by Dennis Crouch
I like to think of collateral estoppel alongside binding precedent. Both involve a set of (at least) two separate cases, and the question is always whether some conclusion drawn in the earlier action forecloses relitigation of same (or similar) issue in the latter action. Precedent is broader because it applies across-the-board regardless of the parties involved. Collateral estoppel (also known as issue preclusion) is narrower because it only applies when the party being estopped in the latter action was also a party in the earlier action. But, when that same-party requirement is met, collateral estoppel can be much more powerful.
Most importantly, collateral estoppel has the magical quality of binding both horizontally and upwardly. This means that issues decided at the district or administrative court level can be binding on all other courts: district courts, administrative courts, appellate courts, and even the Supreme Court. On the other hand, precedent feels the gravitational pull and flows downhill. District court decisions are not binding precedent because they are at the bottom. The Supreme Court precedent flows down to bind appellate courts whose precedent binds their assigned district courts. Although both the Supreme and Appellate Courts respect their own prior precedent, it is not binding in the same sense as those courts are also self-empowered to modify their own precedent. Another important feature of collateral estoppel is that it applies to both issues of fact and law. Although some courts have disagreed on this point, precedent is only binding for questions of law.
The Federal Circuit’s recent Uniloc preclusion decision in a tough one and exemplifies an important feature of collateral estoppel power: it is most strident when the original court decision is wrongly decided, but still binds later courts (including later appellate courts). Uniloc USA, Inc. v. Motorola Mobility LLC, 52 F.4th 1340 (Fed. Cir. 2022). Even though the original lower court decision violated precedent, it still binds the parties.
Some background and how this Played out for Uniloc: HP sold several patents to Uniloc back in 2017 who then sued Apple, Motorola, and Blackboard for patent infringement. Uniloc has a litigation financing relationship with Fortress with the patents serving as collateral for the deal. The terms of the Fortress financing became important for the subsequent litigation. Basically, if Uniloc failed to meet certain revenue goals, Fortress would be given a license to the patents (including a right to sublicense). Uniloc then failed to meet the goal, and, by the terms of the deal, those license rights arguably automatically passed to Fortress.
After some venue-action, Uniloc v. Apple ended up before Judge Alsup in the Northern District of California. Apple’s attorneys were able to convince the judge that the license to Fortress meant that Uniloc no longer held full rights to the patent and thus lacked standing to sue. In my view, Judge Alsup’s conclusion is wrong on the merits. A patentee who has non-exclusively licensed its patents still has standing to sue. Judge Lourie detailed the law on this in his “additional-views” while the lead opinion simply stated that “there is considerable force to Uniloc’s argument” on the merits. Uniloc v. Motorola, 52 F.4th at 1350 (Fed. Cir. 2022). Uniloc appealed the Apple decision and (again, in my view) would have likely won on appeal. . . BUT, the parties reached a settlement in the midst of appellate briefing and asked the appellate court to dismiss the appeal. Uniloc did not seek vacatur of the lower court decision as part of the settlement. So, at the end of the day, Uniloc v. Apple ended with district court judgment remaining in force, including the decision that the Fortress license left Uniloc without standing.
The two other cases, Uniloc v. Motorola and Uniloc v. Broadcom, were still pending in D.Del on December 4, 2020 when Judge Alsup issued his dismissal order of the Apple case. Two days later, on December 6, Motorola moved to dismiss on the same ground of lack-of-standing due to the license. Oddly, in its briefing Motorola indicated that Apple decision was “not binding” but did provide important authority. I expect that Motorola’s “not binding” remark was an indication of its precedential value rather than its preclusive impact. Judge Connolly dismissed the Motorola case, on standing grounds (and without referring to collateral estoppel). Acting sua sponte, Judge Connolly then dismissed the Broadcom in a short opinion that might have been based upon issue preclusion (from the prior Motorola decision), but without actually stating as-such and without Broadcom making that claim. One bottom line here, although the parties and the court were all aware of the prior Apple decision, nobody suggested at the district court level that it had issue preclusive effect.
Uniloc appealed the dismissals — arguing that the two decisions were wrongly interpreted the Constitutional requirements of standing with regard to patent infringement actions. At that point, the two defendants for the first time raised the prospect that Apple had preclusive effect.
Forfeiture/Waiver: From the setup, you should be thinking forfeiture or Waiver. Ordinarily a new defense cannot be raised for the first time on appeal. Rather, a party planning to appeal some issue should first timely raise the issue to the district court. This rule of appellate procedure applies to the res judicata doctrines of issue and claim precl. Arizona v. California, 530 U.S. 392 (2000) (“res judicata [is] an affirmative defense ordinarily lost if not timely raised”). Here though, we also have the express statement from Motorola that the Apple case is “not binding.” That appears to create waiver.
On appeal, the Federal Circuit excused the delay by Motorola and Broadcom and found no forfeiture or waiver. In particular, the court noted that, although the Apple decision was final, it wasn’t final-final. For issue preclusion to attach, the original action must be “final,” which generally means that the district court reached final judgment on all issues in the case. But, the Federal Circuit here held that there is no forfeiture unless the first case is what I call final-final–not only reached final judgment, but also exhausted all direct appeals. That extra step is important for this case because the Apple appeal was still pending as Motorola/Broadcom district court was deciding those cases. The Federal Circuit noted “strong policy reasons supporting a finding of no forfeiture of collateral estoppel when the argument was first raised after the appeal process of the preclusive case was concluded, and we conclude that forfeiture is not appropriate here.”
As to waiver based upon the “not binding” statement from Motorola, the Federal Circuit concluded that the intent of that statement was with reference to stare decisis and binding precedent rather than binding as a matter of preclusion.
Finally, the court also issued a catch-all statement that–even if there was an unexcused delay, the court has discretion to excuse the forfeiture. “[W]e have discretion to excuse any forfeiture.” Id. That claim by the court seems to be an overstatement since an appellate court excusing forfeiture or waiver should consider the equities of the situation before moving forward.
Merits of Issue Preclusion: On the merits, it seems to be a fairly open-and-shut case of issue preclusion:
- Same Issue: The issue of the license and the resulting no-standing decision was litigated decided in the first case (Apple); and that same issue is now at issue in the second case (Motorola).
- Actually Litigated: The issue was actually litigated in Apple.
- Subject to a Valid and Final Judgment: The Apple district court properly entered final judgment (and it was not reversed on appeal).
- Essential: The license/standing issue was essential to the Apple judgment.
Traditional collateral estoppel had a same-parties rule, parties in first case be the same as those in the second case. In its 1979 Parklane Hosiery decision, the Supreme Court opened the door to “nonmutual issue preclusion” that allows non-parties to assert collateral estoppel. The new rule is a same-party rule and is singular rather than plural; result being that issue preclusion can apply so long as the party against whom nonmutual issue preclusion is being invoked was a party to the previous proceeding that already decided the same issue. Here, Uniloc was a party to the first case, and the estoppel is being applied against Uniloc in the second case–so this fits the same-party rule. Courts distinguish between nonmutual defensive and nonmutual offensive issue preclusion; being more resistant to applying issue preclusion in the offensive scenario. Here, the preclusion is defensive in nature–Motorola is using the doctrine to as a defense against Uniloc’s infringement claim. With all this together, the appellate court easily concluded preclusion applied and thus that the defendants should automatically win on the issue of standing.
Two aspects of this case make me uncomfortable. The first part is an underlying feature of issue preclusion — that sometimes the first court reaches the wrong answer, and issue preclusion can still apply. Here, the wrong-answer is amped-up because it is a Constitutional decision stripping the plaintiff of standing. Second, and perhaps more importantly is that the appellate court skips over a core rule that prohibits the application of collateral estoppel to work an injustice. In its opinion, the Federal Circuit repeatedly cited Wright & Miller, but skipped that entire section of the treatise. Here, the seeming error by Uniloc was settling the Apple appeal. Of course, that was a global dispute involving lots of other patents. But, by the time of the Apple settlement, the Delaware court had already issued its decisions in the two later cases and, despite a full opportunity, nothing in any of those cases suggested that the Apple district court decision would have a collateral estoppel impact. It is not clear to me that these considerations are enough to flip the outcome, but it seems that the court should have at least considered the issues.