This appeal represents a small part of the massive litigation generated by the 1989 Exxon Valdez oil spill into the waters of Prince William Sound, Alaska. The dispute we consider here arises from the punitive damages claims filed against Exxon2 by private parties injured by the spill and consolidated into a single mandatory class action in federal court. Aligned on one side in this appeal are Exxon and a group of plaintiff seafood processors known as the Seattle Seven. The Seattle Seven reached a $64 million settlement agreement with Exxon in the immediate aftermath of the Valdez spill. On the other side are the remaining class plaintiffs, referred to in this opinion as “plaintiffs.”
The critical factual element is the settlement agreement between Exxon and the Seattle Seven. The Seattle Seven, who process seafood caught in Prince William Sound, sued Exxon for compensatory and punitive damages after the spill forced their operations to shut down for significant periods of time. The settlement agreement they reached with Exxon did not include a release and therefore did not formally terminate the Seattle Seven’s claims against Exxon. The Seattle Seven agreed, however, that they would not execute on any compensatory damages award entered in their favor and also would pay or “cede” back to Exxon any punitive damages they might recover. The agreement was subsequently modified to permit the Seattle Seven to retain a portion of the punitive damages award received.
Although both the district court and the plaintiffs knew that there had been a settlement agreement between Exxon and the Seattle Seven, neither knew of the existence of the cede back provision. Acting in its own best interest, Exxon chose not to inform the punitive damages jury either. On September 16, 1994, the jury assessed punitive damages against Exxon in the amount of $5 billion. The plan of allocation the plaintiffs eventually proposed for this a-ward, and that the district court approved, did not include the Seattle Seven.
The central issue for us to decide is whether the jury should have been told of the cede back provision during the last phase of the punitive damages trial. The district court, agreeing with the class plaintiffs, held that Exxon’s failure to affirmatively disclose this information to the jury merited exclusion of the Seattle Seven *793from the plan of allocation. Exxon and the Seattle Seven appeal this ruling.
Exxon’s liability for any punitive damages, and the amount of punitives the jury imposed are challenged in related appeals. We here assume without deciding, for purposes of this appeal, the validity of the judgment against Exxon. We do not intimate what the result of that appeal will be.
BACKGROUND
The oil tanker Exxon Valdez ran aground on the Bligh Reef in Prince William Sound, Alaska on the evening of March 23, 1989. Damage to the Valdez’s cargo holds caused it to spill 11 million gallons of oil into the Sound, resulting in a great environmental disaster. The spill grievously injured both the environment and the economic livelihood of those individuals who relied on the theretofore abundant marine life of the region for their livelihood.
The State of Alaska and the United States brought actions against Exxon for the injury to the environment. Those cases were resolved by entry of a consent decree on October 8,1991, under the terms of which Exxon agreed to pay at least $900 million to restore damaged natural resources. See Eyak Native Village v. Exxon Corp., 25 F.3d 773, 775 (9th Cir.1994).
The hundreds of private civil actions filed in federal court were consolidated before Judge H. Russel Holland of the District of Alaska. First the plaintiffs, and then Exxon moved the district court to certify a mandatory punitive damages class. Judge Holland granted Exxon’s motion on April 19, 1994. Alaska’s state courts agreed to recognize the class action as the only avenue through which any plaintiff, whether in state or federal court, could recover punitive damages from Exxon. See Chenega Corp. v. Exxon Corp., 991 P.2d 769, 775 (Alaska 1999).
The Seattle Seven, the largest of the region’s seafood processors, sued Exxon in 1989. Exxon sought to reach a settlement as quickly as possible, but its negotiations with the Seattle Seven and other plaintiffs revealed a roadblock posed by the increasing likelihood that a mandatory punitive damages class would be certified. Claims for compensatory damages could be easily disposed of by exchanging payment for releases, but a plaintiffs release of its slice of the future lump-sum punitive damages award merely reduced the number of claimants sharing the punitive damages pie, not the size of the pie itself. Exxon thus actually faced a financial disincentive to settle, because any amount of money it paid to persuade a plaintiff to forgo its slice would nevertheless be included in the amount of the eventual award.
On January 8, 1991, the Seattle Seven and Exxon settled the Seattle Seven’s claims for the 1989 and 1990 fishing seasons in exchange for a payment of $63.75 million. To avoid the punitive damages dilemma, the parties included in the agreement a “cede back” provision. The provision stated that the Seattle Seven would not release their punitive damages claims against Exxon but would instead remain parties to the litigation in order to receive their share of an eventual punitive damages award, which they would then cede back to Exxon. The existence of a settlement agreement was made known to the rest of the subsequent punitive damages class, but its terms were kept confidential.
The mandatory punitive damages class action was tried to a jury in three phases in 1994. The first determined that Captain Joseph Hazelwood’s behavior had been reckless, a necessary prerequisite for an award of punitive damages. The second phase assessed the amount of compensatory damages attributable to the spill to give the jury guidance in fixing the appropriate amount of punitive damages. For purposes of this appeal, we need not question the determinations during those phases. The third phase fixed the amount of punitive damages.
Before the third phase began, the parties entered into an Impact Stipulation. *794This described the harm the Valdez spill had caused private parties and quantified part of it by referring to the total amount already paid by Exxon to private parties in compensation (approximately $300 million). This figure included the approximately $64 million paid to the Seattle Seven under the 1991 settlement agreement.
In the third phase of the punitive damages proceedings, the plaintiffs emphasized to the jury the magnitude of the harm and the resulting need for punishment and deterrence. Exxon, for its part, sought to demonstrate that it had already accepted corporate responsibility by pointing to the fact that in many cases, it had paid money to injured parties without requiring anything in return but a receipt and without requiring releases. Exxon’s president testified that Exxon had paid “over $300 million” receiving only receipts in return, and thus, that it had received nothing of value in return for its payments. Exxon’s counsel reiterated this in his closing argument. In fact, however, because the $300 million figure included amounts paid to parties such as the Seattle Seven, who did agree to settle their claims, these statements were inaccurate. The amount paid in return for nothing but receipts was actually somewhere around $168 million.
Exxon’s apparent strategy to maximize to the jury what Exxon had already paid in order to minimize punitive damages did not work well. On September 16, 1994, the jury awarded punitive damages in the sum of $5 billion, at that time the largest award of its kind in history.
The next step was to allocate those damages among the plaintiffs in a manner proportionate to their injury. The original plan of allocation, drawn up by the non-settling plaintiffs, did not include the Seattle Seven because the Seven’s lack of a financial interest in the recovery meant that they also lacked motivation to pursue a stake in the award. In order to create a financial incentive for the Seattle Seven, in 1996 Exxon negotiated a modification to the 1991 settlement agreement with the Seattle Seven. The modification permitted the Seattle Seven to retain $12.4 million of their punitive damage allocation rather than ceding it all back to Exxon.
The Seattle Seven then filed an objection to the proposed allocation with the district court, contesting their exclusion from the plan. At this point, the reason for the requested modification, the cede back provision, became known both to the district court and to the plaintiffs. The plaintiffs began vigorously to oppose inclusion of the Seattle Seven in the plan of allocation.
The district court agreed with the plaintiffs, originally taking the position that the cede back provision itself was unlawful as against public policy. The Seattle Seven and Exxon moved for reconsideration, supporting their motion with declarations of numerous legal luminaries, including former U.S. Attorneys General, judges of various U.S. Courts of Appeal, law professors, and an Alaska Supreme Court Justice, all to the effect that cede back agreements are ethical, enforceable, and necessary for the orderly administration of justice in mass tort cases.
Upon reconsideration, the district court agreed that the cede back agreement was not in and of itself unethical, but held that the Seattle Seven were nonetheless barred from participating in the allocation of damages because the jury was not told of the agreement’s existence. The court’s order stated that the problem was Exxon’s failure to tell the jury “the whole story” regarding the agreements. The court emphasized its belief that the jury should have been entitled to determine how much Exxon should actually pay in punitive damages, out of its own pocket, stating: “Punitive damages are imposed to punish the conduct which juries determine to be reckless. The court has no doubt that the Exxon Valdez jury would be outraged if Exxon, through the Seattle Seven settlement agreement, rather than the claim*795ants, were to wind up with almost 15% of the punitive damages award.”
In this appeal, appellants Seattle Seven and Exxon contend both that cede back agreements are lawful and that for their proper administration, they must not be disclosed to juries. Otherwise, appellants argue, the jury in order to compensate for them or to prevent the defendant from paying less than what the jury believes is appropriate punishment will inflate the punitive damages award.
The appellee plaintiffs defend the district court’s reasoning, arguing that such agreements are unethical and unenforceable unless juries are told of them. They also contend that even if juries should ordinarily not be told, disclosure in this particular case was warranted by Exxon’s exaggerated statements to the jury regarding the amount paid to claimants without releases in return. We review approval of the plan of allocation for abuse of discretion and any necessary legal questions de novo. See In re Mego Financial Corp. Securities Litigation, 213 F.3d 454, 460 (9th Cir.2000).
There are accordingly three principal issues that we must consider in the disposition of this appeal: (1) the lawfulness and enforceability of cede back agreements like the one in this case; (2) whether, if lawful and enforceable, they should generally, as a matter of law, be kept from the jury; and (3) if they should ordinarily be kept secret, whether there were circumstances present in this case that should have required Exxon to tell the jury about the existence of this particular agreement. We hold that cede back agreements are enforceable; that in accordance with the general principle that indemnification arrangements should not be allowed to affect a jury’s determination of damages, cede back agreements should not be disclosed to the jury; and, finally, that there are no circumstances in this case that would have warranted disclosing the terms of this cede back provision to the jury. We therefore conclude that the district court abused its discretion in approving a plan of allocation that denied enforceability of the settlement agreement between Exxon and the Seattle Seven and that barred the Seattle Seven from receiving any allocation of punitive damages.
I. ENFORCEABILITY OF THE CEDE BACK PROVISION
In recent years, federal courts have become all too familiar with the peculiar problems posed by mass tort litigation. See, for example, Amchem Products, Inc. v. Windsor, 521 U.S. 591, 117 S.Ct. 2231, 138 L.Ed.2d 689 (1997) (asbestos); Valentino v. Carter-Wallace, Inc., 97 F.3d 1227 (9th Cir.1996) (epilepsy medication); In re Agent Orange Product Liability Litigation, 818 F.2d 145 (2d Cir.1987). Such litigation clogs dockets for decades, creating burdens on the judicial system and delaying relief for injured parties. As a result, the general policy of federal courts to promote settlement before trial is even stronger in the context of large-scale class actions. See Franklin v. Kaypro Corp., 884 F.2d 1222, 1229 (9th Cir.1989) (stating that the fact that “there is an overriding public interest in settling and quieting litigation ... is particularly true in class action suits.”). It is unfortunately also true, however, that such settlements are difficult to reach. “[Ojbtaining a settlement in multi-party litigation may be quite complex.” Id. at 1225. It will frequently be very close to impossible for a mass tort defendant to achieve a settlement with every potential plaintiff. The resulting presence of non-settling defendants, non-settling plaintiffs, or both, may seriously affect the parties’ incentives to settle in the first place.
In addition to encouraging individual settlements, combs have encouraged the use of mandatory class actions to handle punitive damages claims in mass tort cases. Mandatory class actions avoid the unfairness that results when a few plaintiffs-those who win the race to the court*796house-bankrupt a defendant early in the litigation process. They also avoid the possible unfairness of punishing a defendant over and over again for the same tortious conduct. As a result, mandatory classes have been endorsed by many courts and commentators. See In re A.H. Robins Co., Inc., 880 F.2d 709, 738 (4th Cir.1989) (recognizing that “the ‘trend’ of the authorities is clearly in the direction of a more liberal approach to the certification of the mass tort action”). See also II The American Law Institute, Enterprise Responsibility for Personal Injury 263 (Reporters’ Study 1991) (recommending federal legislation to create nationwide mandatory punitive damages classes); Richard A. Seltzer, Punitive Damages in Mass ToR Litigation: Addressing the Problems of Fairness, Efficiency and Control, 52 Fordham L.Rev. 37, 61 (1983) (arguing that only class actions provide a practical means for resolving the problems that accompany punitive damage awards in mass tort litigation).
One drawback to the mandatory class action, however, is that it makes it even more difficult to settle the claims of any individual plaintiff. Because punitive damages in a mandatory class action are awarded in one lump sum, a defendant has a serious disincentive to settle with any plaintiff unless it can negotiate a settlement with them all, a staggering feat if not a practical impossibility. Partial settlement merely reduces the number of plaintiffs who share an eventual award. It does not reduce the award’s amount. Because a defendant like Exxon would presumably be indifferent as to whether it paid 10,000 plaintiffs $500,000 each or 500,000 plaintiffs $10,000 each, the creation of mandatory punitive damages classes cuts against the strong judicial policy of encouraging settlement in class actions.
We deal here with multiple plaintiffs suing one defendant, but an analogous problem frequently occurs in the more typical situation of a single plaintiff with claims against multiple defendants. When a plaintiff is able to settle with fewer than all of the defendants, the question becomes how to determine what share of a jury’s total assessment of damages a non-settling defendant should pay. Courts agree that the non-settling defendant does not have to pay the entirety of any eventual damages award. They diverge, however, on the issue of apportionment, taking three distinct approaches. See McDermott, Inc. v. AmClyde, 511 U.S. 202, 215-17, 114 S.Ct. 1461, 128 L.Ed.2d 148 (1994) (explaining the three approaches).
Under the first approach, the non-settling defendant pays the entire amount of the award less the actual amount the plaintiff has already received from the settling defendant, even if this total turns out to be in excess of the non-settling defendant’s, share of the fault as determined by the jury. The non-settling defendant then retains the right to seek contribution from the settling defendant in order to bring total payments in line with allocation of fault. This is called the “pro tanto with contribution” approach, and it creates little incentive for any defendant to settle.
The second approach is known as “pro tanto without contribution.” Under this approach, the non-settling defendant pays the entire amount of the award less the amount of the settlement and does not retain the right to seek contribution. This helps ensure that the plaintiff receives the full amount of damages and maintains incentives to settle, but can result in the non-settling defendant paying more than its share of fault.
Finally, under the “proportionate share” approach, the non-settling defendant pays only the amount of the award that is allo-cable to its share of the fault, as determined by the jury. The proportionate share approach is the law in the Ninth Circuit, has been adopted by the Supreme Court for use in maritime actions, and is the approach recommended by the American Law Institute. See Kaypro, 884 F.2d at 1231. See also McDermott, 511 U.S. at *797217, 114 S.Ct. 1461; Restatement (Third) of Torts: Apportionment of Liability § 16.
The main advantage of the proportionate share approach is that it is the only one of the three that combines fairness to all parties with an appropriate balance of individual incentives to settle. The effect of proportionate share apportionment, however, is that the actual amount of damages the plaintiff receives will deviate from the amount awarded by the jury, unless the amount of the settlement exactly matches the settling defendant’s share of fault as subsequently determined by the jury. If the jury later determines that the settling defendant’s share of fault is less than the amount paid in settlement, this will result in a windfall to the plaintiff. If the jury’s allocation is higher, this will result in a shortfall.
This case differs from the typical situation in that we do not have a single plaintiff seeking to recover a single award from multiple defendants. The Exxon Valdez punitive damages class involves multiple plaintiffs seeking to recover a single award from a single defendant. Neither our court nor leading authorities have addressed this situation. The potential distortion of settlement incentives that occurs when some parties settle and some do not is the same, however, as with the multiple defendant situation. If Exxon could have been sure that the district court would eventually adopt a form of the proportionate share approach, permitting non-settling plaintiffs to recover damages only in proportion to their allocation of harm and allowing the remaining punitives to go uncollected, settlement incentives would have been preserved and the cede back provision would not have been necessary. Exxon had no such certainty, though (and indeed, the district court eventually refused to adopt such a method). Exxon therefore sought to achieve a proportionate share result without the court’s assistance by adding the cede back provision to its settlement agreement with the Seattle Seven.
An analogous type of cede back agreement has been used in the multiple-defendant context and is called a Pierringer release, after the leading case to consider it, Pierringer v. Hoger, 21 Wis.2d 182, 124 N.W.2d 106 (Wis.1963). Pierringer releases have been approved in Wisconsin and Minnesota. See id.; Frey v. Snelgrove, 269 N.W.2d 918, 922 (Minn.1978).
Pierringer releases have the effect of reaching a proportionate share result and have principally been used in jurisdictions that adhere to the “pro tanto with contribution” approach. A plaintiff who settles with one of multiple defendants agrees to indemnify that defendant for any eventual contribution action brought by the non-settling defendants after the entry of judgr ment. See Peter B. Knapp, Keeping the Pierringer Promise: Fair Settlements and Fair Trials, 20 Wm. Mitchell L.Rev. 1 (1994).
The Supreme Court has recognized that Pierringer releases mitigate the adverse effect on settlement that exists under a “pro tanto with contribution” regime, but has worried that the prospect of indemnity actions might add “yet another potential burden on the courts.” McDermott, 511 U.S. at 212, 114 S.Ct. 1461. As a practical matter, however, Pierringer releases do not require the actual litigation of a contribution action, followed by an indemnity action. The portion of the judgment allo-cable to the settling defendant’s fault is simply considered uncollectible, just as it would be under the proportionate share approach. See Austin v. Raymark Ind., Inc., 841 F.2d 1184, 1190 (1st Cir.1988) (stating that through a Pierringer release, a non-settling defendant “effectively obtain[s] its contribution from the settling defendants by having assessed against it only its own percentage of liability”). Similarly, in the multiple plaintiff context, a non-settling plaintiff through a cede back agreement obtains only its proportionate share of the entire punitive damages award.
*798What Exxon and the Seattle Seven did, therefore, was use a Piemnger device to obtain the functional equivalent of a proportionate share allocation of damages. Since both the Ninth Circuit and the Supreme Court have endorsed the proportionate share approach because of its superiority in blending fairness to the parties with incentives to settle, we cannot hold such an agreement unenforceable as a matter of public policy. Far from being unethical, cede back agreements make it easier to administer mandatory class actions for the assessment of punitive damages and encourage settlement in mass tort cases. As a result, such agreements should typically be enforced.
II. WHETHER CEDE BACK AGREEMENTS SHOULD GENERALLY BE DISCLOSED TO THE JURY
The district court in this case held that cede back agreements, though ethical, should be disclosed to the jury because the jury should be able to take them into account in assessing punitive damages. Exxon and the Seattle Seven contend, however, that juries should never be told. They argue persuasively that the salutary purposes of such agreements would be frustrated if the jury knew about the terms of the agreement and were permitted to offset them by increasing damages. We agree.
If a jury was told that the defendant would eventually get back a portion of the punitive damages assessed, the jury would likely compensate by imposing more damages, thereby assuring that the defendant would pay the entire amount deemed appropriate by the jury. This is exactly what the district court believed the jury should have been permitted to do in this case. Yet if that were to be the result of the settlement agreement, from the defendant’s perspective there would be no point in settling in the first place. The defendant would still have to pay the full amount assessed by the jury, in addition to the amount paid in settlement.
This is what we have recognized elsewhere in our law: that a jury should assess damages but not determine how much defendants should “actually” pay or how much plaintiffs should “actually” receive. In Larez v. Holcomb, 16 F.3d 1513 (9th Cir.1994), we held that it was prejudicial error to inform a jury deliberating on an award of punitive damages that the defendant would be indemnified by his employer for any such award. See id. at 1520-21. We stated that a jury’s task is to arrive at a “dispassionate” determination of the proper award, and that evidence of indemnification might have tempted the jury to inflate the award out of sympathy for the plaintiff. See id. at 1519. Similarly, we held in Brooks v. Cook, 938 F.2d 1048 (9th Cir.1991), that juries should not be told about the availability of attorneys’ fees when fixing an award for a prevailing plaintiff. We explained that “the fear is that a jury, informed of plaintiffs right to additional funds, will view the money as a windfall and take steps to offset it.” Id. at 1052.
These principles are not unique to this circuit. It is uniformly held that absent exceptional circumstances, a jury deliberating on the amount of a damages award is not to consider where the funds that constitute that award will come from, or where they will end up. For example, the states of Georgia and Oregon both have enacted tort reform measures that provide that large portions of punitive damages awards (75 percent in Georgia and 50 percent in Oregon) go to the state or state-designated charities rather than the prevailing plaintiff. Both states have held that it is prejudicial error to inform the jury of this ultimate outcome, because of the temptation for the jury to inflate the award in order to more fully compensate the plaintiff. See Ford v. Uniroyal Goodrich Tire Co., 267 Ga. 226, 476 S.E.2d 565, 570-71 (Ga.1996); Honeywell v. Sterling Furniture Co., 310 Or. 206, 797 P.2d 1019 (Or.1990). Similarly, the Supreme Court of Minnesota has held that juries should *799not be informed that a finding of a certain percentage of comparative negligence on the part of the plaintiff will serve to reduce the damages award by that percentage. See Rosenthal v. Kolars, 304 Minn. 378, 231 N.W.2d 285, 288 (Minn.1975). Juries are also not to be told of statutory caps on damages, or, in antitrust and RICO cases, that damages will eventually be trebled. See Sasaki v. Class, 92 F.3d 232, 237 (4th Cir.1996) (holding that informing a jury of a statutory cap on damages for one of plaintiffs claims may have led to an increased award on the non-capped claim); HBE Leasing Corp. v. Frank, 22 F.3d 41, 45-46 (2d Cir.1994) (holding that informing a jury of RICO’s treble damage provision may confuse or prejudice jury into lowering the award in order to counteract the trebling effect); Pollock & Riley, Inc. v. Pearl Brewing Co., 498 F.2d 1240, 1242-43 (5th Cir.1974) (treble damages); Semke v. Enid Automobile Dealers Ass’n, 456 F.2d 1361, 1370 (10th Cir.1972) (treble damages); Weiss v. Goldfarb, 154 N.J. 468, 480, 713 A.2d 427 (N.J.1998) (holding that informing a jury of a statutory cap on damages that applied to one defendant may have led the jury to increase the amount of damages allocable to other, non-capped defendants). The message from these cases is the same: juries are to be kept free of any outside influence that might lead them to inflate or reduce their damages award in order to “secure justice” for the parties. See Rosenthal, 231 N.W.2d at 288.
In cases involving settling and non-settling defendants, several states provide by statute that in allocating fault, the jury cannot be told of a settlement, or its tenns. See Conn. Gen.Stat. § 52-216a; Fla. Stat. ch. 768.041; Maine Rev. Stat. Ann. § 163; N.H.Rev.Stat. Ann. § 507:7 — i; N.Y. C.P.L.R. 4533-b (McKinney). The reason for this is the danger that the jury will adjust its award of damages according to the amount of the settlement. See Builder’s Square, Inc. v. Shaw, 755 So.2d 721, 725 (Fla.Ct.App.1999). If the jury perceives that the settlement amount is low, for example, it might be tempted to increase the amount of fault allocable to the non-settling defendants in order to maximize the plaintiffs recovery.
In Minnesota, a state that has recognized the validity of agreements similar to the cede back provision in this case, the Supreme Court has held that the existence of a Pierringer release may be admissible for limited purposes such as to show witness bias. The court must exercise its discretion, however, as to what details of the agreement should be provided to the jury, and “as a general rule the amount paid in settlement should never be submitted.” Frey v. Snelgrove, 269 N.W.2d 918, 923 (Minn.1978). Thus, the existence of an agreement between the parties is clearly relevant should the settling defendant attempt to testify at trial in a manner favorable to the plaintiff. The details of the agreement should not be disclosed, though, and the evidence admitted should be limited to avoid distorting the jury’s deliberations on damages. See id. See also 23 Wright & Graham, Federal Practice and Procedure § 5311 (commenting that when evidence of a settlement agreement is admissible to show witness bias, “it is true that often the fact of a compromise will suffice to show bias and that exploring the details will increase the prejudice to the opposing party without any appreciable effect on the credibility of the witness”); Restatement (Third) of Torts: Apportionment of Liability § 24, Reporters’ Note, comment i (recommending that when the fact of a settlement agreement is admissible to show witness bias, the agreement itself should not be admitted into evidence, and that if it is, potentially prejudicial portions of it should be redacted).
Therefore, it is clear that cede back agreements should generally not be revealed to juries deliberating on punitive damages. The only remaining question is whether any particular circumstances in this case warranted an exception from this general rule.
*800III. SPECIAL CIRCUMSTANCES
The plaintiffs argue that even if evidence of a cede back agreement would ordinarily be kept from a jury, Exxon should have volunteered it in this case to correct the false impression Exxon created when it announced that it had paid out $300 million in compensation requiring only receipts in return when the correct figure was approximately $168 million. They further argue that the fact that the exaggeration went unconnected justifies the district court’s decision to exclude the Seattle Seven from the plan of allocation and render the cede back provision completely unenforceable.
The plaintiffs rely on Lawson v. Trow-bridge, 153 F.3d 368 (7th Cir.1998). Lawson holds that otherwise inadmissible evidence of indemnification may be admitted on cross-examination to impeach a testifying defendant who intimates to the jury that he will be financially ruined by a large damages award. See id. at 379. Lawson does not concern the enforceability of the underlying indemnification agreement.
We will not extend Lawson beyond its holding to justify the district court’s exclusion of the Seattle Seven from the plan of allocation on the basis of Exxon’s trial conduct. Denying the Seattle Seven recovery on the basis of Exxon’s actions would be manifestly unfair.
As for Exxon itself, we do not condone its conduct. The exaggerations it made at trial, however, have little, if anything, to do with the cede back provision. Exxon stated that it had paid out $300 million requiring only receipts in return; the only correction necessary was the alteration of that figure to $168 million. There would have been no need to mention the cede back provision, and thus no need to disturb the general rule that such agreements should be ordinarily be kept from the jury. Furthermore, the only conceivable prejudice the plaintiffs could have suffered when Exxon overstated its corporate benevolence would have been a downward adjustment by the jury of the damages award, and the proper remedy for such prejudice would be a new trial on punitive damages. The plaintiffs have never claimed that $5 billion was too low an award, however, nor have they ever sought a new trial. Refusing to enforce the cede back provision is not a remedy that relates to the error complained of.
IY. OTHER ARGUMENTS
The plaintiffs raise on appeal two alternative arguments in support of the district court’s decision to exclude the Seattle Seven from the plan of allocation. First, the plaintiffs claim that because the Seattle Seven “settled” their claims against Exxon, they have no right to claim a share of punitive damages regardless of the existence of the cede back provision. It is clear, however, that the settlement agreement did not contain a release of any claims; it merely “settled” the matter of what would happen to the claims once the claims were finally adjudicated. Second, the plaintiffs argue that public policy precludes any agreement that diminishes the deterrent effect of a punitive damages award. This argument runs contrary to the law of this circuit, which permits indemnification agreements and the settlement and release of punitive damages claims. As exemplified by Larez, it is not up to the jury to decide how much a defendant must actually pay at the end of the day, or how much a plaintiff will actually receive. We therefore reject both of the plaintiffs’ alternative arguments.
CONCLUSION
We hold that the district court abused its discretion in approving the plan of allocation over the Seattle Seven’s objection. Cede back agreements are lawful and enforceable, and generally should not be disclosed to the jury. No special circumstances in this case justify the district court’s refusal to enforce the cede back agreement between Exxon and the Seattle Seven. As a result, the existence of the *801cede back agreement cannot justify exclusion of the Seattle Seven from the plan of allocation.
The approval of the Allocation Plan is VACATED and the matter is REMANDED.
. We follow the practice of other panels of this court who have decided cases involving the Exxon Valdez, and use "Exxon” to refer to Exxon Corporation, Exxon Shipping Company, Exxon Transportation Company, and any other related entity. See Eyak Native Village v. Exxon Corp., 25 F.3d 773, 774 n. 1 (9th Cir.1994).