holding Co., 328 U.S. 395, 398, 66 S. Ct. 1086, 90 L. Ed. 1332 (1946) (quoting Brown v. Swann, 35 U.S. (10 Pet.) 497, 503, 9 L. Ed. 508 (1836)). Thus, the bankruptcy courts have broad authority to act in a manner that will prevent injustice or unfairness in the administration of bankruptcy estates. See Pepper, 308 U.S. at 307-08, 60 S. Ct. 238 ("[I]n the exercise of its equitable jurisdiction the bankruptcy court has the power to sift the circumstances surrounding any claim to see that injustice or unfairness is not done ...."); In re Combustion Eng'g, Inc., 391 F.3d 190, 235 (3d Cir.2004) ("Bankruptcy courts are courts of equity, empowered to invoke equitable principles to achieve fairness and justice in the reorganization process." (internal quotation omitted)). To this end, they may, when necessary, "eschew[ ] mechanical rules," Holmberg v. Armbrecht, 327 U.S. 392, 396, 66 S. Ct. 582, 90 L. Ed. 743 (1946), "modify creditor-debtor relationships," United States v. Energy Res. Co., 495 U.S. 545, 549, 110 S. Ct. 2139, 109 L. Ed. 2d 580 (1990), and "craft flexible remedies that, while not expressly authorized by the [Bankruptcy] Code, effect the result the Code was designed to obtain," Cybergenics Corp., 330 F.3d at 568. See also 11 U.S.C. § 105(a) (authorizing bankruptcy courts to "tak[e] any action or mak[e] any determination necessary or appropriate to enforce or implement court orders or rules, or to prevent an abuse of process."); but see In re Combustion Eng'g, 391 F.3d at 236 ("The general grant of equitable power contained in § 105(a) cannot trump specific provisions of the Bankruptcy Code, and must be exercised within the parameters of the Code itself.").

Section 1113 of the Bankruptcy Code illustrates the role that the bankruptcy courts take in ensuring fairness during the course of bankruptcy proceedings. Under this provision, a bankruptcy court may permit a debtor to unilaterally reject or modify an existing collective bargaining agreement only if the court determines, inter alia, that the debtor has made a proposal to the union that "assures that all creditors, the debtor and all affected parties are treated fairly and equitably." 11 U.S.C. § 1113(b)(1)(A) (emphasis added). Likewise, § 1114, which prohibits a debtor from unilaterally terminating or modifying retiree benefits without a court order, requires the debtor to make a proposal for the modification of benefits to the authorized representatives of the debtor's retirees that treats all affected parties "fairly and equitably." Id. § 1114(f)(1)(A). These sections also require the bankruptcy court to determine that "the balance of the equities clearly favors" the motions before granting them. Id. § 1113(c)(3); id. § 1114(g)(3). The provisions underscore the importance of equitable principles for bankruptcy courts, particularly in bankruptcies involving unionized workers and employee retirement benefits.8

We will not require the equitable principles under which bankruptcy courts operate to be discarded when courts are deciding whether to approve a pension plan termination under the reorganization test. See In re U.S. Airways Group, Inc., 296 B.R. 734, 746 (Bankr.E.D.Va.2003) (holding that the requirement under 29 U.S.C. § 1341(c)(2)(B)(ii)(IV) that the bankruptcy court "approve the termination" of an ERISA plan authorized the court to consider the "equities in the case"). We are convinced that, in light of the unique facts of this case, the Bankruptcy Court could not have applied the plan-by-plan approach without by producing an unfair result that would have violated principles of equity.

Had the Bankruptcy Court applied the reorganization test on a plan-by-plan basis it would have had to pick and choose between the six plans that Kaiser sought to terminate, deciding that certain plans would remain active. Some of Kaiser's workers would receive their full pension benefits, while others would receive no more than the amount guaranteed under ERISA. It is likely that even workers within the same union would be treated differently from each other because they participated in different plans, not all of which would necessarily be eliminated. Without some statutory basis or other principled rationale for this result, such disparate treatment smacks of arbitrariness. Under § 1113, "the focus of the [bankruptcy court's] inquiry as to `fair and equitable' treatment should be whether the [debtor's] proposal would impose a disproportionate burden on the employees," Wheeling-Pittsburgh, 791 F.2d at 1091, as compared to "creditors, debtor and all of the affected parties," 11 U.S.C. § 1113(b)(1)(A). Here, the PBGC asked the Bankruptcy Court to burden certain employees disproportionally as compared to other employees. This strikes us as an unfair result, and it is one that a bankruptcy court sitting in equity should not impose absent a clear mandate from Congress.

When pressed at oral argument, the PBGC offered no basis on which the Bankruptcy Court could make the difficult decision of which of Kaiser's four small plans should be terminated under a plan-by-plan approach. It stated only that a debtor can make this choice in whatever manner it wishes. But this does little more than restate the problem before us. Here, Kaiser has voluntarily chosen to terminate all six pension plans. Thus, the "solution" the PBGC suggests does not apply in this case.

It did apply in the case of US Airways Group. There, U.S. Airways sought to terminate only one of its four pension plans. The affected workers were the company's pilots, who had "already given up more in pay and benefits than any other employee group." US Airways Group, 296 B.R. at 744. Not surprisingly, the pilots felt "a particularly keen sense of having been shabbily treated" by the request to terminate their pension plan. Id. They argued that it would be "highly unfair" to terminate the pilots' plan "while all the remaining employee groups would keep their current pension benefits." Id. Nevertheless, the bankruptcy court approved the termination under the reorganization test, contingent on the removal of the contractual bar in the pilots' CBA. Id. at 745-46.

Unlike the case before us, the court in US Airways Group was asked to apply the reorganization test to just one pension plan. Under those circumstances, any unfairness inherent in the termination was the result of the debtor's business decision about which plan to terminate, not the bankruptcy court's roll of the dice as to which plan should or should not be terminated.9

The PBGC contends that the result of applying an aggregate approach is just as unfair as terminating some plans and leaving others in place. Under the aggregated analysis, more plans would be terminated, and more workers impacted, than what is required for Kaiser to emerge from Chapter 11. The Bankruptcy Court itself acknowledged that it was not necessary for Kaiser to terminate all of its plans to reorganize, yet it terminated them all anyway. Surely, the PBGC argues, it is not fair or equitable to strip an employee's pension benefits without any economic justification. Faced with a choice of burdening some of the participants in Kaiser's plans and burdening them all, the PBGC contends that equity weighs in favor of the former.10

We are not unsympathetic to this view. There is undoubtedly a tension between treating similarly situated workers alike and doing the least that is necessary for the company to emerge from bankruptcy. However, we are persuaded that, on the whole, an aggregate approach is more in line with the objectives of the Bankruptcy Code.

As a practical matter, voluntary terminations under the reorganization test always require an employer to bargain with the representatives of its employees or retirees. See 11 U.S.C. § 1113(c)(2); id. § 1114(g)(2). When an employer seeks to terminate multiple plans, participants in one plan will be less likely to agree to a termination if doing so would open the door to a decision by a bankruptcy court to single out their plan for termination under the plan-by-plan approach, while leaving the employer's other plans intact. See U.S. Airways Group, 296 B.R. at 744 (describing union's objection to being singled out for termination of its pension benefits). And debtors will generally be unable to select one plan among many for unilateral termination without the consent of its participants because doing so would violate the fair and equitable requirements of § 1113 and § 1114. Cf. Wheeling-Pittsburgh, 791 F.2d at 1091 (defining inquiry into "fair and equitable" in terms of a "disproportionate burden" on employees).

The consequence is that employers that could conceivably restructure their pension liabilities and successfully reorganize will have a harder time doing so under a plan-by-plan approach. This would, in turn, lead to a higher number of liquidations and, by extension, a higher number of overall plan terminations. The result would be to leave all interested parties — the PBGC, workers, retirees, and creditors — worse off as compared to the same number of reorganizations. An approach that results in unnecessary liquidations is neither fair nor consistent the Bankruptcy Code's preference for reorganization. See Nordhoff Invs., Inc. v. Zenith Elecs. Corp., 258 F.3d 180, 190 (3d Cir.2001) (noting the "strong public policy in favor of maximizing debtor's estates and facilitating successful reorganization"); In re Baker & Drake, Inc., 35 F.3d 1348, 1354 (9th Cir.1994) ("Congress's purpose in enacting the Bankruptcy Code was not to mandate that every company be reorganized at all costs, but rather to establish a preference for reorganizations, where they are legally feasible and economically practical.").

The PBGC has leveled several other arguments against our reading of ERISA. It contends that the legislative history, its administrative interpretation, and public policy support its view that ERISA mandates a plan-by-plan approach. As with the PBGC's statutory analysis, we do not find its arguments on these points to be persuasive.

The PBGC points to a legislative trend to tighten the restrictions on pension plan terminations as support for a plan-by-plan approach to the reorganization test. In 1986, Congress established the four distress tests in the Single-Employer Pension Plan Amendments Act ("SEPPAA"). Prior to the enactment of SEPPAA, "a plan could be terminated voluntarily at any time, regardless of the relationship between its assets and liabilities." E. Thomas Veal & Edward R. Mackiewicz, Pension Plan Terminations 68 (2d ed.1998). If the plan could not provide the guaranteed benefits to pensioners, the PBGC would pay the benefits and assess liability against the plan's sponsor. Id. at 68-69. Congress found that this system "encourage[d] employers to terminate plans, evade their obligations to pay benefits, and shift unfunded pension liabilities" to the PBGC. H.R.Rep. No. 99-241, part 2, at 59 (1985), reprinted in 1986 U.S.C.C.A.N. 685, 717-18. SEPPAA was intended to heighten the requirements for plan terminations.

In 1987, Congress passed the Pension Protection Act of 1987 ("PPA") to restrict pension plan terminations further. PPA made employers liable, for the first time, to the PBGC for the full amount of unfunded benefit liabilities to all participants and beneficiaries under the plan. See 29 U.S.C. § 1362. It also applied the reorganization test to "[t]he plan sponsor or any member of its controlled group," whereas SEPPAA applied it only to each "substantial member" of the control group. See 29 U.S.C. § 1341(c)(2)(B); Veal & Mackiewiz, supra, at 252 n. 3. Both changes increased the burden on employers seeking to terminate pension plans.

The PBGC contends that SEPPAA and PPA, taken together, reflect a clear congressional purpose to limit the circumstances under which pension plans may be voluntarily terminated to instances where sponsors are suffering "severe hardship." H.R.Rep. No. 99-300, at 278 (1985), reprinted in 1986 U.S.C.C.A.N. 929; H.R.Rep. No. 99-241, part 2, at 59-60 (1985), reprinted in 1986 U.S.C.C.A.N. 717-18. In addition, Congress intended to reduce the financial burdens on the PBGC and increase the chance that a plan's participants will receive their full expected benefits. In the context of the reorganization test, the PBGC argues that these objectives can only be achieved if bankruptcy courts employ a plan-by-plan approach, which prevents terminations that are economically unnecessary.

Our reading of the legislative history does not convince us that Congress mandated a plan-by-plan analysis. At most, the legislative history demonstrates that Congress had a general intent to make it more difficult for employers to terminate pensions; however, that is hardly determinative of whether, or how, the reorganization test should be applied in the multiplan context. As discussed above, we think it likely that a plan-by-plan analysis would actually increase the overall number of terminations and therefore conflict with the legislative intent on which the PBGC relies. In any event, such a general legislative purpose provides no guidance on the mechanics of applying the reorganization test workably on a plan-by-plan basis. We view the absence of any such guidance in the text of ERISA as more indicative of congressional intent than anything the PBGC has cited in the legislative record.

The PBGC claims that we should defer to its interpretation of the reorganization test under Chevron U.S.A. Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 104 S. Ct. 2778, 81 L. Ed. 2d 694 (1984). We disagree.

Congress has delegated to the PBGC the power to adopt rules and regulations that are necessary to carry out the purposes of Title IV of ERISA, see 29 U.S.C. § 1302(b)(3), and the Supreme Court has accorded Chevron deference to the PBGC's interpretation of ERISA on other occasions. See PBGC v. LTV Corp., 496 U.S. 633, 652, 110 S. Ct. 2668, 110 L. Ed. 2d 579 (1990) (deferring to the PBGC's anti-follow-on policy); Mead Corp. v. Tilley, 490 U.S. 714, 725, 109 S. Ct. 2156, 104 L. Ed. 2d 796 (1989) (according Chevron deference to the PBGC's interpretation of 29 U.S.C. § 1344). However, deference to the PBGC here is improper because the PBGC has neither the expertise nor the authority to determine when a plan should be terminated under the reorganization test. Issues relating to an employer's bankruptcy and reorganization are within the expertise of bankruptcy courts, not the PBGC. Thus, ERISA grants the bankruptcy courts alone the power to determine when an employer "will be unable to pay all its debts pursuant to a plan of reorganization and ... continue in business outside of chapter 11." 29 U.S.C. § 1341(c)(2)(B)(ii)(IV). Where Congress delegates to the courts, rather than administrative agencies, the power to make determinations under a statute, Chevron deference does not apply. See United States v. Mead Corp., 533 U.S. 218, 226-27, 121 S. Ct. 2164, 150 L. Ed. 2d 292 (2001) (holding that administrative implementation of a particular statutory provision qualifies for Chevron deference when Congress delegated authority to agency "to make rules carrying the force of law"); Murphy Exploration and Prod. Co. v. U.S. Dep't of Interior, 252 F.3d 473, 478 (D.C.Cir.2001) (stating that "Chevron does not apply to statutes that ... confer jurisdiction on the federal courts" because "such statutes do not grant power to agencies"); Bamidele v. INS, 99 F.3d 557, 561 (3d Cir.1996) (refusing to defer to agency's interpretation of a legal issue that is the province of the courts).

Furthermore, even if we were to hold that the PBGC had the authority to interpret § 1341(c)(2)(B)(ii)(IV), "[w]e will not defer to `an agency counsel's interpretation of a statute where the agency itself has articulated no position on the question.'" Connecticut General Life Ins. Co. v. Comm'r, 177 F.3d 136, 143-144 (3d Cir. 1999) (quoting Bowen v. Georgetown Univ. Hosp., 488 U.S. 204, 212, 109 S. Ct. 468, 102 L. Ed. 2d 493 (1988)); see also Southco, Inc. v. Kanebridge Corp., 390 F.3d 276, 300 (3d Cir.2004) (en banc) (Roth, J., dissenting) (noting that agency position advanced for the first time in litigation would "be entitled to no deference whatsoever"). To merit deference, an agency's interpretation of the statute must be supported by regulations, rulings, or administrative practice. Bowen, 488 U.S. at 212, 109 S. Ct. 468. It appears to us that the PBGC first adopted the view that ERISA requires a plan-by-plan analysis during the course of this litigation.11 Chevron deference is not appropriate under these circumstances. Id. at 213, 109 S. Ct. 468.

The PBGC contends that it has promulgated a series of rules that require that the voluntary termination of pensions occur on a plan-by-plan basis. See 29 C.F.R. § 4041.41; Distress Terminations and Standard Terminations of Single-Employer Plans, 57 Fed.Reg. 59206 (1992); Single-Employer Plan Terminations Under the Pension Protection Act, 53 Fed.Reg. 1904 (1988); Special Procedures Relating to the Reorganization Distress Test, 52 Fed.Reg. 38290 (1987); Distress Terminations of Single-Employer Plans and Standard Terminations of Single-Employer Plans, 52 Fed.Reg. 33318 (proposed Sept. 2, 1987); The Effects of the Single-Employer Pension Plan Amendments Act of 1986 on Voluntary Plan Terminations Initiated On or After January 1, 1986 and Before April 7, 1986, 51 Fed.Reg. 12489 (1986). But these rules, like ERISA's text, merely refer to the termination requirements for a "plan." They neither state that bankruptcy courts should apply the reorganization test on a plan-by-plan basis nor provide any guidance as to how courts would employ such an approach in the context of multiplan terminations. In short, contrary to the PBGC's urging, the rules do not reflect an administrative determination that ERISA requires a plan-by-plan analysis.12 Because the PBGC cannot point to regulations, rulings, or administrative practices in which it adopted the position that it asserts here, both the deliberateness and authoritativeness of its statutory interpretation are suspect. Smiley v. Citibank (South Dakota), N.A., 517 U.S. 735, 741, 116 S. Ct. 1730, 135 L. Ed. 2d 25 (1996).

Policy issues have provided the subtext for several of the PBGC's arguments in favor of a plan-by-plan approach and, at times, the PBGC has brought these issues to the surface explicitly. There is no question that this case implicates significant policy concerns that potentially affect millions of American workers and hundreds of businesses. Furthermore, the PBGC is an important government entity whose interests are not lightly ignored. We therefore address its policy arguments briefly and comment on their role in our analysis.

The policy concerns in this case have formed two parallel tracks. First, the PBGC has contended repeatedly that an aggregate approach to the reorganization test harms American workers who participate in ERISA plans because it subjects them to plan terminations that are economically unnecessary. Its fear is that our holding today will make it too easy to terminate pension plans that are actually affordable and will create an incentive for employers to terminate plans that they would otherwise maintain. As a result, more workers will lose their fully vested pensions and receive only the benefits guaranteed by ERISA.

The second concern is that our holding will negatively impact the PBGC itself. We take judicial notice of the fact that the PBGC's financial health has deteriorated sharply in recent years. At the end of the 2005 fiscal year, the PBGC's liabilities exceeded its assets by $23.1 billion, a swing in its net position of nearly $33 billion since 2000. PBGC Performance Report at 4; Congressional Budget Office, A Guide to Understanding the Pension Benefit Guaranty Corporation 1 (2005), available at http:// www.cbo.gov/ftpdocs/66xx/doc6657/09-23-GuideToP BGC.pdf. The PBGC noted in its brief that "[t]he issue presented in this case recurs in other large bankruptcy cases in which the agency is a guarantor of pension benefits." Interpreting ERISA in a way that triggers more plan terminations, and thereby increases the burdens on the PBGC, could undermine the PBGC's already shaky financial position and "pose a considerable risk to the single-employer termination insurance program." (PBGC Br. at 3.)

We do not downplay the significance of either argument. They provide sound policy rationales for the result for which the PBGC advocates and highlight the important interests at stake in this case. Moreover, there is no question that the aggregate approach may, in some cases, lead to results that are less than ideal for workers and for the PBGC. Nevertheless, "[w]e do not sit here as a policy-making or legislative body." DiGiacomo, 420 F.3d at 228. There are no clear answers to the difficult policy issues involved in this case and, in any event, their resolution is better left to Congress than the courts. We have taken Congress's failure to provide a shred of guidance on how to apply a plan-by-plan approach as indicative of its intent. If Congress perceives our holding to be in error, the cure is to amend ERISA. See Griffin, 458 U.S. at 576, 102 S. Ct. 3245.

For the reasons stated above, we conclude that when an employer in Chapter 11 bankruptcy seeks to terminate multiple pension plans voluntarily under the reorganization test, Congress intended the bankruptcy courts to apply the test to all of the plans in the aggregate. Consequently, we will affirm the order of the District Court upholding the Bankruptcy Court's conclusion that Kaiser had satisfied the reorganization test with respect to all six plans that it sought to terminate.