Concept
Nonparty Preclusion
Claim preclusion
Nonparty Preclusion
A judgment ordinarily binds the parties to the suit and no one else. Taylor v. Sturgell states the rule in the Supreme Court's words: “one is not bound by a judgment in personam in a litigation in which he is not designated as a party or to which he has not been made a party by service of process.”1 The reason is the right to be heard. A person who was not a party generally has not had a “full and fair opportunity to litigate” the claims and issues the suit settled, so applying claim or issue preclusion to that person runs against the “deep-rooted historic tradition that everyone should have his own day in court.”2
The rule has a price. Suppose Peter sues Mary over a collision and loses, and Paul, injured in the same collision, then sues Mary on his own claim. Making Mary defend a second time seems unfair to her. Precluding Paul would be more unfair, because Paul has not had his day in court: he did not take part in Peter's case, he may have better evidence against Mary than Peter offered, and he may have a stronger reason to litigate vigorously, as when his injuries are more serious than Peter's.3 The system accepts the second suit as the lesser unfairness.
The exceptions are few and narrow. Each covers a nonparty so closely tied to a party to the first judgment that the nonparty is burdened and benefited by it as though a party. Call the party to the first suit P1, and call the nonparty who is now suing, and whom the defendant asks the court to bind, P2.4 The Court grouped the recognized exceptions into six categories:5
- P2 agreed to be bound by the determination of issues in an action between others.6
- P2 and P1 share a preexisting substantive legal relationship, such as preceding and succeeding owners of property. Courts call such ties privity.7
- P1, a party with the same interests, adequately represented P2, as in a properly conducted class action or a suit by a trustee or guardian.8
- P2 assumed control of the first litigation, although P1 was the named party.9
- P2 now sues as P1's designated representative or agent, so P1 would be relitigating through a proxy.10
- A special statutory scheme, such as bankruptcy or probate, binds nonparties, provided the proceeding comports with due process.11
The categories sort by the reason each gives for binding P2. The first two look to P2's own agreement and P2's legal tie to P1. The third looks to what P1 did for P2. The fourth and fifth look to who actually ran the first suit and who stands behind the second. The sixth looks to a legislative command. Falling within a category settles only who can be bound. The party asserting preclusion must still show the same claim in both suits, a valid and final judgment on that claim, and, in most jurisdictions, that the judgment was on the merits.12
Taylor supplies the example that runs through this unit. Herrick sued the Federal Aviation Administration under the Freedom of Information Act for agency records and lost. His friend Taylor then sued for the same records, and the two shared a lawyer.13 The courts below held Taylor to Herrick's loss under “virtual representation,” a doctrine several circuits had adopted in varying forms and one broader than any exception the Court had approved.14 The case asks what every category answers: whether someone with no formal role in the first suit may fairly be held to its result.
The court that rendered the first judgment determines which law governs who is bound. Federal common law fixes the preclusive effect of a federal-court judgment, and for a federal-question judgment such as Herrick's FOIA suit it supplies a uniform federal rule, subject to due process limits.15 A state-court judgment instead receives the same full faith and credit in every court in the United States that it has in the courts of the rendering state, so 28 U.S.C. § 1738 makes that state's preclusion law the starting point.16 A later section of this chapter develops the choice.
The subsections follow the Court's order. They begin with the exception that depends most plainly on the nonparty's own choice and end with the doctrine the Court refused to adopt.
Agreement to Be Bound
The first category rests on the nonparty's own choice. Under Restatement (Second) of Judgments § 40, a person who agrees to be bound by the determination of issues in an action between others is bound according to the terms of the agreement.17 The standard illustration is the test case. When different plaintiffs sue the same defendant in separate actions over the same transaction, all the parties to all the actions may agree that the defendant's liability will be definitely determined, one way or the other, in a single suit chosen to stand for the rest.18
Consent answers the objection that ordinarily shields outsiders. A nonparty keeps the right to a day in court, and a person who has agreed to abide by someone else's trial has chosen to give that right up.19 Both sides gain from the bargain. Claimants with the same grievance would otherwise prove the same facts in many suits, and the defendant would risk inconsistent answers about identical conduct; one trial settles the question for everyone. Because the agreement runs to all the parties to all the actions, it binds in both directions. A defendant who loses the test case cannot contest liability again, and claimants who lose cannot relitigate it.
The agreement also fixes how far the preclusion reaches. In California v. Texas, the Court dismissed certain defendants from a suit on the strength of a stipulation that each of them would be bound by the Court's final judgment on a specified issue.20 The stipulation named an issue, so the binding effect attached to that issue.
Everything in this category depends on a promise that was actually made. A homeowner who watches the first trial, testifies in it, shares Alvarez's lawyer, or hopes Alvarez wins has promised nothing. Each of those facts describes a connection, and none is an undertaking to be bound. A problem that piles up such connections is testing whether you will mistake closeness for consent. Look for language of commitment, such as a stipulation or a test-case arrangement, and then read its terms to see how far the commitment goes. The last subsection returns to what happens when courts try to infer binding effect from closeness alone.
Preexisting Legal Relationships
The second category depends on a legal tie that exists before the first suit begins. The Court explained that nonparty preclusion may rest on a variety of preexisting “substantive legal relationship[s]” between the nonparty and a party.22 Qualifying relationships include, but are not limited to, preceding and succeeding owners of property, bailee and bailor, and assignee and assignor.23 A bailor is an owner who hands goods to another, the bailee, to hold for a time; an assignor transfers a right, such as a claim for payment, to an assignee. Indemnitor and indemnitee, where one has promised to cover the other's liability, is another pairing.24 A nonparty in such a relationship is said to be in privity with the party, and the judgment burdens and benefits the nonparty as it would a party.25
Every relationship on the list shares one feature: the nonparty's rights in the subject of the suit come from, or lead to, the party's rights in it. That feature supports preclusion twice over. First, it explains why the second suit is really the first claim again. A buyer of land takes the seller's title as it stands, so if the seller has already lost a suit over where a boundary runs, the buyer never acquired a better claim to the disputed strip. If the buyer could sue again, any losing owner could escape a judgment by selling the land. Second, it makes binding the buyer tolerable, because a buyer can search the record before purchasing and price the risk of a decided dispute. The Court noted that these exceptions originated “as much from the needs of property law as from the values of preclusion by judgment,” which is why the property examples come first.26
The tie runs both ways. A nonparty linked to a party in this way is burdened and benefited by the judgment as if a party, so had Okafor won, Hu could have used that victory against Lindqvist.27
The error that costs points is treating privity as a synonym for closeness.
No preexisting legal relationship links Pruitt to Voss. Pruitt's claim for his injuries was his own from the moment of the collision and never passed through Voss, whereas a buyer's claim to the land is the claim the seller held and passed along. Matching interests and one lawyer show alignment, and in Taylor the Court declined to treat a strong incentive to litigate and a shared lawyer as making a friend an adequate representative.28
The requirement of a relationship that exists independent of the litigation follows from the day-in-court tradition, which leaves no other principled stopping point. A tie created by the lawsuit itself, such as a shared lawyer, shows only that two people cooperated. If closeness sufficed, a person could lose a claim without a hearing because someone nearby sued first. The result may seem unfair to Delgado, who must meet a second suit over the same accident, but barring Pruitt would be worse, because he has had no chance to be heard.29 The last subsection shows the Court refusing to stretch these ties to mere closeness.
Adequate Representation
The third category covers a nonparty whom someone else represented in the first suit. The Court confirmed that, “in certain limited circumstances,” a nonparty may be bound because she was “adequately represented by someone with the same interests who [wa]s a party” to the suit. Representative suits with preclusive effect on nonparties include properly conducted class actions and suits brought by trustees, guardians, and other fiduciaries.32
The earlier categories turned on what the nonparty did or already was. This one turns on what a party did for the nonparty, which is to stand in the nonparty's place and litigate the nonparty's interest. Representation can bind an outsider only to the extent that it replaces the opportunity the outsider lacked, so the question in every case is whether the stand-in gave the absent person something equal to a hearing.
Taylor states the minimum. The interests of the nonparty and her representative must be aligned, and either the representative must have understood herself to be acting in a representative capacity or the original court must have taken care to protect the nonparty's interests.33
Each condition does separate work. Alignment gives the representative a reason to make the arguments the absent person would make, because a litigant who gains by conceding a point that hurts someone else, or who has nothing at stake in an argument, has no incentive to press it. The second condition addresses a different weakness: two people can want the same result and still differ about whom the suit is for. A litigant who sues only for personal benefit chooses the evidence, the theories, and the settlement terms to serve that benefit and owes the absent person nothing. When the litigant knows the suit is brought for others, or a court is watching, someone is checking that the absent person's interests are handled. Problems about friendly litigants with matching interests usually leave this second condition out, and it is the one to look for.
Suits by fiduciaries satisfy both conditions through the nature of the office. A fiduciary is a person the law charges with managing another's affairs for that other's benefit, as a trustee does for the beneficiary of a trust and a guardian does for someone who cannot manage legal affairs alone.34 The beneficiary's interest is the very thing the office exists to protect, so alignment is built into the role, and a trustee or guardian who sues necessarily understands the suit as brought in that capacity.
A class action is the paradigm of court-supervised representation. Named plaintiffs sue for a defined group, and the court controls whether the suit may go forward on that basis, so the procedure builds in the protections the doctrine demands. Taylor itself refused to bind a friendly nonparty because doing so would have turned the first suit into a class action with none of the structural protections of Rule 23.35
Certification divides Okoye from Hollis. The Court's phrase “properly conducted class actions” signals that the label “class action” is not enough; the representation must meet the two conditions.36
An informal alliance cannot take the place of those protections. In Taylor, Herrick had a strong incentive to win and shared a lawyer with Taylor, and the Court still refused to treat him as Taylor's adequate representative.37 Those facts show aligned interests, which is the first condition. They show nothing about whether Herrick sued for Taylor or whether any court protected Taylor, which is the second. The Court drew the line at legal relationships that either preexisted the litigation or were created under a court's supervision.38 Okoye's stand-in was created and supervised by a court before judgment. A friend's role as stand-in would be a label applied afterward, by a defendant who won the first suit and wants the result extended.
Control and Proxy Suits
Two categories reach people who were never named in the first suit but are tied to it by what they did or by whom they now act for. The fourth binds a nonparty who “assume[d] control” over the litigation in which the judgment was rendered, and the fifth forbids a bound party from “relitigating through a proxy.”40 Control looks back at the first suit and binds the person who steered it. The proxy rule looks forward at the second suit and bars the person who speaks for someone the first judgment already bound.
A nonparty who took control of the first case has already received what the day-in-court tradition protects. Such a person has had “the opportunity to present proofs and argument,” and so has “had his day in court” though not a formal party, as Restatement (Second) of Judgments § 39 comment a explains.41 Controlling a suit means choosing the evidence, the witnesses, the theories, and whether to appeal or settle, so the controller held every power a party holds. The rule also closes a loophole. If control carried no consequence, a person could direct a case from behind a nominal party and refile in his own name after a loss, leaving the defendant facing an opponent who cannot lose.
The inquiry asks who in fact directed case one, and the name on the caption does not settle it. P2 may have assumed control of the first litigation “even though P1 was the named party.”42
Connections short of control do not suffice. A friend with matching goals and a common lawyer shows alignment, and alignment is not control. The question is who gave the instructions. A witness, a spectator, and a friend with a stake in the result all stand outside the first case, and control is the fact that moves a person inside it.
The proxy category addresses the reverse problem. The person now suing is a stranger to case one, but the person standing behind that stranger is not. Preclusion is in order when a person who did not take part in the first litigation later sues as the designated representative of a person who was a party to it. And although the Court's decisions had not addressed the point directly, preclusion also seems clear when a nonparty later sues as an agent for a party bound by a judgment.43 The reason is that the real party in interest has already had her day in court.44 If the bar reached only the person who lost, a loser could defeat claim preclusion by recruiting a relative to file the same complaint.
The two categories differ in what must be shown. Control requires proof about case one: that the nonparty in fact directed it. Proxy requires proof about case two: that the nonparty sues as the designated representative or agent of the bound party and not on a claim of the nonparty's own. Exam answers go wrong on that second requirement, because a nonparty may be closely connected to the loser and still assert a right that belongs to the nonparty alone. Ask whose claim is being pressed.
When case one was a state-court judgment, 28 U.S.C. § 1738 gives it the same full faith and credit in every court in the country that it has in the rendering state's own courts. The rendering state's preclusion law is therefore the starting point for deciding who is bound, including how that state treats a nonparty who controlled the suit or sues as an agent.45
Statutory Schemes
The sixth category covers special statutory schemes, such as bankruptcy and probate proceedings, that bind people who were not parties to the proceeding, provided the proceedings comport with due process.47 The first five categories find the reason for binding a nonparty in the nonparty's own conduct or position: a promise, a tie to a party, a representative, control of the first suit, or a proxy. Here the source of the binding effect is the statute itself. The legislature has said who is bound by the outcome of a particular kind of proceeding, so a court looks for that command and for the process the scheme promises.
The purpose of such schemes explains why the law lets them bind outsiders. Bankruptcy and probate are collective proceedings, built to settle the claims of many interested people at once. A bankruptcy gathers a debtor's creditors to divide a limited pool of assets, and a probate proceeding fixes who takes a decedent's property so that heirs, creditors, and later buyers can rely on the answer. A proceeding of that kind cannot do its work if every interested person may reopen the result in a separate suit.
The due process condition keeps the exception from swallowing the rule. The general rule rests on a deep-rooted tradition that everyone should have his own day in court, so the exception reaches only proceedings that comport with due process, and the federal common law of preclusion is itself subject to due process limits.48 A statute that declared a person bound by a proceeding the person never learned of would cut that person off without a chance to be heard, which is the injury the tradition forbids.
The ordinary civil suit sits on the other side of the line. With no statute, only the parties and those within one of the other categories are bound. In Taylor, the courts below bound a nonparty on virtual representation, a doctrine the Supreme Court had never approved and none of the six categories includes.49 An answer that invokes the sixth category must name the particular scheme that binds nonparties and confirm that the person to be bound received the process the scheme promises. The number of people a dispute affects does not create a statutory scheme.
Virtual Representation Rejected
Virtual representation bound a nonparty to a party's result without a legal relationship between them. The six categories generally rest on an express or implied legal relationship between the nonparty and the party who litigated the first case, and this doctrine subjected a nonparty to preclusion in the absence of one.51 Taylor was the Court's first look at the doctrine, although several circuits had adopted it in varying forms and the courts below relied on it.52 The exception so styled is broader than any the Court had approved.53
The two suits in Taylor were as alike as two suits can be. Both asked whether the Freedom of Information Act entitled a requester to the FAA records, and the particular identity of the claimant made no difference to that question.54 That sameness is why the courts below were tempted to treat Herrick's loss as Taylor's.
The Court nonetheless rejected the preclusion claim. Binding Taylor would, in effect, have turned Herrick's suit into a class action with none of the structural protections of Rule 23.55 A class judgment reaches absent people because a court supervised the representation. Herrick's suit had no such supervision, so the label “representative” would have been attached afterward by a defendant who wanted the result. The Court also refused to stretch adequate representation to fit Herrick. Representation justifies preclusion only if, at a minimum, the interests of the nonparty and her representative are aligned and either the representative understood herself to be acting in a representative capacity or the original court took care to protect the nonparty. Herrick's strong incentive to litigate and his shared lawyer did not meet that test. They show that two people wanted the same outcome, not that Herrick sued for Taylor or that a court watched over Taylor's interests.56
The refusal protects predictability. Virtual representation asks a judge to decide after case one whether two people stood close enough that one spoke for the other, and only hindsight can answer. A friend deciding whether to sue could not know whether a stranger's earlier loss would bar the claim. The recognized categories avoid that uncertainty because each rests on something fixed before judgment or on a safeguard a court supplied: Taylor drew the line at legal relationships that either preexisted the litigation or were created under a court's supervision.57 Friendship is neither, and a shared lawyer is a private arrangement no court examined. A stranger to the first suit whose facts fit none of the six narrow exceptions is not bound, however close the stranger stood to a party.58
The FOIA setting pressed the best argument for loosening the rule. A request for agency records is public in the sense that nothing about the requester changes the outcome, so indistinguishable requesters can bring the identical claim in succession, and the government must keep defending a question it has already won. The Court's limited view of preclusion is in tension with that situation, yet the Court refused the preclusion claim anyway.59 A public claim does not turn a friend into a party.
The two scenes differ by a single fact. Dalton stood close to Reyes's suit, while Ostrow ran Lind's, and only the second fact fits a category. An exam stem that piles up friendship, shared interests, and a shared lawyer is testing whether you will treat closeness as a ground for binding a nonparty. Check the facts against the six categories one at a time.