EASTERBROOK, Chief Judge.
Reynolds Group Holdings acquired Pactiv Corp. in 2010. The acquisition agreement, which made Pactiv a wholly owned subsidiary, calls for severance pay to any non-union employee let go without cause, within a year, as a result of the transaction.
After the closing, Pactiv established a severance-pay plan. The clause in the acquisition agreement was skeletal; the plan itself contained many implementing terms, including a requirement that the departing worker execute a “separation agreement”, “in a form acceptable to the Company,” releasing all other claims against Pactiv.
Within a year of the acquisition Pactiv directed Chad Rupert to move to a new city. He declined, and Pactiv acknowledged that this entitled Rupert to severance pay under the plan. It sent him a separation agreement — which contained some unwelcome terms. Pactiv demanded that Rupert promise that for the next year he would not work for any of its rivals in research and development, solicit the sale of any competing goods and services, or try to hire any of Pactiv’s staff. Observing that he had not previously been subject to a restrictive covenant, and contending that he should not be required to submit to one under the severance plan, Rupert declined to sign. Pactiv withheld the severance benefits, and Rupert sued. So did Pactiv, seeking a declaration that its decision is lawful.
The district court held that Rupert is entitled to the benefits because the formal plan, governed by ERISA (the Employee Retirement Income Security Act), lacks any language conditioning severance pay on signing a restrictive covenant. 2012 WL 5974986, 2012 U.S. Dist. LEXIS 158413 (N.D.Ill. Nov. 1, 2012). Plans and their material terms must be in writing. 29 U.S.C. § 1102(a)(1); Curtiss-Wright Corp. v. Schoonejongen, 514 U.S. 73, 83, 115 S.Ct. 1223, 131 L.Ed.2d 94 (1995). This plan contemplates a separation agreement “acceptable to the Company” but does not say or imply that Pactiv can use this as a hook to add any limit it wants. Then the benefits staff could use the lure of severance pay to demand that the worker hand over the deed to the worker’s vacation home or promise to run errands [1001]*1001for the staffer’s family. Perhaps it would be within the scope of this plan to use the separation agreement to obtain a departing employee’s promise to honor the terms of a covenant already signed; the plan itself makes benefits contingent on keeping Pactiv’s secrets confidential; but the agreement tendered to Rupert demanded more on his part. ERISA’s writing requirement prevents the employer from making greater demands as a matter of discretion. Cirulis v. UNUM Corp. Severance Plan, 321 F.3d 1010 (10th Cir.2003), holds exactly this about a situation similar to Rupert’s.
As it happens, however, Rupert did not ask the district judge to award benefits under Pactiv’s plan. He did not rely on § 1102(a)(1) or Cirulis. Instead he asked for benefits under the acquisition agreement. He repeatedly told the district court that the plan is irrelevant to his claim and that he does not seek benefits under it. We asked Rupert’s lawyer at oral argument whether he had ever reserved reliance on the plan as a fallback position; counsel said no. The district court rejected the theory that Rupert actually advanced (we discuss it below) but nonetheless decided in his favor on the theory that he abjured.
Judges sometimes have the authority to relieve parties of their forfeitures (though Rupert never asked), but if they do this they must notify the other side, so that it can meet the argument. For its part, Pactiv moved for summary judgment under ERISA, contending that despite appearances some of the language in the plan does allow it to insist on restrictive covenants. Since Rupert had foresworn any argument based on the plan, Pactiv did not elaborate — and the district judge, taking up the subject without the benefit of briefs from either side, did not discuss any of the language from the “Return of Severance Payments” section of the plan that Pactiv has brought to our attention and would have brought to the district judge’s, had it known that a claim based on the plan was about to be adjudicated.
Many decisions in this circuit hold that a district judge must notify the litigants, and invite the submission of evidence and legal arguments, before resolving a case on a ground the parties have bypassed or using a procedure they did not propose. See, e.g., Goldstein v. Fidelity & Guaranty Insurance Underwriters, Inc., 86 F.3d 749 (7th Cir.1996); Choudhry v. Jenkins, 559 F.2d 1085 (7th Cir.1977). See also Celotex Corp. v. Catrett, 477 U.S. 317, 326, 106 S.Ct. 2548, 91 L.Ed.2d 265 (1986) (“district courts are widely acknowledged to possess the power to enter summary judgments sua sponte, so long as the losing party was on notice that she had to come forward with all of her evidence”). The norm is that judges must not take litigants by surprise. The 2010 amendment to Rule 56(f) makes this explicit:
After giving notice and a reasonable time to respond, the court may:
(1) grant summary judgment for a nonmovant;
(2) grant the motion on grounds not raised by a party; or
(3) consider summary judgment on its own after identifying for the parties material facts that may not be genuinely in dispute.
Fed.R.Civ.P. 56(f) (emphasis added). If judges could decide suits without warning on the basis of considerations the litigants were not contesting, litigation would be even less manageable than it is already. Lawyers would need to submit evidence and legal arguments on issues that appeared to be irrelevant, on the off chance that the judge would second guess the parties’ litigation strategies. That would produce delay, bloat, and expense. As a norm, waivers are forever. If a waived or [1002]*1002forfeited issue is to come back to life, the revival must be preceded by notice. Litigants then safely can limit their submissions to the subjects genuinely in dispute.
In this court Rupert has thought better of the choices he made earlier. His appellate brief relies heavily on Cirulis. He insists that he did not need to ask the district court for relief under the plan, because that issue had been raised in the pleadings. Yet motions for summary judgment, and responses to them, supersede the pleadings. A litigant can’t use pleadings filed on Day 1 to displace choices made on Day 2; time moves in only one direction. What’s more, even in this court Rupert does not respond to Pactiv’s argument that some language in the severance plan allows it to insist on restrictive covenants. Instead he observes that the language on which Pactiv relies is missing from the acquisition agreement.
Free access — add to your briefcase to read the full text and ask questions with AI
EASTERBROOK, Chief Judge.
Reynolds Group Holdings acquired Pactiv Corp. in 2010. The acquisition agreement, which made Pactiv a wholly owned subsidiary, calls for severance pay to any non-union employee let go without cause, within a year, as a result of the transaction.
After the closing, Pactiv established a severance-pay plan. The clause in the acquisition agreement was skeletal; the plan itself contained many implementing terms, including a requirement that the departing worker execute a “separation agreement”, “in a form acceptable to the Company,” releasing all other claims against Pactiv.
Within a year of the acquisition Pactiv directed Chad Rupert to move to a new city. He declined, and Pactiv acknowledged that this entitled Rupert to severance pay under the plan. It sent him a separation agreement — which contained some unwelcome terms. Pactiv demanded that Rupert promise that for the next year he would not work for any of its rivals in research and development, solicit the sale of any competing goods and services, or try to hire any of Pactiv’s staff. Observing that he had not previously been subject to a restrictive covenant, and contending that he should not be required to submit to one under the severance plan, Rupert declined to sign. Pactiv withheld the severance benefits, and Rupert sued. So did Pactiv, seeking a declaration that its decision is lawful.
The district court held that Rupert is entitled to the benefits because the formal plan, governed by ERISA (the Employee Retirement Income Security Act), lacks any language conditioning severance pay on signing a restrictive covenant. 2012 WL 5974986, 2012 U.S. Dist. LEXIS 158413 (N.D.Ill. Nov. 1, 2012). Plans and their material terms must be in writing. 29 U.S.C. § 1102(a)(1); Curtiss-Wright Corp. v. Schoonejongen, 514 U.S. 73, 83, 115 S.Ct. 1223, 131 L.Ed.2d 94 (1995). This plan contemplates a separation agreement “acceptable to the Company” but does not say or imply that Pactiv can use this as a hook to add any limit it wants. Then the benefits staff could use the lure of severance pay to demand that the worker hand over the deed to the worker’s vacation home or promise to run errands [1001]*1001for the staffer’s family. Perhaps it would be within the scope of this plan to use the separation agreement to obtain a departing employee’s promise to honor the terms of a covenant already signed; the plan itself makes benefits contingent on keeping Pactiv’s secrets confidential; but the agreement tendered to Rupert demanded more on his part. ERISA’s writing requirement prevents the employer from making greater demands as a matter of discretion. Cirulis v. UNUM Corp. Severance Plan, 321 F.3d 1010 (10th Cir.2003), holds exactly this about a situation similar to Rupert’s.
As it happens, however, Rupert did not ask the district judge to award benefits under Pactiv’s plan. He did not rely on § 1102(a)(1) or Cirulis. Instead he asked for benefits under the acquisition agreement. He repeatedly told the district court that the plan is irrelevant to his claim and that he does not seek benefits under it. We asked Rupert’s lawyer at oral argument whether he had ever reserved reliance on the plan as a fallback position; counsel said no. The district court rejected the theory that Rupert actually advanced (we discuss it below) but nonetheless decided in his favor on the theory that he abjured.
Judges sometimes have the authority to relieve parties of their forfeitures (though Rupert never asked), but if they do this they must notify the other side, so that it can meet the argument. For its part, Pactiv moved for summary judgment under ERISA, contending that despite appearances some of the language in the plan does allow it to insist on restrictive covenants. Since Rupert had foresworn any argument based on the plan, Pactiv did not elaborate — and the district judge, taking up the subject without the benefit of briefs from either side, did not discuss any of the language from the “Return of Severance Payments” section of the plan that Pactiv has brought to our attention and would have brought to the district judge’s, had it known that a claim based on the plan was about to be adjudicated.
Many decisions in this circuit hold that a district judge must notify the litigants, and invite the submission of evidence and legal arguments, before resolving a case on a ground the parties have bypassed or using a procedure they did not propose. See, e.g., Goldstein v. Fidelity & Guaranty Insurance Underwriters, Inc., 86 F.3d 749 (7th Cir.1996); Choudhry v. Jenkins, 559 F.2d 1085 (7th Cir.1977). See also Celotex Corp. v. Catrett, 477 U.S. 317, 326, 106 S.Ct. 2548, 91 L.Ed.2d 265 (1986) (“district courts are widely acknowledged to possess the power to enter summary judgments sua sponte, so long as the losing party was on notice that she had to come forward with all of her evidence”). The norm is that judges must not take litigants by surprise. The 2010 amendment to Rule 56(f) makes this explicit:
After giving notice and a reasonable time to respond, the court may:
(1) grant summary judgment for a nonmovant;
(2) grant the motion on grounds not raised by a party; or
(3) consider summary judgment on its own after identifying for the parties material facts that may not be genuinely in dispute.
Fed.R.Civ.P. 56(f) (emphasis added). If judges could decide suits without warning on the basis of considerations the litigants were not contesting, litigation would be even less manageable than it is already. Lawyers would need to submit evidence and legal arguments on issues that appeared to be irrelevant, on the off chance that the judge would second guess the parties’ litigation strategies. That would produce delay, bloat, and expense. As a norm, waivers are forever. If a waived or [1002]*1002forfeited issue is to come back to life, the revival must be preceded by notice. Litigants then safely can limit their submissions to the subjects genuinely in dispute.
In this court Rupert has thought better of the choices he made earlier. His appellate brief relies heavily on Cirulis. He insists that he did not need to ask the district court for relief under the plan, because that issue had been raised in the pleadings. Yet motions for summary judgment, and responses to them, supersede the pleadings. A litigant can’t use pleadings filed on Day 1 to displace choices made on Day 2; time moves in only one direction. What’s more, even in this court Rupert does not respond to Pactiv’s argument that some language in the severance plan allows it to insist on restrictive covenants. Instead he observes that the language on which Pactiv relies is missing from the acquisition agreement. That effectively retreats to his original theory of liability — that the acquisition agreement between Pactiv and Reynolds Group entitles him to benefits.
Rupert favors this argument for two reasons: first, the acquisition agreement lacks any of the qualifications (including the need for a signed separation agreement) found in the post-acquisition plan; second, the acquisition agreement has an attorneys’-fees clause more favorable to prevailing parties than the one in ERISA. But the stumbling block, from Rupert’s perspective, is § 9.6(b) of the 'acquisition agreement, which provides that the agreement cannot be invoked by anyone other than Pactiv and Reynolds Group. Section 9.6(b) has a few exceptions, but § 6.4(d), which calls for severance benefits, is not among them. Relying on both the general disclaimer in § 9.6(b) and the absence of § 6.4(d) from the listed exceptions, the district court rejected Pactiv’s claim to be a third-party beneficiary of the acquisition agreement.
In the district court, and again on appeal, Rupert contended that § 9.6(b) is ineffectual under Delaware law. (The acquisition agreement has a choice-of-law clause pointing to Delaware.) It is a poor argument. The Delaware decisions on which Rupert relies stand for the normal principle that contracts can create third-party beneficiaries. None of the decisions concludes that parties are forbidden to limit the set of non-signatories who acquire rights under a contract. Pactiv and Reynolds Group may have included § 6.4(d) because awarding severance benefits would make the transaction more valuable to themselves. Merging businesses often fear that uncertainty will cause essential employees to jump ship. Promising to compensate workers adversely affected by an acquisition makes it easier for employees to conclude that they will be well treated; this improves the chance that people will remain, and improves the morale (and hence the productivity) of those who do remain. We agree with the district court that Delaware would not see § 6.4(d) as having value only to workers, and thus as necessarily creating third-party benefits despite § 9.6(b). (This makes it unnecessary to decide whether any rule of Delaware law enforcing for workers’ benefit a promise in corporate documents about welfare-benefit plans would be preempted by ERISA. See Bartholet v. Reishauer A.G., 953 F.2d 1073 (7th Cir.1992) (ERISA preempts claims, based on state law, concerning promises to create welfare-benefit plans).)
Although Rupert’s main argument is that he is entitled to benefits , as a third-party beneficiary of § 6.4(d), he briefly contends that § 6.4(d) is itself an ERISA plan. If it is, then § 9.6(b) becomes irrelevant. Workers are not third-party beneficiaries of a welfare-benefit plan governed [1003]*1003by ERISA. Participants are the direct beneficiaries.
Pactiv does not contend that Rupert has waived or forfeited this possibility. It is a difficult one to evaluate, because the language in the acquisition agreement is skeletal. Here is the text of § 6.4(d), captioned “Severance Policy”:
1. A Non-Union Employee terminated without cause shall be paid, in addition to all amounts accrued or required by law to be paid, a sum equal to (i) the number of years such Non-Union Employee has been employed by the Company (or its Affiliates, including entities acquired by the Company), rounded to the nearest month times two (such number the “Severance Period”) times (ii) such Non-Union Employee’s weekly compensation, provided that the amount computed by clause (i) shall not be less than four nor greater than 26.
2. If such Non-Union Employee is eligible for an annual bonus in the year of termination, such person shall be paid a pro-rated amount of his or her bonus, at “target” if such bonus is incentive based, as well as any accrued but unpaid bonuses.
In contending that this language creates a welfare-benefit plan, Rupert relies on Halliburton Co. Benefits Committee v. Graves, 463 F.3d 360 (5th Cir.2006), rehearing denied, 479 F.3d 360 (5th Cir.2007).
Halliburton held that language in a merger agreement amended an existing welfare-benefit plan. Because the plan predated the merger, language disclaiming third-party beneficiaries did not matter; as we pointed out above, employees are the direct beneficiaries of ERISA plans. The court made it explicit, when denying rehearing, that the holding is “limited to the specific language used” (479 F.3d at 361) in amending that particular plan. Pactiv, unlike Halliburton and its merger partner, did not have a severance plan before the acquisition. This creates problems, because the acquisition agreement lacks many terms that are either mandatory in all ERISA plans (such as provision for amendment and identification of the person holding the amending power, 29 U.S.C. § 1102(b)(3)) or are needed to make a plan work. Who is the administrator? How are claims made and resolved? Are any conditions attached to the benefits? What do phrases such as “without cause” mean? It is not defined in the acquisition agreement. Is refusal to accept a reassignment “cause” for discharge? Who pays the claims, from what pool of funds? It can be hard to know just how much may be omitted while creating an ERISA plan. See, e.g., Fort Halifax Packing Co. v. Coyne, 482 U.S. 1, 107 S.Ct. 2211, 96 L.Ed.2d 1 (1987); Bowles v. Quantum Chemical Co., 266 F.3d 622, 631-33 (7th Cir.2001); Collins v. Ralston Purina Co., 147 F.3d 592, 595-97 (7th Cir.1998); Donovan v. Dillingham, 688 F.2d 1367 (11th Cir.1982) (en banc).
We need not decide whether it is proper to treat the agreement as a stand-alone plan, because it does not stand alone. It was implemented after the acquisition by an eight-page plan providing the essential details. That Rupert seeks to recover under a plan (the one he sees in the acquisition agreement) implies to us that he has not waived his claim under ERISA, even if his papers in the district court forfeited the benefits of the completed plan. The district judge thought it proper to relieve Rupert of that forfeiture, and although we have explained that the court erred by doing this without notice to Pactiv, the court did not abuse its discretion by electing not to enforce the forfeiture. After all, Pactiv itself asked for a decision on the issue (this was what it sought a declaratory judgment about).
[1004]*1004Under the circumstances, the best course is a remand so that the district court can give the required notice and then entertain arguments based on the language of the post-acquisition plan. .It will be necessary to consider the contention that the agreement is itself a plan only if the district judge decides in Pactiv’s favor on the post-acquisition plan.
Vacated and Remanded