Raymond ZIELINSKI, et al., individually and on behalf of all persons similarly situated, Plaintiffs-Appellants, v. PABST BREWING COMPANY, INC. Defendant-Appellee, 463 F.3d 615

POSNER, Circuit Judge. This case raises in a novel setting the vexing question of the

interpretation of ERISA welfare plans that specify no date of termination. Back in 1971, Blue Cross-Blue Shield offered a prescription-drug plan that reimbursed enrolled members’ entire drug costs subject only to a $2 deductible for a 34-day supply of the drug. Beginning in 1973 several Milwaukee breweries, including Pabst and Schlitz, agreed in successive collective bargaining agreements with a local union to provide the coverage described in the Blue Cross-Blue Shield plan to their retired employees plus (for six months after the retiree’s death) spouses and dependents.

In 1981 Schlitz closed its Milwaukee brewery and signed a shutdown agreement with the union that supplanted the then-current collective bargaining agreement but provided that “the Company shall continue to provide the health and welfare benefits for retirees [and their spouses and dependents] described in” the prescription-drug provision of the collective bargaining agreements if they retired on or before January 1, 1982. The shutdown agreement contained no termination date.

Pabst succeeded Schlitz as the obligor of the shutdown agreement in 1999 and five years later precipitated — by reducing the benefits specified in the 1971 Blue Cross-Blue Shield plan — this lawsuit on behalf of more than 500 Schlitz employees who had retired on or before January 1, 1982, plus spouses and dependents.

The reductions were as follows: the $2 deductible was limited to generic drugs, and only for a 30-day supply; for brand-name drugs the deductible was raised to $35 or (for “non-preferred” brands) $50; and an annual cap of $3,000 was placed on reimbursement of drug costs and those costs would for the first time be counted against the $20,000 lifetime ceiling on major-medical coverage. Administrators of the prescription-drug plan had made changes in the plan before but most of the changes had expanded rather than contracted benefits. Not all, though.

For example, one had required enrollees to substitute generic for brand-name drugs except when their doctor authorized the purchase of a particular brand-name drug. The plaintiffs seek injunctive relief under ERISA and also damages under the Labor Management Relations Act, 29 U.S.C. § 185(a), which provides a damages remedy for the breach of a labor agreement, including shutdown agreements, see 29 U.S.C. § 185(a); Baker v. Kingsley, 387 F.3d 649, 659-60 (7th Cir.2004); Diehl v. Twin Disc, Inc., 102 F.3d 301, 305 (7th Cir.1996), and is not preempted by ERISA. 29 U.S.C. § 1144(d); Bidlack v.

Wheelabrator Corp., 993 F.2d 603, 604-05 (7th Cir.1993) (en banc) (lead opinion); Bugher v. Feightner, 722 F.2d 1356, 1358-60 (7th Cir.1983); Morse v. Adams, 857 F.2d 339, 343 (6th Cir.1988). The district judge granted

summary judgment for Pabst. He reasoned that because there was no statement in either the collective bargaining agreements that preceded the shutdown agreement or in that agreement that the benefits conferred by the prescription-drug program were vested, the plaintiff class had no legally enforceable right to them. We do not find this reasoning persuasive. The shutdown agreement contains as we said no termination date, and we cannot find any basis for interpolating one.

It is hard to read “shall continue” otherwise than as creating a contractual obligation that indeed continues until six months after the last retiree dies (if he has a surviving spouse, or dependents). The presumption against vesting on which the judge relied to overcome “shall continue” has reference mainly to suits to enforce collective bargaining agreements. Litton Financial Printing Division v. NLRB, 501 U.S. 190, 207-08, 111 S.Ct. 2215, 115 L.Ed.2d 177 (1991); Bidlack v. Wheelabrator Corp., supra, 993 F.2d at 606-07; Des Moines Mailers Union, Teamsters Local No. 358 v.

NLRB, 381 F.3d 767, 769-70 (8th Cir.2004); International Union, United Automobile, Aerospace & Agricultural Implement Workers of America, U.A.W. v. Skinner, 188 F.3d 130, 141 (3d Cir.1999). Those, being short-term agreements, are presumed not to create rights or duties that continue after the agreement’s termination date. The rights that the plaintiffs assert in this case originated in collective bargaining agreements but were carried forward into the shutdown agreement, which, unlike a collective bargaining agreement, has no end date. It is true that in Diehl v.

Twin Disc, Inc., supra, 102 F.3d at 306-09, we analyzed a shutdown agreement in the same manner and with the same presumption against vesting that courts use in interpreting collective bargaining agreements. Although it was not a collective bargaining agreement, it was short term — only 13 months, shorter indeed than the term of the standard collective bargaining agreement, which is three years. So we applied the presumption derived from cases involving benefits language in collective bargaining agreements.

But the presumption could not overcome the shutdown agreement’s explicit provision of benefits “for the lifetime of the pensioner” and the equally unambiguous provision of lifetime benefits to retirees. Id. at 306-07. So we went on to consider, as we shall here, precisely what benefits the plaintiffs were entitled to.

The shutdown agreement in this case is not short term — in fact has no end date (though it would expire eventually, as we’ll see) — and, in any event, no more than the language in Diehl would its language of “shall continue” allow an inference that Pabst’s contractual obligations terminated any earlier than six months after the death of the last retiree. And so Pabst remains obligated to provide prescription-drug benefits.

But also as in Diehl it does not follow, as the plaintiffs argue, that the scope of the obligation was forever fixed by the terms of the 1971 Blue Cross-Blue Shield program — that the chain of incorporations from the shutdown agreement through the collective bargaining agreements back to that plan binds Pabst to every detail of a plan that no longer even exists.

Because workers could retire from Schlitz after 25 years of service — and a worker who retired at 50 in 1982 might well live another 30 years — the shutdown agreement was a very long-term contract, though the provider’s total obligations would diminish as retirees died off, and eventually would cease altogether. The longer the term of a contract, the more likely it is to contain provisions that adjust for future contingencies, such as an escalator clause to adjust for inflation. See PSI Energy, Inc. v. Exxon Coal USA, Inc., 991 F.2d 1265, 1266-67 (7th Cir.1993); Karen Eggleston, Eric A.

Posner & Richard Zeckhauser, “The Design and

Interpretation of Contracts: Why Complexity Matters,” 95 Nw. U.L.Rev. 91, 126 n. 101 (2000). The shutdown agreement contains no such provisions. But the likeliest reason it does not is that it does not actually specify any long-term benefits.

It merely incorporates by reference the benefits provision of a collective bargaining agreement, and since a collective bargaining agreement is short term, the parties rarely bother to negotiate provisions designed to cope with the profound uncertainties about what the distant future may hold. “A short-term contract ... is more likely to have ‘unforeseen’ events disrupt party expectations, precisely because the parties are unlikely to focus attention on the possibility of change in the interval between the signing of the contract and its expiration.” Subha Narasimhan, “Of Expectations, Incomplete Contracting, and the Bargain Principle,” 74 Calif.

L.Rev. 1123, 1192 n. 211 (1986). This is true of collective bargaining agreements as well as other short-term contracts. See Harry Shulman, “Reason, Contract, and Law in Labor Relations,” 68 Haro. L.Rev. 999, 1004-05 (1955). How the parties to a contract actually perform their contractual undertakings is often and here persuasive evidence of what the parties understood the contract to require.

That the words “shall continue” in the shutdown agreement were not meant to create an inflexible obligation is thus suggested by the frequent changes that the plan administrators made in the drug program' — some of which, at least, were disadvantageous to the participants (we gave an example earlier) — without evoking a peep of protest from anyone. The administrators would hardly have been likely to increase benefits had they thought the increase would set a new floor for the indefinite future. But that is not to imply a tacit agreement that the administrators could cut the benefits to zero.

At once defending this extreme position and flinching from it, Pabst’s lawyer told us at argument that as long as the cut did not reduce the benefits to a de minimis level, it would not violate the shutdown agreement. But de minimis covers a range that ends just above zero, and “shall continue at a level arbitrarily close to zero” is not a plausible

interpretation of “shall continue.” Pabst’s lawyer acknowledged however the possibility of an intermediate position- — that any reduction in benefits below the level specified in the shutdown agreement had to be reasonable. And his brief had referred us to affidavits from an expert on health insurance that the changes that Pabst made in the plan it inherited were reasonable in light of changes in health care since 1971 that had affected health insurance; the expert stressed the greatly increased availability of generic substitutes for brand-name drugs.

Adamant that the level of benefits could not be reduced by even a penny, the plaintiffs presented no counterevidence; and in this court, continuing in that vein, they dismiss the expert’s affidavits as irrelevant. Pabst is on to something. The most sensible

interpretation of the shutdown agreement, the

interpretation that steers between implausible extremes, is that it obligates the company to provide prescription-drug benefits at a level “reasonably commensurate” (to quote earlier cases, broadly similar to this one) with the 1971 Blue Cross-Blue Shield plan. Diehl v. Twin Disc, Inc., supra, 102 F.3d at 311; Barker v. Ceridian Corp., 122 F.3d 628, 638 (8th Cir.1997); Poole v. City of Waterbury, 266 Conn. 68, 831 A.2d 211, 233-34 (2003).

Holding Pabst to the literal terms of that ancient plan in today’s marketplace would give the retirees an insanely generous plan relative to today’s norms, rather than a plan that would provide them with the same coverage, mutatis mutandis, that they would have had in 1974. (The Blue Cross-Blue Shield plan was created in 1971, but the breweries did not start enrolling their retirees in it until 1974.) That is one reason to doubt the plaintiffs

interpretation. The indefinite duration of the shutdown agreement, coupled with its structure — the chain of incorpo-rations going back to 1971 — is another. But the most compelling reason is the sheer impossibility of determining Pabst’s obligations if the only guide is the old Blue Cross-Blue Shield brochure — yet it is the only evidence of the plan’s terms.

According to the brochure, the plan provides reimbursement for “covered drugs.” This implies that not all drugs in existence in 1971 were eligible for reimbursement, which in turn implies that not all newly developed drugs would be eligible either; an affirmative decision by the plan to “cover” them would be required. Most prescription-drug plans limit the circumstances in which they will reimburse the cost of a given drug (for example, minoxidil for hypertension but not for hair loss), and exclude some drugs from coverage altogether.

The brochure’s language suggests that the Blue Cross-Blue Shield plan was like other plans in this respect. But being just a brochure, it does not spell out which drugs are covered or the criteria for determining coverage of future drugs and future (different, or better understood, or more treatable) medical conditions. Maybe “the contract” to which the brochure alludes, whose “coverage, exclusions, amendments and other provisions” qualify the brochure’s statements, would answer these questions.

But neither party has produced or invoked the contract, suggesting that it’s been lost; in any event, no arguments based upon it have been made. So the drug provision in the shutdown agreement contains gaps. Filling gaps is a standard activity of courts in contract cases. William B. Tanner Co. v. Sparta-Tomah Broadcasting Co., 716 F.2d 1155, 1158-60 (7th Cir.1983); Dobson v. Hartford Financial Services Group, Inc., 389 F.3d 386, 399 (2d Cir.2004); Black Horse Lane Associates, L.P. v. Dow Chemical Corp., 228 F.3d 275, 285-86 (3d Cir. 2000); Restatement (Second) of Contracts § 204 and comment b (1981).

As we explained in Haslund v. Simon Property Group, Inc., 378 F.3d 653, 655 (7th Cir. 2004), “If contracting parties had to provide for every contingency that might arise, contract negotiations would be interminable. Contracts can be shorter and simpler and cheaper when courts stand ready to fill gaps and resolve ambiguities in the minority of contracts that get drawn into litigation.” Particularly apt are our words in the William B. Tanner ease: “While we find no ambiguity of expression to be present in the contract, we do find a deficiency or omission of scope.” 716 F.2d at 1159.

But the gap in this case can be filled only if “reasonable commensurability” is ascertainable here by the methods of litigation, an issue the district judge did not address. (Nor did he mention the expert’s affidavits.) The qualification is implicit in Diehl, where the shutdown agreement also incorporated the health-benefits provision of an old collective bargaining agreement. For we didn’t try to apply the standard of reasonable commensurability to the facts of that case, and we won’t try here; but we will try to provide a little guidance to the district judge.

The Blue Cross-Blue Shield brochure notes that “in a recent year, Americans spent $3.7 billion for 1.1 billion out-of-hospital prescriptions.” This is an average of $3.36 per 34-day supply (apparently the quantity to which the quotation refers, though like much of the original plan this is obscure). That is 1.68 times the $2 deductible.

Were the deductible increased to the point at which the current average price (of a brand-name drug — for remember that Pabst has not raised the deductible for generics, though it has reduced its obligation to reimburse by limiting reimbursement to the cost of a 30-day supply) were 1.68 times the deductible, the members of the plaintiff class would be receiving a benefit reasonably commensurate with the benefits provided by the shutdown agreement. (So if the average price today is $20, the deductible would be $11.90 ($20/1.68).) With this adjustment, there would be no reason to cut down the quantity of drugs covered by the deductible from 34 to 30 days.

Calculating a benefits ceiling reasonably commensurate with the ceilings in the shutdown agreement is more difficult — for there were no ceilings in the agreement. But average drug prices were much lower in 1981 (not to mention 1971) and drug usage less extensive. “Prescription drug expenditures have grown at double-digit rates every year since 1980.” Craig Copeland, “Prescription Drugs: Issues of Costs, Coverage, and Quality,” Employee Benefits Research Institute Issue Brief, Apr. 1999, p. 3, http:// www.ebri.org/pdf/briefspdfi0499ib.pdf. That is why there is now a Medicare prescription-drug plan.

Supposing without deciding that all or virtually all prescription-drug programs created in 2004 (the year of the challenged reduction in benefits) contained yearly and lifetime ceilings, the highest of those ceilings would provide reasonable commensurability with the benefits provision of the shutdown agreement. A further adjustment would be appropriate to reflect the new Medicare program.

The shutdown agreement actually requires deduction of benefits received under a prescription-drug program created under future “hospital-surgical legislation,” which does not describe the new Medicare program but perhaps gestures, if blindly, toward it. In sum, the benefits to which the plaintiff class is entitled are those specified in the shutdown agreement but adjusted — to the extent possible without wild conjecture — for changes to which the parties to the agreement would have agreed had they focused at the outset on the duration of the commitment made by the employers.

The district judge will want to pay particular attention to the pharmaceutical-benefits packages that employees in the brewing industry negotiate through their unions today, as the brewers’ union did lo these many years ago. Those packages should approximate, at least roughly, the value of the 1971 Blue Cross-Blue Shield plan updated to 2004. The judgment is vacated and the case remanded for further proceedings consistent with this opinion.