The primary issue presented for our review in this consolidated appeal concerns the attempted assertion by a multiemployer pension plan against an individual employer withdrawing from the plan of withdrawal liability as computed under the "modified presumptive method," pursuant to the Multiemployer Pension Plan Amendments Act, 29 U.S.C. Sec. 1381 et seq. Specifically, we must determine whether the fact that a plan has no "unfunded vested benefits" as of the year preceding an employer's decision to withdraw from the plan immunizes the withdrawing employer from such liability.

* A

The Employers-International Longshoremen's Association Pension Welfare and Vacation Fund for the North Carolina Ports Area ("the Fund") is a labor-management fund authorized to provide employee benefits pursuant to the Labor-Management Relations Act, 29 U.S.C. Sec. 151 et seq., and the Employee Retirement Income Security Act, 29 U.S.C. Sec. 1001 et seq. The Fund was formed through an agreement between several North Carolina locals of the International Longshoremen's Association ("ILA"), members of the North Carolina Longshoremen's Employers Association, and several stevedoring enterprises doing business in North Carolina. The trustees of the Fund have adopted a pension plan for the purpose of providing retirement and death benefits to eligible employees. The plan is financed by contributions the Fund receives from employers of longshoremen who are members of ILA locals, pursuant to collective bargaining agreements between the employers and the locals. Until August 1, 1987, when they withdrew from the Fund, two such employers were Almont Shipping Company, Incorporated ("Almont") and Stevedores, Incorporated ("Stevedores").1

Several of the Fund's trustees are also trustees of the Employers-International Longshoremen's Association, AFL-CIO, Pension Fund for the North Carolina Area ("the North Carolina Fund"). Like the Fund, the North Carolina Fund provides benefits for North Carolina longshoremen. Under a somewhat more complex arrangement, the North Carolina Fund used to receive contributions from the South Atlantic International Longshoremen's Association/Employers District Escrow Fund ("the Escrow Fund") and the South Atlantic International Longshoremen's Association/Employers Guaranteed Annual Income Fund ("the GAI Fund"). The Escrow Fund and the GAI Fund were, in turn, financed by contributions from waterfront employers. For tax purposes, the funding arrangement was amended on September 30, 1985. As a result of the amendment, contributions that previously were channeled through the Escrow Fund and the GAI Fund are now channeled through local port escrow funds. Since the fiscal year ending September 30, 1985, neither the Escrow Fund nor the GAI Fund has contributed to the North Carolina Fund.

To understand the dispute that has arisen among the parties it is necessary to pause to examine the general framework of the Multiemployer Pension Plan Amendments Act, 29 U.S.C. Sec. 1381 et seq. ("MPPAA" or "Act"). The Act, which took effect in 1980, was intended "to protect the financial base of pension plans from the erosion that occurred when a participating employer withdrew from a multiemployer pension plan that contained unfunded vested benefits." Masters, Mates & Pilots Pension Plan v. USX Corp., 900 F.2d 727, 730 (4th Cir.1990); see generally Pension Benefit Guaranty Corp. v. R.A. Gray & Co., 467 U.S. 717, 720-25, 104 S. Ct. 2709, 2713-16, 81 L. Ed. 2d 601 (1984) (discussing the Act's genesis). Unfunded vested benefits, or "UVBs," are defined as "the amount by which the value of future benefits vested (nonforfeitable) in covered employees exceeds the value of a plan's assets." Masters, Mates & Pilots, 900 F.2d at 730; see 29 U.S.C. Sec. 1393(c). MPPAA imposes "withdrawal liability" upon withdrawing employers "so that an employer who withdraws from a pension plan pays its proportionate share of the plan's UVBs." Masters, Mates & Pilots, 900 F.2d at 730. Specifically, MPPAA provides:

If an employer withdraws from a multiemployer plan in a complete or partial withdrawal, then the employer is liable to the plan in the amount determined under this part to be the withdrawal liability.

29 U.S.C. Sec. 1381(a). The Act then provides:

The withdrawal liability of an employer to a plan is the amount determined under section 1391 of this title to be the allocable amount of unfunded vested benefits....

29 U.S.C. Sec. 1381(b)(1). Section 1391 then provides several complex formulas for computing withdrawal liability. One of these formulas, known as the "modified presumptive method," although complex in its particulars, basically computes withdrawal liability as the sum of (a) the withdrawing employer's share of the plan's UVBs as of September 26, 1980, amortized over a fifteen-year period and (b) the employer's share of the plan's total UVBs for all years after September 25, 1980. 29 U.S.C. Sec. 1391(c)(2)(A)-(C).2

When Almont and Stevedores stopped contributing into the Fund and the Escrow Fund and the GAI Fund stopped contributing into the North Carolina Fund, the Fund and the North Carolina Fund (collectively "the Funds") attempted to charge withdrawal liability. To determine the withdrawal liability of the departing entities (collectively "the Employers"), the Funds used the modified presumptive method. In the case of each Fund, there were UVBs as of September 26, 1980, but there were no UVBs as of the year preceding the withdrawal of the Employers.3 Also in the case of each Fund, the UVBs as of September 26, 1980, were so great that when amortized and offset against the surplus of assets as of the year preceding withdrawal, pursuant to the modified presumptive method, the formula resulted in a withdrawal liability for each Employer.4 Thus, the Funds considered themselves to be entitled to withdrawal liability payments from the Employers.

Each of the Employers asserted that it owed no withdrawal liability. The Employers did not allege error in the Funds' arithmetic. Instead, the Employers argued that the statutorily prescribed formulas should not apply where, as here, a plan has no UVBs as of the end of the year preceding a participant's withdrawal. Almont and Stevedores also argued that the Fund should have applied a different computation formula. The Escrow and GAI Funds advanced the additional contention that they were not "employers" and were therefore not required to contribute anything, irrespective of whether the absence of UVBs in the year preceding withdrawal precluded application of the formula.

Pursuant to 29 U.S.C. Sec. 1401(a), the Fund, Almont and Stevedores submitted their dispute to arbitration. The arbitrator affirmed the Fund's assertion of withdrawal liability. Pursuant to 29 U.S.C. Sec. 1401(b)(2), Almont and Stevedores brought suit in the United States District Court for the Eastern District of North Carolina seeking to vacate the arbitrator's decision. The court, which also had before it a suit brought by the Escrow Fund and the GAI Fund to resolve the dispute with the North Carolina Fund, vacated the arbitrator's decision and held for the Employers in both cases. See Wise v. Ruffin, 716 F. Supp. 213 (E.D.N.C.1989) (court's opinion in the North Carolina Fund case). The court agreed with the Employers that, under MPPAA, employers withdrawing from a multiemployer pension plan owe no withdrawal liability when the plan has no UVBs at the end of the year preceding withdrawal. Because of the basis of the court's holding, there was no need to consider either the arguments concerning the particular computation formula selected or the argument of the Escrow Fund and the GAI Fund that they were not employers under MPPAA. See 716 F. Supp. at 217.

The Funds have appealed.

As is always the case when we are called upon to interpret a statute, we look first to determine if the relevant statutory language is clear. If so, our analysis ordinarily stops and we apply the statute as written. See Griffin v. Oceanic Contractors, Inc., 458 U.S. 564, 570, 102 S. Ct. 3245, 3249, 73 L. Ed. 2d 973 (1982) ("Our task is to give effect to the will of Congress, and where its will has been expressed in reasonably plain terms, 'that language must ordinarily be regarded as conclusive.' ") (citing Consumer Product Safety Comm'n v. GTE Sylvania, Inc., 447 U.S. 102, 108, 100 S. Ct. 2051, 2056, 64 L. Ed. 2d 766 (1980)). Therefore, we must determine whether the statute and, in particular, Sec. 1381(b)'s provision that "the withdrawal liability of an employer to a plan is the amount determined under section 1391 of this title to be the allocable amount of unfunded vested benefits," are clear.

To support their position, the Employers point to the language in Sec. 1381(b) that provides that withdrawal liability is to be "the allocable amount of unfunded vested benefits." The Employers argue that, because the amount of UVBs as of the end of the year preceding withdrawal was zero, their allocable amount of those UVBs must also be zero. The Funds counter that Sec. 1381(b) specifically provides that the "allocable amount of unfunded vested benefits" is to be "determined under section 1391." Therefore, the Funds conclude, given that the plain language of the statute provides that the allocable amount of UVBs is to be determined under the Sec. 1391 formulas, and that there is no statutory provision calling for disobedience to the formulas merely because the UVBs at the end of the year preceding withdrawal were zero, the plain language of the statute supports their position. The Employers reply that Sec. 1381(b)'s reference to the Sec. 1391 formulas implicitly assumes the existence of UVBs as of the end of the year preceding withdrawal.

We agree with the Funds. The explicit provision that the "allocable amount" of UVBs is to be determined under the Sec. 1391 formulas leads unambiguously to the conclusion that the specific provisions of the formulas, which, in the case of the modified presumptive method, direct consideration of a plan's UVBs as of September 26, 1980, apply. The Employers' argument amounts to a request that we read into the words "allocable amount" the notion that allocable amount is to be based only on the UVBs as of the year preceding withdrawal when those UVBs have a value of zero. Honoring the Employers' request would not constitute resolving a statutory ambiguity but instead would require that we ignore and disregard, substituting a departure for, clear statutory language. Congress did not exercise its option to insert a provision calling for disobedience to the formulas where the UVBs for the year preceding withdrawal are zero. We, of course, have no such option. Because we construe the Employers' argument as an attempt to have us rewrite plain congressional language, we are compelled to reject their argument.

It is true that "in rare cases the literal application of a statute will produce a result demonstrably at odds with the intentions of the drafters, and those intentions must be controlling." Griffin, 458 U.S. at 571, 102 S.Ct. at 3250. " 'In expounding a statute, we must not be guided by a single sentence or member of a sentence, but look to the provisions of the whole law, and to its object and policy.' " Philbrook v. Glodgett, 421 U.S. 707, 713, 95 S. Ct. 1893, 1898, 44 L. Ed. 2d 525 (1975) (citing United States v. Heirs of Boisdore, 49 U.S. (8 How.) 113, 122, 12 L. Ed. 1009 (1849)). The language and logic of MPPAA as a whole and MPPAA's legislative history, however, support, rather than contravene, our decision to give effect to the statute's plain language.

* As even a cursory glance reveals, the formulas prescribed by Congress in Sec. 1391 are meticulous and complex. See, e.g., supra n. 2. Where Congress has defined "allocable amount of unfunded vested benefits" with such specificity, we are particularly hesitant to tamper with that definition by reading into the term defined a contrary specific definition such as the one proposed by the Employers, i.e., that "allocable amount of UVBs" means only the UVBs as of the year preceding withdrawal in those situations where there were no UVBs as of that year, but not in other circumstances. Given that "where there is no clear intention otherwise, a specific statute will not be controlled or nullified by a general statute," see Morton v. Mancari, 417 U.S. 535, 550-51, 94 S. Ct. 2474, 2482-83, 41 L. Ed. 2d 290 (1974), it necessarily follows that specific language will not be controlled by general language when that general language is not even explicit. We simply can not infer from the words "allocable amount of unfunded vested benefits" the conclusion that Congress took care to account for all of the contingencies addressed in Sec. 1391 but simply forgot to insert the sweeping provision that the absence of UVBs as of the end of the year preceding withdraw shuts off the elaborate Sec. 1391 formulas. Such an oversight by Congress is even less likely in light of the fact that the absence of UVBs as of the end of the year preceding withdrawal is foreseeable and, therefore, not likely to have escaped congressional attention.5

Moreover, the Sec. 1391 formulas, as well as other parts of MPPAA, drip with the concept that computations of withdrawal liability often involve consideration of years other than simply the year preceding withdrawal. See 29 U.S.C. Sec. 1391(b) (describing the "presumptive method," calling for consideration of plan's UVBs as of September 26, 1980); id. Sec. 1391(c)(2) (describing the modified presumptive method, also calling for consideration of plan's UVBs as of September 26, 1980); id. Sec. 1391(c)(3) (describing the "rolling-5 method," calling for consideration of contributions made during the five years preceding withdrawal). The Employers would have us hold that Congress intended, though it did not so state, that consideration of those years should suddenly become irrelevant if a plan has no UVBs at the end of the year preceding withdrawal. We think it more likely that Congress would have announced such an intention, as it did on the other occasions within the same section in which it provided that the relevant time for a given calculation was to be the year preceding withdrawal. See 29 U.S.C. Sec. 1391(c)(4) (describing the "direct allocation method," explicitly calling for determination of UVBs as of the plan year preceding the year of withdrawal); see also id. Sec. 1389(a)(1) (Act explicitly provides that for purposes of de minimis rule, UVBs are to be "determined as of the end of the plan year ending before the date of withdrawal").6

In addition to its other shortcomings, the Employers' position leads to absurd results. For example, consider a situation where application of one of the statutorily prescribed formulas leads to a finding of one million dollars in withdrawal liability. Under the Employers' interpretation of the Act, if the plan has one penny of UVBs as of the end of the year preceding withdrawal, the withdrawing employer owes one million dollars in withdrawal liability; however, if the plan is one penny wealthier and has zero UVBs as of the end of the year preceding withdrawal, the withdrawing employer owes nothing. It simply can not be that Congress silently intended for such a dramatic shift in liability to be based solely on whether the plan's UVB total falls, however slightly, above or below the zero UVB threshold. The example highlights our more accurate interpretation rejecting the conclusion that Congress intended for the relevance vel non of the UVBs as of 1980 to depend on subsequent events such as the existence vel non of a positive UVB total as of the year preceding withdrawal. We are not prepared to infer from congressional silence an anomalous conclusion that would constitute a departure from the overall statutory scheme.

The Employers argue that it is the Funds' position that can lead to absurd results. The Employers point to 29 U.S.C. Sec. 1389, which creates a de minimis rule that "provides for a reduction of unfunded vested benefits allocable to an employer that withdraws from a plan." Ben Hur Construction Co. v. A.S. Goodwin, 784 F.2d 876, 880 (8th Cir.1986). The rule provides that, in certain situations, an employer's withdrawal liability is to be reduced by 3/4 of 1 percent of the plan's UVBs determined as of the end of the year preceding withdrawal.7 The Employers ask us to compare the effect of the de minimis provision under the Funds' interpretation of the statute in two different situations: "Case A," in which an employer who would otherwise owe ten thousand dollars in withdrawal liability withdraws from a plan with one million dollars in UVBs as of the end of the year preceding withdrawal; and "Case B," in which the facts are the same except the plan has zero UVBs as of the end of the year preceding withdrawal. The Employers note that, under the Funds' position, the Case A employer, though withdrawing from an underfunded plan, will owe less in withdrawal liability than the Case B employer who withdraws from a fully-funded plan.8 The Employers contend that, "[c]learly Congress neither contemplated nor intended such an arbitrary and capricious result." The Employers' claim, apparently, is that Congress could not have intended for the better funded plan to be entitled to a greater withdrawal liability. The Employers point out that if, as they assert, the absence of UVBs as of the end of the year preceding withdrawal precludes the existence of withdrawal liability, a more logical result is reached because the employer in Case B would owe nothing, which, obviously, is less withdrawal liability than what the employer in Case A would owe.

We acknowledge that the Employers' position may make Sec. 1389's application more logical in the context of the hypothetical they present. However, the problems that the Employers highlight under the Funds' interpretation resurface under even the Employers' interpretation whenever the UVBs as of the year preceding withdrawal rise above zero. For example, even under the Employers' interpretation, if there was one penny in UVBs as of the end of the year preceding withdrawal, there would be no de minimis reduction because 3/4 of 1 percent of one penny effectively equals zero. That result, with which the Employers could not reasonably quarrel, would embody the ostensibly unpalatable proposition that, at least insofar as the de minimis reduction is concerned, withdrawal liability increases as the plan becomes better funded. Thus, the feature of increasing the de minimis reduction as a plan's UVBs increase, and its effect of increasing the withdrawal liability owed to better funded plans, simply inhere in Sec. 1389; the application of the de minimis reduction to the Employers' hypothetical is no more illogical than its application to any other set of hypotheticals whose results are compelled by any credible theory of the statute.

It may be that Congress intended to use UVB totals as a measure of a plan's size and reasoned therefrom that the de minimis reduction should be greater when employers depart from plans with high UVB totals. Such a conclusion would follow from the assumption that UVB total accurately represents plan size because the larger the plan, the more de minimis, i.e., less significant, the impact of any one withdrawal. As noted above, however, we need not delay our analysis with speculation about Congress' motives because to the extent there is any illogic in the application of the de minimis rule under the Funds' interpretation, the Employers' interpretation cures that illogic only in the one limited situation in which a plan's UVBs as of the year preceding withdrawal equal zero. Thus, the Employers' argument here is really no different from the argument, rejected above, that Congress intended, though it did not say so, that a zero UVB total at the end of the year preceding withdrawal deactivates the Sec. 1391 formulas.

In light of the foregoing, the interpretation advanced by the Employers can not be squared with MPPAA as a whole.

The plain meaning interpretation advanced by the Funds has the additional virtue of being responsive to MPPAA's purposes as reflected in the statute's legislative history.

Consistent with its overall concern for adequate funding of multiemployer pension plans, Congress intended for MPPAA to eliminate employers' incentives to withdraw from multiemployer plans. See H.R.Rep. No. 96-869, 96th Cong., 2nd Sess., pt. 1 (Education and Labor Committee Report), at 67 (1980), reprinted in 1980 U.S.Code Cong. & Admin.News 2918, 2935 ("The purpose [of withdrawal liability] is to relieve the funding burden on remaining employers and to eliminate the incentive to pull out of a plan which would result if liability were imposed only on a mass withdrawal by all employers."); id., pt. 2 (Ways and Means Committee Report), at 10, 1980 U.S.Code Cong. & Admin.News at 3001 (Committee concerned that "present law provides an undesirable incentive for employers to withdraw from plans and an unfair burden on the employers who continue to maintain the plans").

Under the interpretation advanced by the Employers, a calculating employer who knew that application of the congressionally prescribed formulas would otherwise lead to a substantial withdrawal liability would be encouraged to withdraw after a year in which the plan had no UVBs. Such an employer would successfully avoid the effect of the statutory formulas, under the Employers' interpretation, because a zero UVB total for the year preceding withdrawal would deactivate the formulas. The employer would be inclined to withdraw because it would recognize that the value of the UVBs calculated as zero for one year may not be zero the next. We can not accept the proposition that Congress silently intended for a statute otherwise designed to encourage participation in multiemployer plans to have the effect of encouraging withdrawal from such plans. We find more consistency with MPPAA's legislative history in the plain meaning interpretation advanced by the Funds, under which the fact that a plan's UVBs fall to zero for one instant in a continuous period of time does not create an incentive to withdraw because it does not eviscerate withdrawal liability.9

MPPAA's legislative history also evinces a concern for the evil of saddling new participating employers with UVB liability created in the plan prior to their arrival. See H.R.Rep. No. 96-869, 96th Cong., 2nd Sess., pt. 1 (Education and Labor Committee Report), at 77 (1980), reprinted in 1980 U.S.Code Cong. & Admin.News 2918, 2945 (the presumptive method "would impose liability only with respect to those years in which an employer contributes to a plan" and "protects newly entering employers from unfunded liabilities built up before they entered the plan, in order to reduce the fears of prospective contributors caused by current allocation rules."); see also id., pt. 2 (Ways and Means Committee Report), at 15, 1980 U.S.Code Cong. & Admin.News at 3004 (Committee "intends that [withdrawal liability] rules not deter new employers from entering a multiemployer plan").

This congressional concern provides another explanation for the relevance to the withdrawal liability calculation of prior years. The formula that the Employers would have us deactivate effectively requires existing participating employers to be liable for UVBs incurred prior to the arrival of a new employer. Employers will be less likely to join multiemployer plans if they know that existing employers will avoid liability for UVBs incurred prior to the new participant's arrival simply by departing after a year on which the plan's UVBs have decreased to zero. Such a disincentive to plan participation runs afoul of MPPAA's purpose as expressed in the legislative history.10

The Employers read MPPAA's legislative history differently, arguing that the withdrawal liability provisions are intended to protect against only those withdrawals that "threaten vested pension benefits." Brief of GAI Fund and Escrow Fund at 24. "A withdrawal from a multiemployer plan that has no shortfall, as in this case, has no deleterious financial effect on the plan's ability to pay benefits and, therefore, imposes no financial burden on the [Pension Benefit Guaranty Corporation]. The legislative history contains no indication whatsoever that Congress was concerned with employer withdrawals under these circumstances." Id. at 24-25.

We reject the Employers' analysis of the legislative history for several reasons. First, their analysis is unresponsive to our observations that their interpretation, as compared to the Funds' plain meaning interpretation, encourages withdrawals and creates a disincentive to plan participation. Second, their analysis characterizes the legislative history too narrowly by implying that Congress was concerned exclusively with troubled plans. Indeed, if Congress' concern was so narrow, it is that much more peculiar that there is no provision explicitly deactivating the formulas when withdrawal follows a fully funded year. Most importantly, however, to give force to the Employers' restrictive interpretation of the legislative history would make withdrawal liability imposition litigious and unworkable. For if Congress did not intend for withdrawal liability to be imposed on a fully funded plan, it probably did not intend for such liability when the plan was only one penny short of fully funded status. And what about a plan with one dollar in UVBs or one hundred dollars in UVBs or one thousand dollars in UVBs? If we were to incorporate into our analysis the Employers' reading of the legislative history, we would be forced to draft a rule to accompany our holding in which we, as a court, would develop a formula for determining when a given withdrawal does or does not really "threaten vested pension benefits." Or, we would have to draft a rule providing that withdrawal liability may be imposed only for "those withdrawals that would have the result of foisting the burden of funding onto remaining employers and thus would create an incentive for other employers to withdraw," see Brief of GAI Fund and Escrow Fund at 29, and we could leave the elaboration of the standard to future litigation. Indeed, if we should allow the general principle that withdrawal liability should be no greater than the threat posed by a given withdrawal as determined by the amount of UVBs as of a single point in time, i.e., the year preceding withdrawal, we should do away with the Sec. 1391 formulas altogether and require that the withdrawal liability equal some allocable portion of the UVB total as of that year. Until we are elected to Congress, such finagling with the Act will remain beyond our authority.

Finally, we observe that, rather than leading to absurd results, the plain language interpretation advanced by the Funds has an additional salutary effect, which, admittedly, is not discussed in MPPAA's legislative history.

In Masters, Mates & Pilots, we noted the legitimacy of the moving market average approach to asset valuation, under which the value of an entity's assets at any one time is determined by examining the market value of the entity's assets over an extended period of time. See 900 F.2d at 730-33. Such an approach "moderates the impact of severe fluctuations in the stock market," id. at 731, by recognizing the fact that values fluctuate over time. A value showing everything vested at one point in time may change so as to demonstrate a substantial lack of vesting at some future point in time. Spreading the valuation process also serves to spread the risk.

Congress' decision to provide for consideration of UVBs in prior years seems to incorporate the notion that a plan's health is most safely valued by a method that looks not merely to the amount of UVBs at one point in time (the end of the year prior to withdrawal), but instead looks to the amount of UVBs over a longer period of time. That approach avoids unfortunate results such as a sudden skyrocketing of a plan's assets, leading to an illusory appearance of an absence of UVBs, followed by an equally sudden fall in asset value that reveals that the plan is actually underfunded. Cf. Republic Industries v. Teamsters Joint Council, 718 F.2d 628, 632 n. 2 (4th Cir.1983) (noting contention that unfunded vested liability may result when "the assets of the fund ... diminish as a result of market forces, general economic conditions or imprudent investments"), cert. denied, 467 U.S. 1259, 104 S. Ct. 3553, 82 L. Ed. 2d 855 (1984). Although we rest our decision most securely on MPPAA's plain language, logic, and legislative history, we derive additional assurance from the reasonableness, from the perspective of economic theory, of consideration of plan health over a range of years, rather than at one discrete point in time.

We recognize that our conclusion contradicts the position of the Pension Benefit Guaranty Corporation ("PBGC"), see Notice of interpretation, 51 Fed.Reg. 47,342 (1986), and that the PBGC ordinarily has considerable authority for purposes of interpreting MPPAA. See Belland v. Pension Benefit Guaranty Corp., 726 F.2d 839, 843 (D.C.Cir.) (citations omitted), cert. denied, 469 U.S. 880, 105 S. Ct. 245, 83 L. Ed. 2d 183 (1984). In addition, at first blush, our conclusion appears to run counter to the conclusion reached by the First Circuit in Berkshire Hathaway, Inc. v. Textile Workers Pension Fund, 874 F.2d 53 (1st Cir.1989), although it does appear to find support in the conclusion of the Eighth Circuit in Ben Hur Construction Co. v. A.S. Goodwin, 784 F.2d 876 (8th Cir.1986). We conclude, however, that the PBGC is not entitled to deference in this context and that the First Circuit's conclusion in Berkshire Hathaway should not disturb our result.

* In its most recent pronouncement on the matter, found in a notice of interpretation, the PBGC has changed its position and ruled that an employer that withdraws from a plan that was fully funded at the end of the year preceding withdrawal owes no withdrawal liability, irrespective of the result of application of the statutorily prescribed formulas. Compare Notice of interpretation, 51 Fed.Reg. 47,342 (1986) (stating current position) with PBGC Opinion Letter 83-19 (August 11, 1983) (stating former position). The notice concludes that "while [29 U.S.C. Sec. 1391] is ambiguous as to whether an employer that withdraws from a plan that has no unfunded vested benefits as of the end of the preceding plan year may incur withdrawal liability, reference to the legislative history and other provisions of Title IV clears up this ambiguity and compels the conclusion that no withdrawal liability is to be assessed in such a case." 51 Fed.Reg. at 47,344.

Our review of the PBGC's current interpretation of the law is guided by several familiar principles. First, "[i]f the intent of Congress is clear, that is the end of the matter; for the court, as well as the agency, must give effect to the unambiguously expressed intent of Congress.... [I]f the statute is silent or ambiguous with respect to the specific issue, the question for the court is whether the agency's answer is based on a permissible construction of the statute." Chevron, U.S.A., Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 842-43, 104 S. Ct. 2778, 2781-82, 81 L. Ed. 2d 694 (1984) (footnotes omitted). Second, "[a]n agency interpretation of a relevant provision which conflicts with the agency's earlier interpretation is 'entitled to considerably less deference' than a consistently held agency view." I.N.S. v. Cardoza-Fonseca, 480 U.S. 421, 446 n. 30, 107 S. Ct. 1207, 1221 n. 30, 94 L. Ed. 2d 434 (1987) (citations omitted).

Given such a standard of review, we elect not to defer to the PBGC's current position. The cornerstone of the PBGC's position and the origin of the PBGC's ostensible entitlement to deference under Chevron lie in the proposition that the statute is ambiguous. To us, nothing could be clearer than Sec. 1391(c)(2)(B)'s explicit directive to include in the withdrawal liability computation a portion of an amortization of "the plan's unfunded vested benefits as of the end of the last plan year ending before September 26, 1980." Nothing in the PBGC's notice persuades us that we have overlooked some source of ambiguity. The notice concedes that the PBGC's former position, with which we do agree, was at least "apparently straightforward." 51 Fed.Reg. at 47,343. However, the notice asserts in one terse conclusory sentence, "the use of the term 'allocable' at least suggests the prior existence of unfunded vested benefits to be allocated." Id. This statement ignores the elaborate and unambiguous statutory definitions of "allocable" that Sec. 1391 provides and essentially begs the ultimate question, conclusively answered by the statute, of the date at which the existence of UVBs is to be measured. Yet, the notice immediately proceeds to employ various tools of statutory interpretation that are not applicable to unambiguous statutes such as the one it was analyzing. Given the clarity of the statute, we believe that the notice should be viewed merely as a statement of why the PBGC believes Congress should amend the statute and not as an interpretation of an ambiguous provision that is entitled to Chevron deference. See Chevron, 467 U.S. at 843, n. 9, 104 S. Ct. at 2781, n. 9 ("The judiciary is the final authority on issues of statutory construction and must reject administrative constructions which are contrary to clear congressional intent. If a court, employing traditional tools of statutory construction, ascertains that Congress had an intention on the precise question at issue, that intention is the law and must be given effect.").

In Berkshire Hathaway, the First Circuit found that MPPAA did not permit imposition of withdrawal liability under the computation method employed by the plan in that case, where the plan was fully funded as of the end of the year preceding withdrawal. However, unlike in the case before us, in which the plan has used the modified presumptive method, the plan in Berkshire Hathaway employed a variation of the "direct allocation method," see 29 U.S.C. Sec. 1391(c)(4), as the means of computing withdrawal liability. In contrast to the modified presumptive method, the direct allocation method used in Berkshire Hathaway does not so explicitly require inclusion of UVB amounts for a year other than the year immediately preceding withdrawal. To the contrary, the statutory description of the method used by the plan in Berkshire Hathaway contains explicit directives to base computations upon the year preceding withdrawal.11 The particular provision that formed the basis of the plan's argument in Berkshire Hathaway provided:

The plan's unfunded vested benefits which are attributable to participants' service with the employer is the amount equal to the value of nonforfeitable benefits under the plan which are attributable to participants' service with such employer ... decreased by the share of plan assets ... which is allocated to the employer....

29 U.S.C. Sec. 1391(c)(4)(B); see 874 F.2d at 55-56. The plan in Berkshire Hathaway argued from congressional silence that it would be wrong to conclude that the quoted provision assumes the existence of some plan unfunded vested benefits before computing each employer's share. Thus, unlike the Funds before us, who simply ask that we apply pellucidly clear statutory directives to include UVBs as of 1980 in the withdrawal liability calculation, the plan in Berkshire Hathaway asked the court to inject a certain meaning into language that was, at best, from the plan's perspective, ambiguous.

Thus, the conflict between our holding and that of the distinguished panel in Berkshire Hathaway may be more apparent than real. Given the difference in the formulas at issue, the First Circuit's application of the "equally plausible and perhaps more intuitive," see 874 F.2d at 56, interpretation of the PBGC is not necessarily at odds with the conclusion we reach. Undeniably, the Berkshire Hathaway court believed that the PBGC position was entitled to deference. See 874 F.2d at 55. However, the court's decision ultimately rested upon its conclusion that the PBGC's position comported with the statutory language at issue in the case before it. See id. ("Perhaps most importantly, PBGC's interpretation is consistent with the statutory language, and Congress expressly delegated substantial regulatory authority to PBGC relating to withdrawal liability.") (emphasis added). Accordingly, our conclusion will not be shaken by the fact that the First Circuit, in another context, deferred to a PBGC notice of interpretation to which we, in a different context, choose not to defer.12 And, of course, our detailed treatment of Berkshire Hathaway should not detract attention from the fact that the Eighth Circuit interpreted the direct allocation method in the same manner in which we interpret the modified presumptive method, albeit without the PBGC opinion before it. See Ben Hur (decided prior to the PBGC's shift in positions).

For the foregoing reasons, the judgments of the district court are

AFFIRMED IN PART, REVERSED IN PART, AND REMANDED.13