Concept
Clear Error for Judge Findings
Scope of review for judge and jury
Clear Error for Judge Findings
A judge who decides the facts does the work a jury would otherwise do: the judge hears the witnesses, weighs the exhibits, resolves the conflicts among them, and announces what happened. Rule 52(a)(6) governs how an appellate court reviews that work. Findings of fact, whether based on oral or other evidence, must not be set aside unless clearly erroneous, and the reviewing court must give due regard to the trial court's opportunity to judge the witnesses' credibility.1
The standard is the counterpart of de novo review. De novo review governs what the judge decided about the law; clear-error review governs what the judge decided about what happened. The same subject can produce either one, depending on what the judge did. A court that finds the facts needed to decide domicile for a tax claim makes findings reviewed for clear error, while a court that chooses a legal test for domicile presents a question of law reviewed de novo.2
The deference is real, and it has limits. A reviewing court may not reverse a finding simply because it would have found differently, but it may reverse one that is clearly erroneous.3 Judge-found facts therefore occupy a middle position: far harder to overturn than a legal ruling, yet not conclusive.4 The reason is the comparative advantage described earlier. The trial judge saw and heard the witnesses, while the appellate court reads a record that omits their demeanor.5 Sorting the challenged ruling therefore comes first. An appellant who attacks a finding of fact must overcome this deferential standard; one who attacks a legal ruling meets no deference at all.
The subsections below state the standard and the reasons for it, show that it reaches findings drawn from documents and inferences, explain why credibility findings receive the most protection, and show that clear-error review follows the judge's fact-finding function even outside a full bench trial.
The Rule 52(a)(6) Standard
Rule 52(a)(6) packs two commands into one sentence. Findings of fact, whether based on oral or other evidence, must not be set aside unless clearly erroneous; and the reviewing court must give due regard to the trial court's opportunity to judge the witnesses' credibility.6 The first command sets the threshold a challenger must clear. The second tells the appellate court to respect the judge's firsthand view of the witnesses while it applies that threshold. The phrase clearly erroneous is not self-explanatory, so the task here is to learn what it demands and why the standard sits where it does.
The Supreme Court supplied the working definition in Anderson v. City of Bessemer City. A court of appeals there had overturned a district court's fact-findings in a gender discrimination case, and the Supreme Court reversed, using the occasion to explain the standard.7 Anderson started from a formulation first stated in United States v. United States Gypsum Co.: a finding is clearly erroneous when, “although there is evidence to support it, the reviewing court on the entire evidence is left with the definite and firm conviction that a mistake has been committed.”8
Each phrase of that sentence does separate work. The opening clause concedes that evidence supports the finding, so support in the record does not by itself protect a finding. The words on the entire evidence send the court to the whole record, including everything that cuts against the finding. And definite and firm conviction fixes the degree of certainty: doubt about the finding, or a preference for the other side's account, falls short. Together the phrases mean that a finding falls only when the full record leaves the reviewing court firmly convinced that the trial judge erred. That is why an appellate court may not reverse merely because it would have found the facts differently; it may reverse only when the finding is clearly erroneous.9
Anderson then added refinements that decide most appeals. The standard does not entitle a reviewing court to reverse simply because it is convinced that it would have decided the case differently. If the district court's account of the evidence is plausible in light of the record viewed in its entirety, the court of appeals may not reverse it, even though it would have weighed the evidence differently as trier of fact. From that premise the Court drew a corollary: “Where there are two permissible views of the evidence, the factfinder's choice between them cannot be clearly erroneous.”10
The corollary follows from what a record with two permissible views looks like. Such a record supports more than one reasonable account, and the trial judge, who had to decide, picked one. An appellate judge who prefers the other account holds a different opinion about a close call, and a different opinion about a close call falls short of a firm conviction of mistake. The choice therefore stands whichever way the trial judge went.
What moves the answer between the two scenes is whether the record leaves room for the judge's account. On the narrowest statement of the standard, a finding survives unless it rests on a misunderstanding of the law or lacks adequate evidentiary support, and the second finding fails that test.11 Anderson's formula reaches one step further, because a finding with some support can still fall when the whole record shows a mistake, although the plausibility gloss keeps that door narrow.12
The Court's reasons for this much deference reach beyond the one the Rule itself names, the trial judge's chance to assess credibility. Fact-finding is the trial judge's major role, and experience in that role builds expertise. Duplicating the judge's efforts on appeal would add only negligibly to the accuracy of fact determination at a huge cost in judicial resources. And the parties have already concentrated their energies on persuading the trial judge; asking them to persuade three more judges asks too much. The trial on the merits should be the “main event,” not a “tryout on the road.” For these reasons, clear-error review, with its deference to the trier of fact, is the rule, not the exception.13
The standard's history explains its middle position. Equity practice had moved away from the broad de novo review of ancient chancery and treated a trial court's findings as carrying great weight, though not conclusive. Because a single standard was thought desirable for all nonjury cases, whether historically legal or equitable, the broader equity review was adopted. Findings therefore reach the appellate court, in one court's phrase, “well armed with the buckler and shield” of Rule 52(a): they are not conclusive on appeal, but the party seeking to overturn them carries a heavy burden.14
On an exam, sort the appellant's complaint before doing anything else. Trial-procedure problems often state the litigants' requests and objectives cursorily and leave you to work out the reasoning, so the real target of the appeal can be hard to see.15 If the complaint concerns what happened, ask whether the record leaves room for the judge's account; if it does, affirm. The standard reaches findings of ultimate fact as well as basic facts, so relabeling a finding as a conclusion does not escape it.16 If the complaint is that the judge applied the wrong legal rule, the issue is one of law, and the de novo review described earlier governs.
Documentary Evidence and Inferences
The words “whether based on oral or other evidence” in Rule 52(a)(6) answer a question that the credibility clause in the same sentence invites.18 That clause tells the reviewing court to respect the trial judge's opportunity to judge credibility. A judge who decides from exhibits, deposition transcripts, or inferences never watched anyone's demeanor, so the advantage the clause names went unused. It is tempting to conclude that deference shrinks when no witness testified on the point. The Rule says otherwise: the standard attaches to the finding, whatever evidence produced it.
Courts of appeals once reasoned the tempting way. One influential opinion put it this way: “If [the judge] decides a fact issue on written evidence alone, we are as able as he to determine credibility.”19 Various courts of appeals asserted that findings not based on credibility could be reviewed de novo. Anderson traces that theory to Orvis v. Higgins, a 1950 Second Circuit opinion written by Judge Frank, and notes later decisions in the Sixth and Eighth Circuits that followed it.20 The theory had a certain logic, because it rested on the institutional differences, or the lack of them, between trial and appellate courts.21 Its weakness was the text of the Rule, which contained no such exception.
The Supreme Court rejected the theory in Anderson. The clearly erroneous test applies, it held, “even when the district court's findings do not rest on a credibility determination, but are based instead on physical or documentary evidence or inferences from other facts.”22 The Court could not trace the theory back to the text of Rule 52(a), which states that findings of fact shall not be set aside unless clearly erroneous, and the Rule's added emphasis on credibility does not alter that command. As the Court had said in Pullman-Standard v. Swint, Rule 52(a) “does not make exceptions or purport to exclude certain categories of factual findings from the obligation of a court of appeals to accept a district court's findings unless clearly erroneous.”23 Later in 1985 an amendment added “whether based on oral or other evidence” to the Rule, so the holding now appears on the face of the text.24
Demeanor is only one of the trial judge's advantages. Fact-finding is the judge's major role, and experience in that role builds expertise; duplicating the judge's efforts on appeal would add little accuracy at a huge cost in judicial resources.25 The judge has also sifted the whole case, so a document or an inference is read against everything else the judge heard, while the appellate court sees only the limited issues and rulings presented to it.26 And even where an appellate court is as well placed to decide as the trial court, disregarding the trial court's finding impairs confidence in the trial courts and multiplies appeals with their expense and delay.27
Inferences receive the same protection as findings drawn directly from proof. The Supreme Court has applied clear-error review to a finding of negligence, to the determination whether a payment to a taxpayer was a “gift,” and to whether an employer had a discriminatory intent.28 Each is a conclusion the judge draws from underlying facts, and each sits near the boundary between fact and law. If an error of law has impaired the judge's judgment on such a mixed question, the finding should be set aside; but when the finding applies a correct principle of law to the facts, clear-error review governs. In view of that array of precedent, attempts to draw nice distinctions between fact and law in order to broaden review of determinations the trial court makes better seem doomed to fail.29 The later section on mixed questions takes up that boundary in detail.
The sorting move is to ask what the appellant is attacking before asking what kind of evidence the judge used. A complaint about what happened receives clear-error review whether the proof was spoken or written, and a complaint about what the law requires receives de novo review whatever the proof. Findings that rest on credibility receive even more protection, as the next subsection explains.
Credibility Findings
Rule 52(a)(6) contains a clause that the earlier subsections noted but did not develop. After directing that findings of fact not be set aside unless clearly erroneous, it requires the reviewing court to give due regard to the trial court's opportunity to judge the witnesses' credibility.31 The clause singles out credibility because that is where the judge's advantage is most pronounced: it peaks where credibility turns on demeanor evidence that only the trial judge can observe.32 A finding that rests on a witness's believability therefore receives clear-error review with extra weight behind it.
Anderson, which held that clear-error review reaches findings drawn from documents and inferences, drew a further line inside that holding. “[W]hen findings are based on determinations regarding the credibility of witnesses, Rule 52(a) demands even greater deference to the trial court's findings; for only the trial judge can be aware of the variations in demeanor and tone of voice that bear so heavily on the listener's understanding of and belief in what is said.”33 The trial judge watched and listened to the witnesses, while the appellate court reads a transcript that preserves the words and little else, since demeanor evidence is unavailable on appeal.34
United States v. Aluminum Co. of America explains why findings of this kind come close to beyond review. In such cases the appeal must be decided on an incomplete record, because “the printed word is only a part, and often by no means the most important part, of the sense impressions which we use to make up our minds.” An appellate court must have some affirmative reason to reverse anything done below, so “to reverse a finding it must appear from what the record does preserve that the witnesses could not have been speaking the truth, no matter how transparently reliable and honest they could have appeared.”35
Because demeanor is exactly what the record lacks, an appellant cannot argue that a witness seemed shifty or sincere. Any argument for reversal has to come from the content of the record itself. Even where the conflict is between two witnesses' direct testimony, appellate courts ought to be chary of reversing, and on an issue like a witness's own intent, which only the witness can testify about, the finding is “unassailable” except in the most exceptional cases.36 Unassailable describes how rarely an appellant wins; it does not make the finding immune, because the governing text still asks whether the finding is clearly erroneous.37
The dispute over the phrase “moral turpitude” shows the ordinary outcome. P and D's CEO gave conflicting testimony about what the parties meant by that phrase, and the trial judge credited the CEO's version over P's. The judge also found that a reasonable observer would read a photograph as showing poor moral judgment. Because evidence in the record supports the findings, and because the judge's credibility determinations are entitled to deference, the appellate court should find no clear error and should affirm.38
The appellant in that scene wins because the proof of falsity sits in the record itself, in documents the appellate court can read as well as the judge could. A tenant who held only her own word against the landlord's would lose on appeal, since reversing would require the appellate court to pick a witness to believe from a transcript, the one task it is poorly placed to perform.
Credibility also marks a gradient within fact review. An appellate court, the Alcoa opinion observed, will hesitate less to reverse the finding of a judge than that of an administrative tribunal or a jury, but it will still reverse most reluctantly and only when well persuaded. That holds to a considerable degree even when the judge has not seen the witnesses, because the judge's duty is to sift the evidence, put it into logical sequence, and make the proper inferences, and in a record of over 40,000 pages an appellate court could not function at all without ascribing some prima facie validity to those conclusions.39 Findings built on documents therefore receive ordinary clear-error deference, and findings built on a choice between witnesses receive more. Both carry the label clear error, but the appellant's burden differs: a challenger to a document-based finding can point to the paper, while a challenger to a credibility call succeeds only by showing that the record itself makes the testimony impossible to accept.
Findings Outside Bench Trials
Rule 52(a) is written for judicial fact-finding after bench trials, and that placement tempts students to read clear-error review as a bench-trial rule. The reading is a mistake: the clearly erroneous standard applies to most all trial court factual findings, whether or not they were made in the context of a trial.41 The reason lies in why the standard exists. Deference attaches to the judge's position as the person who handled the evidence and resolved the dispute, so the standard follows that fact-finding function wherever the judge performs it. Appellate courts scrutinize a judge's factual determinations deferentially whether the judge made them after a trial without a jury or in order to rule on a motion, and they may not reverse merely because they would have found differently.42
The most common setting outside a bench trial is a jury case in which the judge decides a threshold question that has nothing to do with the merits. Even in a jury trial, a judge may make findings of fact to rule on motions to dismiss for lack of personal jurisdiction, improper venue, insufficient service of process, and similar non-merits defenses, and Rule 52(a)(6) governs review of those findings.43 When the facts bearing on the defense are disputed, the judge must resolve them to rule, even though a jury will later decide the claim itself. A judge doing that work stands where a bench-trial judge stands, so the reasons for deference carry over.
The domicile problem shows the same point in a decided example. When the court found the facts necessary to decide domicile for purposes of a tax claim, those findings drew clear-error review, because Rule 52's standard contains no limitation to fact-finding in a bench trial. When a court instead chose a legal test for domicile, the choice presented a question of law reviewed de novo. The standard attaches to the particular ruling the appellant attacks, not to the proceeding as a whole.44
The same logic reaches the preliminary facts a judge finds on admissibility. In Bourjaily v. United States, the trial judge's finding that the defendant and the declarant were co-conspirators was reviewed for clear error, and it stood because the record contained sufficient evidence to support it by a preponderance.45 A preliminary question of that kind is for the judge, not the jury, because the answer controls what evidence the jury may hear. Keep two rulings apart here. The finding that the two people conspired together is a finding of fact reviewed for clear error. The judge's ultimate decision to admit or exclude the evidence is an evidentiary ruling, and appellate courts review most evidentiary rulings for abuse of discretion, a standard taught later in this chapter.46
The trigger for clear-error review is that a judge actually resolved a factual dispute; the setting matters much less. The other pole is the jury. A fact pattern that places a jurisdictional hearing and a jury verdict in the same lawsuit tests whether you see that two fact-finders produced two sets of facts, each with its own standard. The judge's findings on the jurisdictional motion draw clear-error review under Rule 52(a)(6). The facts a jury found on the merits draw a different standard, because the Seventh Amendment bars reexamining them.47 The appellate court asks only whether a reasonable jury could have reached the verdict, and the jury's role as decision-maker earns that highly deferential review.48 The later section on review of jury findings develops that standard.