Elements of Issue Preclusion

Elements of Issue Preclusion

A party who wants an earlier judgment to settle a question in a new lawsuit has to prove that the judgment qualifies. In its fullest modern form, the showing has six parts: the issue is the same one decided in the first lawsuit; the issue was actually litigated there; the first lawsuit ended in a valid final judgment, or one that is final enough; the issue was necessarily decided; the party to be bound was a party, or in privity with a party, to the first lawsuit; and that party had a full and fair opportunity to litigate.¹ Courts vary widely in how they phrase, order, and acknowledge these requirements, although many now cite Restatement (Second) of Judgments §§ 27 and 28.²

The classical formulation is shorter. It asks for the same issue, raised in a prior action that ended in a valid final judgment; actual litigation and determination of that issue; necessity of its resolution to the judgment; and the same parties or their privies.³ The federal system and most states no longer hold completely to the last requirement, a development covered under nonmutual preclusion below.⁴

Two of the elements need little fresh analysis. Whether the parties in the second suit are the same as, or in privity with, those in the first is decided by the principles that govern claim preclusion, and so is whether the first action ended in a valid final judgment.⁵ The work specific to issue preclusion lies in the elements that examine what the first court did with the question. The subsections below take them in the order a lawyer would test them: whether the issue is the same, whether it was actually litigated, whether it was actually decided, whether the decision was essential to the judgment, how alternative grounds are treated, and whether the judgment is valid and final. Who may be bound, and whether that party had a full and fair opportunity to litigate, come later in the chapter.

The proponent carries the burden on every element.⁶ The demands are heavy for a reason. Claim preclusion reaches claims that were asserted and also transactionally related claims that could have been asserted, while issue preclusion reaches only issues that were actually litigated, not those that could have been.⁷ A proponent therefore has to produce the record of the first case and show how its issue, its contest, its ruling, and its judgment match the question now before the court. A court that cannot tell what the first case settled has no basis for binding anyone to it.⁸

Identical Issue

Nothing else in the analysis matters if the second suit asks a different question from the one the first court answered. The issue as to which relitigation is now sought to be barred must be the same as the one decided in the earlier proceeding.⁹ Preclusion substitutes an old answer for a new hearing, and that exchange is fair only when the old answer responds to the question now in dispute. If the controlling facts, the time period, or the governing law has shifted, the earlier proof may not bear on the new dispute.

The clearest case is a driver who runs a red light and injures two people. One of them sues and wins a judgment after a jury trial. Even if the jury returned only a general verdict, the driver's negligence in that accident is conclusively established, because negligence is an element of a personal injury claim and no plaintiff could win without establishing every element. When the other injured person sues the driver, the driver can be barred from relitigating negligence, since the issue is the same in both cases.¹⁰ Whether someone who was not a party to the first suit may invoke the result is the separate question of nonmutual preclusion, taken up later in this chapter.

Identity is tested issue by issue, so the second suit may rest on an entirely different claim. Whether a defendant drove negligently is the same question in a driver's suit for personal injuries and in a later suit by that driver's spouse for loss of consortium, because both claims rest on a single act of the defendant.¹¹ The same reasoning carries an issue across claims that differ in other ways.

The later installments make the claim in the second suit different from the first, but the invalidity of the loan agreement for lack of disclosures is the same issue in both, because it is the same loan agreement. The issue was also actually litigated, since the parties presented evidence on it in the first suit. The borrower therefore cannot raise the defense a second time.¹²

A single judgment can bind on one issue and leave the next open. Suppose Smith sues Jones for cutting wood on her land, and the court upholds Jones's easement. If Jones cuts wood again and Smith sues over the second incident, the new suit is not barred by claim preclusion, because it concerns a different incident. Smith is nonetheless estopped from trying again to prove that the easement is invalid, because the first action litigated and decided that issue.¹³

An issue may be as small as a single fact. If the first trial produced a finding that the defendant lacked a valid driver's license at the time of the accident, whether made by the judge after a bench trial or left undisputed by the defendant, the defendant can be barred from relitigating that fact in a later case brought by the other injured party.¹⁴ In Otherson v. Department of Justice, the common issue was Otherson's misconduct in abusing aliens, which a criminal conviction had resolved against him and which the government invoked in his later administrative proceeding.¹⁵ Courts once drew a line between ultimate facts and merely evidentiary facts in deciding what could be precluded. More recent commentary criticizes that distinction as difficult, if not impossible, to define and often out of line with the time and energy the parties devoted to contesting the disputed fact.¹⁶ The practical lesson is to test a fact against the other elements, asking whether it was actually litigated and decided and whether it mattered to the judgment, instead of sorting it under a label.

A question of law carries over only if the same law governs the second suit. In one illustration, an earlier suit involved a California retraction statute. Plaintiffs who lived in Virginia and suffered most of their damages there might be governed by Virginia law instead, which might extend the retraction defense to magazines as well as newspapers or might not allow the defense at all. The issue may then be different in the second suit, or irrelevant to it.¹⁷

Time is the most common way an issue stops being identical. When actions arise out of acts that occurred at different times, the issues, even if very similar, are not identical. A ruling that an organization did not function as a religious group in 2007 does not decide its taxable status in 2008, because the character of its activities or objectives may have changed in the meantime.¹⁸ A later date does not by itself defeat identity, however. In one illustration, a buyer, Vanderbilt, won a judgment against a coal supplier, Fisk. When Vanderbilt later sued Fisk for a 2015 breach of the same contract and Fisk again raised the defense that AB coal satisfied the specifications, collateral estoppel should bar him from relitigating it.¹⁹

Changed facts defeat preclusion only when they were essential to the first judgment. In Montana v. United States, the Court confronted the argument that changed facts took the later case outside an earlier decision. It agreed that changes in facts essential to a judgment make collateral estoppel inapplicable, but it did not read the earlier decision, Kiewit I, as resting on the factual assumption that the gross receipts tax would cancel out if public contractors took all available refunds and credits.²⁰

A single judgment can also match a later case on some issues and miss it on others. In Pastre v. Weber, a man sued police officers for using excessive force in arresting him, and he had earlier been convicted of resisting those officers. The conviction necessarily decided that the force the officers used before he was in custody was justified. His civil claim also challenged force used after custody began, which was a different issue, so the conviction had only partial preclusive effect.²¹ Because identity is tested one issue at a time, a court facing a multi-issue case sorts it into its component issues, binds the parties on those that match what the first court decided, and tries the rest.

Actually Litigated

Asking whether an issue is the same says nothing about whether the first court ever heard a contest over it. An issue is actually litigated when it is properly raised in the earlier lawsuit, submitted to the court for a determination, and determined, a process that usually includes a determination by summary judgment but not necessarily one by default judgment.²³ The definition describes a sequence: someone must put the point in dispute, someone must ask the court to resolve it, and the court must do so. A point the parties could have raised but did not was never put in dispute, however obvious or available it was.

That is why this element, arguably the key requirement for issue preclusion, also separates it from claim preclusion. It is litigation of the issue, the adversarial submission of evidence, that makes it likely, though not certain, that the issue was correctly decided.²⁴ Issue preclusion also exposes a litigant to binding effect in suits on different claims, and the common law balanced that broader exposure with a requirement claim preclusion lacks: the issue must actually have been litigated and decided.²⁵

Two further reasons support the requirement. A claimant can usually foresee when transactionally related claims against the same party might arise, but an issue such as ownership of property, validity of an instrument, or a family relationship can arise in settings that have nothing to do with the first transaction. In a $2,500 suit over a pedestrian accident, for example, the factfinder might need to decide who owned the Rolls Royce that struck the pedestrian.²⁶ And giving preclusive effect to issues no one litigated would discourage compromise, make it less likely that stipulations would narrow the issues, and so intensify litigation.²⁷

What the parties never raised stays open. P sues D for one installment of rent under a lease and wins. In a later suit for subsequent installments on the same lease, D is not estopped from denying that she ever executed the lease, because execution was not actually litigated and decided in the first action, even though she could have raised it as a defense the first time.²⁸ In the easement example, likewise, Smith may still argue in the second suit that an implied condition of the easement required payment for the timber, because that theory was not raised and decided in the first action.²⁹

A party who ignores a lawsuit gives up less than it might seem. Judgments entered because a party ignored the litigation, such as a default, are treated as fully operative for claim preclusion, because parties should suffer the consequences of ignoring legitimately commenced litigation, which wastes the time and resources of the court and the adverse party.³⁰ Issue preclusion asks for more, because it requires that the issue was actually contested and decided. Suppose one driver wins a default judgment against the other after a collision because the other never answered. In a later suit over the same collision, issue preclusion does not stop the defaulting driver from trying to prove that the first driver was negligent, since the default rested only on the failure to appear and no one ever litigated fault.³¹

Stipulations are the hardest part of the line to draw, because one word covers two different acts. A stipulation to the truth of a fact establishes that fact by the parties' agreement rather than by judicial resolution, so the fact has not been actually litigated and is not a candidate for issue preclusion.³²

Otherson v. Department of Justice shows a stipulation that does something different. After a first trial on felony charges, the government agreed to drop them and retry Otherson only on misdemeanor federal offenses. In return, he agreed to proceed without a jury on a stipulated record: if six government witnesses were recalled, they would testify on direct and cross-examination as they had at the first trial, and he would introduce none of the defendants' testimony or evidence.³³ When the government later relied on the misdemeanor judgment to bar him from contesting the abuse in the administrative proceeding, he argued that the issues had never been actually litigated. The argument failed because he had not stipulated to the truth of the government's allegations. He had stipulated only to what the witnesses would say, which helped shape the record the factfinder would use and established no fact.³⁴ His lawyer cross-examined the witnesses, and the trial judge evaluated their testimony and decided whether the government proved its case beyond a reasonable doubt, so the facts were contested and decided.³⁵ The test therefore asks what the stipulation does. A stipulation that concedes a fact removes it from actual litigation, while one that only identifies the evidence leaves the contest, and the preclusive effect of its outcome, intact.

Judgments entered by agreement follow the same logic. A consent judgment records the parties' bargain rather than a judicial decision on a contested issue, so it is not a product of actual litigation.³⁶ In Baker v. General Motors Corp., a discharged employee, Elwell, and General Motors submitted an agreed injunction that the Michigan judge signed. The Supreme Court observed that no issue was joined, expressly litigated, and determined in that proceeding, yet the order was still claim preclusive between them: Elwell's wrongful-discharge and related claims had merged in the judgment, and he could not sue again to recover more.³⁷ The same agreed order can therefore end a claim and still decide no issue. Settlement behaves the same way. In most jurisdictions a settlement by consent has no collateral estoppel effect, although the parties may provide otherwise.³⁸ A driver who settled with one injured person through a $2,000 consent judgment may therefore deny negligence when another person hurt in the same collision sues, because the judgment reflects only the parties' agreement.³⁹

Actual litigation does not require a full trial. If the court holds, on a summary judgment motion or after a trial of the issues, that a defendant was negligent, that finding can carry issue preclusion into a later case if the other requirements are met. Without such a finding, issue preclusion should not operate on the point. In practice, deciding what counts as actual litigation and determination can raise difficult questions.⁴⁰ The proponent proves the point from the record of the first case. When Vanderbilt later wants to bar Fisk's repeated defense, he would show the second court the pleadings that raised it, the instructions that submitted it, and the verdict and judgment that rejected it.⁴¹

Actually Decided

A contested issue is not necessarily a decided one. Restatement § 27 gives conclusive effect only to an issue that was actually litigated and determined by a valid and final judgment, so the party asserting preclusion must also show that the first court in fact resolved the point.⁴³ A judgment announces who won and seldom says why, so the practical work lies in discovering which questions the court answered on the way to its result. Help may come from asking whether the first court lacked authority to decide a particular issue, since a court is unlikely to have decided an issue it lacked authority to decide. A variety of techniques have been adopted to identify the issues decided, and to the extent they fail, the opaque judgment fails to preclude relitigation.⁴⁴

The organizing idea is an asymmetry between winning and losing. When a case is litigated to a conclusion and judgment goes to the plaintiff, the plaintiff logically must have prevailed on every element of the claim. When judgment goes to the defendant, the plaintiff may have failed on only one element, and failure on one is enough, so the judgment does not show which particular issues the factfinder decided against the plaintiff.⁴⁵ Preclusion follows that logic because it cannot rest on a guess about what a factfinder concluded.

A general verdict for the plaintiff can therefore be decoded even though it lists no findings, because negligence and the other elements of a personal injury claim must have been established for the plaintiff to win.⁴⁶ Suppose the jury returns a $1,000 verdict for the plaintiff, and the only questions tried were the defendant's negligence and the plaintiff's contributory negligence, which in that jurisdiction would have barred recovery. The verdict necessarily found that the defendant was negligent, that the plaintiff was not contributorily negligent, that the defendant's negligence caused the injury, and that damages were $1,000. Each finding was a prerequisite to the verdict, so each counts as decided even though none appears on the form.

The same inference works in a bench trial. The judge in Otherson's criminal case returned a general verdict of guilty on two misdemeanor counts without specific factual findings, which raised the question how anyone could know what issues were necessarily determined.⁴⁷ The only ground the judge had for doubting the government's version of events was Otherson's cross-examination, which made general attacks on the witnesses' credibility. The agency official concluded that the judge must have found the government's witnesses credible, and therefore must have found that Otherson committed the acts charged. The court found that conclusion perfectly reasonable.⁴⁸

The defense side offers no such logic. Pat sues a city over a towed car and loses on a general verdict. He may have lost because he did not own the car, because the car was never towed, or because the city no longer had it, and the verdict does not say which. If Pat later sues the city on a constitutional theory arising from the same tow, and claim preclusion does not reach that theory, the city ordinarily cannot use the first verdict to bar him from proving ownership, since no one can say the jury actually decided that point.⁴⁹

The outcome of the hypothetical depends on who wins. If the buyer wins, the jury must have found that the supplier proved neither defense, because success on either one would have defeated the claim, and in a later suit the supplier cannot raise either defense again. That is the result in the Vanderbilt-and-Fisk illustration, where Fisk's repeated defense should be barred.⁵⁰ If the supplier wins, the verdict is consistent with success on either defense, so no one can say that either issue was decided in the supplier's favor.⁵¹ Rulings made before the verdict can change that result. Pretrial or pre-verdict motions that remove issues from dispute narrow what the jury could have decided, so if the court grants judgment as a matter of law against the mistake defense and sends only the specifications defense to the jury, a verdict for the supplier must rest on the specifications defense, and that issue is decided in the supplier's favor.⁵² The verdict form is the same in both versions of the case, and what differs is how much the court left for the jury. A grant of summary judgment works the same way, because it represents a finding that there is no genuine issue of material fact on the point and that one party is entitled to judgment.⁵³

The tenant's argument fails. A denial of summary judgment reports only that the evidence leaves a genuine dispute for trial, which is the opposite of the finding a grant represents, and a question that no factfinder ever answers remains undecided.⁵⁴

Parties who foresee later litigation can build the record themselves. The cure for an opaque general verdict is a special verdict or answers to specific interrogatories, and a bench trial in which the judge makes specific findings of fact serves the same function.⁵⁵

When the judgment itself is unclear, Russell v. Place states what the proponent must still show. The precise question must appear, either on the face of the record or by extrinsic evidence, to have been raised and determined in the former suit. If it is uncertain which of several distinct matters the judgment passed on, the whole subject matter of the action is “at large, and open to a new contention” unless extrinsic evidence shows the precise point involved and determined.⁵⁶ The pleadings, instructions, and verdict form discussed above are the usual extrinsic materials.

Courts divide when a general verdict could rest on any of several issues. Under the majority view, when the first judgment may have been based on different findings on more than one issue, no issue receives preclusive effect unless evidence shows which issues were actually litigated and decided. Under the minority view, the second court treats all issues as actually litigated and gives the judgment broad effect.⁵⁷ Extrinsic evidence also runs in one direction only: it may support the verdict and may not contradict it.⁵⁸ A defendant who won a general negligence verdict can try to show from the record that the jury rested on the plaintiff's contributory negligence, but the plaintiff cannot offer evidence that the jury never considered that defense, because the offer would contradict the verdict. If the basis remains unknown after the permitted inquiry, the proponent has not shown that the issue was decided, and the issue is open for trial again.

Essential to the Judgment

A court makes many findings on the way to a judgment, but only those the judgment depends on bind the parties later. Restatement § 27 requires that the determination be essential to the judgment.⁶⁰ The Supreme Court has made the same point by requiring that the issue be actually and necessarily determined by a court of competent jurisdiction before the determination is conclusive in later suits.⁶¹ The working test is a deletion exercise: remove the finding from the first case and ask whether the same judgment would still follow. If it would, the finding was incidental, and it precludes nothing in later litigation.

Two concerns explain the requirement. The first is attention. A finding on which the result does not turn resembles dictum, and limiting preclusion to necessary determinations assures that the rendering court recognized its determination to be significant and worthy of close attention.⁶² The second is review. The party against whom an unnecessary finding was made won anyway, so an appeal would complain of no harm, and an appellate court would likely refuse to hear it, since even a reversal of the finding would leave the outcome unchanged.⁶³ Comment h to § 27 describes such determinations as having “the characteristics of dicta,” which “may not ordinarily be the subject of an appeal by the party against whom they were made.”⁶⁴ The winner therefore has neither the incentive to contest the finding vigorously nor a way to get a mistake corrected, so the full opportunity to litigate that justifies preclusion is not assured.⁶⁵

A finding that sounds favorable to the party invoking it is the usual trap, and Cambria v. Jeffery supplies the standard example. P sued D for common-law negligence, and the court found both parties negligent but denied recovery because P's own negligence barred the claim. When D later sued P over the same accident, P invoked the earlier finding that D was negligent, which together with the contributory negligence rule would have defeated D's claim. The court refused to apply collateral estoppel against D, because the finding of D's negligence was not necessary to the first judgment: P's contributory negligence was enough to dispose of the case.⁶⁶ Deleting that finding leaves the first judgment intact. It was not even logically related to the judgment, since if anything it pointed toward liability, and D, as the winner, could not appeal it.⁶⁷

The finding on the other side of the same verdict has the opposite status. The determination that P was contributorily negligent is the reason P lost, so deleting it would change the result, and it is essential. The pattern appears again when a factfinder finds that neither driver was negligent. Danny wins because he was not negligent, and the added finding that Peter was not contributorily negligent is dictum, because it does not affect the result and the judgment would have been the same without it.⁶⁸

The second step does most of the work in practice, because a finding that hurts the winner almost never holds the judgment up, while findings against the loser usually do. The third step has one complication: a judgment resting on two independent findings survives the deletion of either one, which is the alternative-holdings problem taken up in the next subsection.

Balcom v. Lynn Ladder and Scaffolding Co. shows the test applied to a defendant's victory. An injured plaintiff sued a scaffolding company over a ladder, and the company brought in the ladder's manufacturer, seeking indemnity. The jury found that the manufacturer built the ladder and that it was defective, but that the defect did not cause the plaintiff's injury, so the manufacturer won. The First Circuit held that the manufacturer was free to relitigate who made the ladder, because the verdict rested on the absence of causation and the manufacturer could not have appealed the finding that it built the ladder.⁶⁹

The same result appears when the winner is the plaintiff. In an illustration involving a furnace explosion, Hill sued Villard and won a negligence verdict, while the first case also produced a ruling dismissing his strict-liability claim because the governing law did not recognize that theory. The verdict could not have come out as it did without rejecting the statute of limitations and contributory negligence defenses and finding the design negligent, so each of those determinations was essential. The strict-liability ruling tended to defeat liability rather than establish it, and Hill won despite it and had no reason or ability to appeal it, so in a suit over a second explosion he remains free to argue that strict liability is available.⁷⁰ Even the essential determinations carry forward only if the issue is identical, and two explosions at different times may present different limitations and contributory negligence questions.

Pretrial rulings can be litigated and decided and still fail the requirement. Before trial in one illustration, the court granted summary judgment to the plaintiff, Fortes, on the defendant's defenses that the parties never made the contract and that the contract was illegal, ruling as a matter of law that they made a valid contract. The case went to trial, and Fortes lost because the jury found that he had agreed to provide the stake and failed to do so.⁷¹ In a later suit between them, the defendant may raise those same two defenses again, because the judgment for him rested on the missing stake and he could not have appealed rulings made in a case he won.⁷² Had he lost instead, the same rulings would have been essential to the judgment against him, since Fortes could not recover on a contract that was never made or was void. Essentiality is therefore measured against the judgment actually entered, and a finding cannot be classified without knowing who won and why.

Necessity is measured against the whole judgment, including the relief it grants. In Bifolck v. Philip Morris USA Inc., the Second Circuit held that the necessity requirement is met where the issue is essential to the remedy imposed, even if it is not essential to a finding of liability.⁷³ The same question arises in insurance litigation. When an insurer defends under a reservation of rights, the resolution of the underlying claim against the insured can affect the insured's right to indemnification.⁷⁴ Whether any preclusion follows depends on what the underlying judgment actually and necessarily decided.

Alternative Holdings

The deletion test stops giving a clean answer when a court rests its judgment on two independent grounds, each sufficient by itself. Deleting either ground leaves the judgment standing on the other, so each can be called nonessential, yet the judgment plainly depends on at least one of them. Courts have not agreed whether such a ground is essential and so binds the loser.⁷⁶

The difficulty appears in the old contributory-negligence setting, where any fault by the plaintiff barred recovery. In one illustration, a court rules against a plaintiff for two independent reasons: the plaintiff was contributorily negligent, and the plaintiff's injuries were not caused by the accident. The plaintiff then sues a different defendant over the same accident, and the question is whether the first court's finding of contributory negligence carries over.⁷⁷ Three answers are conceivable: both findings bind, neither binds, or only one binds. The third can be set aside, because no principle explains why one of two equally sufficient grounds should bind while its twin does not. The first Restatement of Judgments chose the first answer and the Restatement (Second) chose the second, so the real disagreement lies between those two.⁷⁸

The Restatement (Second) treats neither ground as essential, and comment i gives two reasons that reinforce each other. The first concerns care. A judge who sees that a claim fails for one reason has less need to test a second reason as rigorously as if the result depended on it, so an alternative determination has, in the Restatement's phrase, “some of the characteristics of dicta.”⁷⁹ The second concerns review. The loser may appeal both grounds, but an appellate court can affirm by upholding either one and never reach the other, so the loser has reason to expect that at least one ground will survive and little reason to pay for an attack on the rest. If preclusion attached to both grounds anyway, losing parties would feel pressure to appeal only to head off the later preclusive effect, and a doctrine meant to lighten litigation would add appeals to it.⁸⁰

Courts that bind the loser on both grounds answer that a ground is no less necessary than its twin, since the judgment rests on each of them.⁸¹ Their position is easiest to defend when the first court's findings are specific. If a jury in a negligence suit returns a special verdict finding the plaintiff contributorily negligent and the defendant free from negligence, the defendant can invoke preclusion against the plaintiff on both issues, because the specific findings give some assurance that the factfinder took each one seriously.⁸² That response to the Restatement's first concern shows why a flat rule in either direction is hard to defend.

Many courts have therefore moved to a case-by-case approach. When the first court considered both issues carefully and the party to be bound had a full and fair opportunity to litigate them, both determinations are treated as essential and both bind, even though either would have supported the judgment alone.⁸³ The second court looks to the record of the first case to see whether each holding received full consideration.⁸⁴ An appeal can supply the same assurance, since the Restatement itself applies preclusion to both findings when the parties appealed and the appellate court affirmed both, and in that case the review the Restatement found missing has actually occurred.⁸⁵ The decisions divide along these lines. The Fifth Circuit in Society of Separationists, Inc. v. Herman followed the Restatement and denied preclusive effect to either finding, while the Third Circuit in Jean Alexander Cosmetics, Inc. v. L'Oreal USA, Inc. examined the earlier litigation to decide whether both findings had been fully considered.⁸⁶ Accounts of the trend disagree as well, with some calling the Restatement position the modern view and others calling the case-by-case approach more widely followed, which confirms that the point is unsettled.⁸⁷ An answer should therefore start by identifying which approach the controlling jurisdiction follows and, where that is unknown, set out the Restatement rule, the contrary rule, and the record-based middle course before applying them.

A jurisdiction that decides case by case would bind the manufacturer on both findings, because the days of evidence and the separate written analysis show that neither ground was treated casually. A jurisdiction following the Restatement (Second) would bind the manufacturer on neither, because the judgment was never affirmed on appeal, so each ground remains open to dispute. The facts are identical in the two jurisdictions, and only the rule about what makes an alternative ground essential changes the answer.

Every approach shares one limit: preclusion needs a determination the first court actually made. Telling a formal alternative holding from a passing remark can be hard, so the second court must confirm that the first court really decided the issue it is asked to treat as settled.⁸⁸ A remark of the second kind is dictum, and dictum precludes nothing under any of the three views.

Valid Final Judgment

The last element packs two separate conditions into one phrase: the judgment must be valid, and it must be final, or at least final enough.⁹⁰ For issue preclusion, a prior adjudication is final enough when it is sufficiently firm to be accorded conclusive effect, a standard more forgiving than the one for claim preclusion.⁹¹ The second court therefore asks how firm the earlier decision was, not whether the first case has run its entire course.

Validity concerns the first court's authority to decide the case. A judgment is valid when the court had subject-matter jurisdiction and personal jurisdiction over the defendant, and the inquiry does not ask whether the court decided correctly or committed error along the way. A wrong ruling is the business of appeal, so a second court that reopened every judgment for error would give up the finality that preclusion exists to secure. A court with no power over the case is different, because its judgment was never entitled to bind anyone. Even then, a court may refuse to give a judgment preclusive effect for lack of jurisdiction only under limited conditions, such as when the defect was never raised in the first proceeding and the defendant defaulted there.⁹² Courts generally analyze validity, like party identity and privity, the same way for both doctrines.⁹³

Finality starts from a shared baseline and then diverges. Finality occurs when the trial court enters judgment, and the taking of an appeal does not affect it unless the appellate court vacates the judgment and orders a trial de novo.⁹⁴ Issue preclusion keeps that rule about appeals and relaxes the rest. Claim preclusion ends an entire claim, including grounds nobody raised, so it can fairly insist on a completed adjudication. Issue preclusion binds a party only on a point that was actually contested and decided, which is a smaller commitment, so it asks only whether the decision on that point was firm enough to rely on.

Treating finality functionally keeps courts out of two opposite mistakes. If the doctrine demanded that every appeal be exhausted, the benefit of finality would be delayed, and a losing party could avoid a decision simply by appealing it. If any tentative ruling counted, parties would be bound by decisions the issuing court itself expected to revisit. Between those extremes, the second court looks at the quality of the earlier decision: whether the court reached it after real deliberation, whether it explained its reasons, and whether the parties could obtain review of it. A pending appeal also leaves the second court a choice. It may stay the later case until the appeal is completed, which is the sounder practice because it avoids the risk of giving binding effect to what may ultimately be found an invalid or erroneous judgment, or it may simply apply preclusion, because the choice is discretionary.⁹⁵

The two scenes differ in the quality of the decision rather than in the existence of an appeal. The first resulted from a full trial and a reasoned opinion, so an appeal does not remove its weight. The second was never meant to be the last word on the point.