Legal Sufficiency Standard

Legal Sufficiency Standard

Rule 50(a)(1) turns on a single phrase. The court may act once a party has been fully heard on an issue and the court finds that a reasonable jury would not have a “legally sufficient evidentiary basis” to find for that party on it.¹ On that finding, the court may resolve the issue against the party and may grant judgment as a matter of law on any claim or defense that, under the controlling law, can be maintained or defeated only with a favorable finding on the issue.² Federal courts call the request a motion for judgment as a matter of law. Older cases and many state courts call it a motion for directed verdict, the name the federal rules dropped in 1991.³

Rule 50 says little about when a basis is legally sufficient, so the phrase takes its meaning from decisions that apply it to particular trial records.⁴ Those decisions share one premise. The inquiry is a question of law: the judge decides whether the record would permit a rational jury to rule for the party opposing the motion, and leaves to the jury which side deserves to win and which witnesses to believe. A judge who blurs the two questions errs in one of two directions. Weighing credibility takes from the jury a decision that belongs to it, and deferring to any scrap of proof sends the jury a case that the law says should end.

The sections below follow the order in which a judge meets the problem. Has the party been fully heard? Could a reasonable jury find for that party? How thin may the party’s proof be before a jury could not rely on it? How should the judge read the record? Who must produce evidence on the issue? And what may the court do once the evidence falls short?

Fully Heard Requirement

Rule 50(a)(1) sets two conditions, and the first concerns timing. The court may act only if a party “has been fully heard on an issue during a jury trial”; only then does it ask whether a reasonable jury would lack a legally sufficient evidentiary basis to find for that party on the issue.⁵ The order makes sense because sufficiency tests the record a party actually made, and no record exists until the party has had its chance to build one. A court that ruled sooner would be predicting what the evidence would show instead of deciding what it did show.

The usual moments follow from the burden of proof. A plaintiff must prove each element, so the defendant may move for judgment as a matter of law when the plaintiff rests, before calling a single witness. Once both sides have rested, either party may move.⁶ The close of the plaintiff’s case and the close of all the evidence are the two moments an exam is most likely to test.

The rule speaks of being fully heard “on an issue,” so the condition is measured issue by issue and not for the case as a whole. The 1991 advisory committee notes, as McSherry v. City of Long Beach quotes them, call early action appropriate “when economy and expedition will be served,” but they bar judgment “against a party who has not been apprised of the materiality of the dispositive fact and been afforded an opportunity to present any available evidence bearing on that fact.”⁷ The issue-by-issue reading matters most when the burden of proof is split. A defendant who carries the burden on a release defense is fully heard on it only after presenting his own proof, so a plaintiff’s motion attacking the defense must wait for that proof even though the plaintiff rested long before.

Rule 50(a)(2) invites a mistake here, because it provides that a motion “may be made at any time before the case is submitted to the jury.”⁸ Read alone, those words let a party move the moment trial opens. Suppose a defendant moves right after the plaintiff’s opening statement, before any witness testifies, and argues that the plaintiff cannot prove causation. May the court grant the motion? It may not. McSherry read the two provisions together: Rule 50(a)(1) sets the standards under which a court may grant judgment as a matter of law, while Rule 50(a)(2) explains when a party may make the motion. The second supplements the first by telling the movant that it may file until the case goes to the jury, but it does not eliminate the requirement that the opponent be “fully heard” on the issue.⁹ The early motion fails because the plaintiff has presented no evidence on causation.

The deadline in Rule 50(a)(2) protects the opponent. The advisory committee notes explain that motions must come before the close of trial so that the responding party has a chance to cure any deficiency in its proof that went unnoticed until called to its attention, and they add that the rule requires the movant to articulate the basis for judgment so that the non-moving party knows what to correct.¹⁰ Rule 50(a)(2) therefore fixes how late a party may move, and the fully-heard condition in Rule 50(a)(1) fixes how early the court may act.

One treatise notes some authority for granting judgment after opening statements and before any evidence, but only when it is clear that no question exists for the jury to consider.¹¹ That authority is thin and sits in tension with McSherry’s reading, so the safer answer on an exam is that the opponent must first have presented its proof on the issue.

Reasonable Jury Test

The fully-heard requirement fixes when a court may ask the sufficiency question, and the reasonable-jury test fixes what it asks. The rule builds the inquiry around a hypothetical jury and not around the judge’s own view of the case.¹³ In Anderson v. Liberty Lobby, Inc., a summary-judgment case, the Supreme Court agreed that the summary-judgment standard mirrors the directed-verdict standard under Rule 50(a): the judge must direct a verdict if, under the governing law, there can be but one reasonable conclusion as to the verdict, and must not direct one if reasonable minds could differ about the import of the evidence.¹⁴ Because the two motions share one inquiry, the Court described it for both as whether the evidence presents a sufficient disagreement to require submission to a jury or is so one-sided that one party must prevail as a matter of law.¹⁵

The test looks to what a jury could do because the jury decides disputed facts. A judge who could take a case away whenever the judge preferred the other side’s evidence would be deciding the case, so the rule lets the judge step in only when the evidence would force a rational jury to a single conclusion. Anderson states the working question this way: the judge asks not whether the evidence unmistakably favors one side but “whether a fair-minded jury could return a verdict for the plaintiff on the evidence presented.”¹⁶ The question carries the trial standard of proof inside it, because it asks whether reasonable jurors could find by a preponderance of the evidence that the plaintiff is entitled to a verdict.¹⁷

Sufficiency is a different question from relevance. Many items of evidence may be relevant under FRE 401 and 402, and the question at the end of the proof is whether, added together, they are enough to let a reasonable jury find for the party who bears the burden of persuasion. If the evidence taken as a whole would support a finding for either side, the jury must decide the case.¹⁸ The same logic explains why a motion often called a no-evidence motion does not turn on whether the opponent offered literally no evidence. A party can offer some evidence and still lose, because the question is whether what it offered was enough for a factfinder reasonably to rule in its favor.¹⁹

Pennsylvania Railroad Co. v. Chamberlain shows the test at the weak end of the range. A railroad brakeman died in a rail yard, and the plaintiff, suing for his death, rested the case on the claim that a violent collision between a string of nine cars and the string he was riding caused his fall.²⁰ The railroad would be liable only if a collision caused the accident. The plaintiff’s only evidence of one was a witness named Bainbridge, who saw the two strings of cars traveling close together and later heard a loud crash. When he looked up, he saw the strings moving in tandem, but from his vantage point and distance he could not tell whether they had touched, and he was not sure the crash had come from those cars rather than from elsewhere in the yard.²¹ Three employees riding the nine-car string testified positively that no collision occurred, and eyewitnesses denied one as well, with no direct rebuttal from the plaintiff.²² The trial court directed a verdict for the railroad, and the court of appeals reversed, finding a jury question in the conflicting testimony. The Supreme Court reversed again, holding that a verdict could not rest on a fact inferred from Bainbridge’s testimony when the uncontradicted testimony of unimpeached witnesses indicated that the inferred fact never occurred.²³

The decision can look like credibility weighing, but the Court ruled on the capacity of the plaintiff’s proof. It concluded that no legally sufficient evidentiary basis supported a verdict for the plaintiff, and it reached that conclusion only after considering the substantial evidence the railroad had offered.²⁴ The employees gave direct accounts from where they watched the cars. Bainbridge supplied a noise and a position from which a collision might be inferred, and nothing more. The line between that kind of ruling and impermissible weighing is thin, so the question to ask is whether the plaintiff’s evidence could carry a rational verdict at all, and not whether the judge finds the defense witnesses more convincing.

The dock scene differs from Chamberlain in one respect: the plaintiff’s witness claims to have perceived the contact itself. Choosing between him and the drivers requires deciding who is telling the truth, and that decision belongs to the jury. Forgetting this is the common mistake. A reviewing judge may think the jury got the facts wrong, but it is not enough that the judge disagrees with the verdict.²⁵

When a case turns on the credibility of witnesses, judgment as a matter of law will almost never be appropriate.²⁶

Picture the possible records on a scale, running from the weakest showing to the strongest. Reasonable minds often disagree about whether a body of proof clears the midpoint, so a wide band around it counts as debatable. Proof anywhere in that band goes to the jury, and the judge defers to whatever verdict results, even if the judge privately thinks the plaintiff’s showing falls short.²⁷ Only proof at the extremes is decided by the judge. A showing as thin as the plaintiff’s in Chamberlain sits at the weak end and supports judgment for the defendant. Proof so overwhelming that only one verdict is possible can support judgment for the plaintiff, although that result is far less common.²⁸ Two questions follow from this one: how little evidence still counts as enough, and how the judge reads the record in the non-movant’s favor. The next two subsections take them up.

Scintilla and Substantial Evidence

The reasonable-jury test leaves open how thin a party’s proof may be before no reasonable jury could rely on it, and courts have answered with two formulations. Under the scintilla test (a scintilla being a spark or trace), the judge denies the motion and sends the case to the jury if there is any evidence at all on which the jury might possibly rule for the non-movant. Under the substantial-evidence test, the judge grants the motion unless the non-movant has produced sufficient, substantial evidence suggesting that the jury might decide in its favor.³⁰ Federal courts follow the second. In Anderson, the Supreme Court repeated an old statement from Munson: judges were once bound to leave a case to the jury if a scintilla of evidence supported it, but the settled rule makes the preliminary question “not whether there is literally no evidence, but whether there is any upon which a jury could properly proceed to find a verdict for the party producing it, upon whom the onus of proof is imposed.”³¹ Anderson then applied the idea to the ordinary civil case: “The mere existence of a scintilla of evidence in support of the plaintiff’s position will be insufficient; there must be evidence on which the jury could reasonably find for the plaintiff.”³²

The Fifth Circuit’s opinion in Boeing Co. v. Shipman is a leading statement of the prevailing view, and it guards against two opposite mistakes. Motions for judgment as a matter of law “should not be decided by which side has the better of the case, nor should they be granted only when there is a complete absence of probative facts to support a jury verdict. There must be a conflict in substantial evidence to create a jury question.”³³ The first half keeps the judge from picking a winner, because weighing conflicting evidence belongs to the jury. The second half keeps the judge from treating any probative fact as enough, which would bring the scintilla rule back in.

The two tests differ because they answer one question differently: how much control a judge should keep over the jury. The scintilla rule protects the jury’s role but lets a verdict stand on a fragment of proof that jurors could use only by guessing. The substantial-evidence test requires the party who bears the burden to offer proof that does real work toward the conclusion it needs, and it still reserves to the jury every question on which reasonable people could differ. Courts have chosen between the two according to their attitudes toward jury control, and the trend has been toward substantial evidence.³⁴

Chamberlain, the decision that rejected the scintilla rule, lets one record be run through both tests.³⁵ The court of appeals found conflicting testimony and, applying the scintilla test, found a jury issue. That result follows from the scintilla rule: Bainbridge’s testimony was some evidence on which a jury might possibly infer a collision, so the case goes to the jury however many witnesses deny it. Under the substantial-evidence test the question is whether his testimony could support a rational finding of a collision against uncontradicted accounts from unimpeached eyewitnesses who watched the cars. It could not, so the Supreme Court held that the verdict could not rest on the inferred fact.³⁶ The record was the same in each analysis, and the threshold alone decided who won.

The two tests part company only in the thin band of proof between nothing and enough. In the dock-worker hypothetical, a witness who watched the truck strike the pallets supplies substantial evidence, so both tests send the credibility contest to the jury. A plaintiff who offers nothing on an element loses under both. A fragment that is merely colorable or not significantly probative will not carry an issue to the jury, a point Anderson made in the parallel setting of summary judgment.³⁷ An answer that denies a federal Rule 50 motion because the plaintiff put on some evidence has applied the scintilla rule and asked the wrong question. The federal question is whether that evidence could carry a rational verdict.

Some courts still apply the more permissive approach. Under the scintilla approach, courts deny the motion if there is a scintilla of evidence supporting the non-movant, and they will not consider the moving party’s evidence even when it is uncontradicted and unimpeached.³⁸ That rule of construction makes a directed verdict almost impossible to obtain.³⁹ Unless a problem states that a local scintilla rule governs, the sufficiency analysis uses the federal test.

Viewing the Evidence

The sufficiency test also leaves open which evidence the judge counts and how the judge reads it.⁴¹ Reeves v. Sanderson Plumbing Products, Inc. states the federal answer. The court draws all reasonable inferences in favor of the non-moving party, and it may not make credibility determinations or weigh the evidence, because “credibility determinations, the weighing of the evidence, and the drawing of legitimate inferences from the facts are jury functions, not those of a judge.”⁴² A party may move at any point before the case goes to the jury, so the record under review is the one that exists when the motion is made, whether at the close of the plaintiff’s case or at the close of all the evidence.⁴³ The standard runs in two directions, one favoring the party who opposes the motion and one preserving the weight of the movant’s proof.

The first direction favors the non-movant. Boeing directs the court to consider all of the evidence, including evidence that does not support the non-mover’s case, but “in the light and with all reasonable inferences most favorable to the party opposed to the motion.”⁴⁴ The reason follows from the question. The test is whether a reasonable jury could find for the non-movant, so the judge has to picture the jury believing the non-movant’s witnesses and choosing the most helpful inferences, then ask whether a verdict resting on those choices would still be rational. A judge who adopted the more likely version of events would be doing the jury’s work. That is why the same formulation assigns to the jury, and not the court, the task of weighing conflicting evidence and inferences and judging the credibility of witnesses.⁴⁵

The benefit extends to inferences, which is why a case can reach the jury even when no one disputes the underlying facts. In Railroad Co. v. Stout, a child was hurt while playing on an unlocked, unfenced railroad turntable at the edge of a sparsely populated rural settlement. No one contested any of those facts, the jury found negligence, and the Supreme Court rejected the railroad’s argument that the judge should have taken the case from the jury.⁴⁶ Deciding whether agreed facts amount to negligence requires applying an open-ended standard to them, and that application belongs to the jury even on an agreed record. Lavender v. Kurn makes the same point about disputed records: when fair-minded people may draw different inferences, “a measure of speculation and conjecture is required on the part of those whose duty it is to settle the dispute by choosing what seems to them to be the most reasonable inference.”⁴⁷ The inference must still rest on evidence. In Galloway v. United States, a veteran claimed insurance benefits on the theory that insanity had left him totally and permanently disabled before his policy lapsed in May 1919. He showed a disability sometime before 1930, but reaching his theory required an inference across years of unproven history, and the Court affirmed a directed verdict for the government.⁴⁸ Inference can bridge gaps, but not that one. Short of that edge, a judge may not reject a reasonable inference merely because the judge would draw a different one.

The second direction keeps the motion from becoming a dead letter. Courts have long agreed that the judge considers the evidence favorable to the non-movant, the inferences from it, and all undisputed evidence. The disagreement concerned whether the movant’s testimonial, disputed evidence should count as well.⁴⁹ Under the federal rule the court credits the movant’s evidence to the extent that it is uncontradicted and unimpeached and comes from disinterested witnesses, meaning witnesses with no stake in the outcome.⁵⁰ Each condition removes a reason the jury might rationally disbelieve. A contradicting witness forces the jury to choose between accounts, impeachment gives the jury a reason to discount an account, and a witness with a stake in the outcome may be doubted even when no one rebuts the story. Testimony that clears all three hurdles leaves the jury no rational basis for rejecting it, so crediting it means that no credibility issue exists and does not mean that the court is weighing the evidence.

A few jurisdictions go the other way and decide sufficiency from the opponent’s evidence alone, a practice that makes judgment as a matter of law almost impossible to obtain.⁵¹ Chamberlain shows why that practice fails. Read alone, Bainbridge’s account of a loud crash looks like some evidence of a collision, and its insufficiency became apparent only against the substantial contrary testimony the railroad offered.⁵² A federal judge must therefore hold the movant’s evidence against the non-movant’s in deciding sufficiency, which is a modest form of evaluation that stops short of judging credibility.

A direct conflict in testimony changes the picture, because a direct conflict goes to the jury regardless of how many witnesses stand on each side.⁵³ Had Bainbridge testified that he saw the two strings of cars collide while six defense witnesses denied it, the record would have presented a credibility contest, and a reasonable jury could have believed him. Chamberlain therefore cannot be read to mean that more witnesses win.⁵⁴

The first trial resembles Chamberlain in structure, because the movant’s proof came from sources with no stake in the outcome and nothing in the record answered it. The second presents a head-on conflict with an interested witness on one side, which is the situation the jury exists to resolve.

In a federal court sitting in diversity, the question of whose sufficiency standard applies has two layers. Where a state statute or constitution commits an issue solely to the jury, the answer is clear: state law cannot alter the essential character or function of a federal court, so the federal judge may still grant the motion under the federal standard.⁵⁵ Where the state’s standard for taking a case from the jury differs from the federal one, courts have disagreed. The better view is that the federal test governs even if state law sets another standard, though the matter is not definitely settled.⁵⁶ An answer should apply the federal standard and flag the open question only when the facts supply a state standard that differs.

Burden of Production

The preceding subsections asked how a court judges a record. This one asks who must build it. A party has the burden of production on an issue when, if the evidence is left as it stands, the issue will be concluded against that party as a matter of law.⁵⁸ That is the party Rule 50(a)(1) exposes when it lets the court resolve an issue against a party who has been fully heard.⁵⁹

The burden of production has a companion, the burden of persuasion, and the two blur easily because both usually fall on the party who must prove the point. The party with the burden of persuasion must convince the factfinder, to the required degree of certainty, that it should prevail. In the ordinary civil case that means proving liability by a preponderance of the evidence, while some civil claims demand clear and convincing evidence and a criminal prosecution demands proof beyond a reasonable doubt.⁶⁰ The burden of production is a different duty: to produce enough evidence to avoid an adverse judgment as a matter of law. The judge decides whether production is satisfied, and the jury decides whether persuasion is satisfied.⁶¹ A defendant who moves for judgment as a matter of law therefore does not take on the plaintiff’s task of proving the case. The movant has its own obligation, though, because the motion must specify the judgment sought and the law and facts that entitle the movant to it.⁶²

On the underlying issue, production comes first and persuasion second. A jury resolves only those questions that reasonably can be disputed, so the party with the burden must first offer evidence sufficient to let the jury consider each essential element. If the evidence falls short on any element, the judge resolves the issue and the jury never reaches it.⁶³ Meeting the burden of production wins nothing by itself. It keeps the party from judgment as a matter of law and places the dispute before the jury, which may still rule against the party that produced enough. A defendant faces the same sequence on matters it must prove. In a slander case, for example, the defendant has the burden of persuasion on truth, so it must produce evidence of truth before the jury may weigh it.⁶⁴

The persuasion standard supplies the yardstick for the production question, because a judge cannot say whether a jury could reasonably find for a party without knowing how convinced the jury must be. In the ordinary civil case the judge asks only whether some reasonable juror could find the fact by a preponderance of the evidence.⁶⁵ Where the governing law demands more, the judge’s question rises with it. Anderson approved the view that it is implicit in adopting the clear-and-convincing standard for certain kinds of cases that the judge has a concomitant duty to consider the applicable burden when deciding whether to send a case to the jury. In a libel suit by a public figure, for example, the evidence must be of enough caliber and quantity to allow a rational factfinder to find actual malice by clear and convincing evidence.⁶⁶

The standard of proof therefore moves the line between the questions the judge answers and the questions the jury answers. Criminal practice shows the far end of that movement. A judge can direct an acquittal unless some reasonable juror could find guilt beyond a reasonable doubt, yet because of the defendant’s right to trial by jury, a prosecutor can never obtain a peremptory ruling on the elements of the offense.⁶⁷

May a plaintiff who has produced nothing on an element still reach the jury by hoping the jurors will disbelieve the defendant’s denial? No, and the reason shows what production requires. A party who must prove a fact by a preponderance of the evidence cannot rely solely on the jury’s disbelief of the adversary’s denial of that fact.⁶⁸ In the standard illustration, the plaintiff alleges that the defendant drove through a red light but offers no affirmative evidence of it. The defendant takes the stand and says the light was green, the plaintiff does not cross-examine, and nothing else in the record bears on the signal. The court must direct a verdict for the defendant on that point.⁶⁹ Rejecting the defendant’s testimony would remove the defendant’s evidence without adding anything to the record, so a verdict for the plaintiff would rest on a guess.

The two scenes differ by a single piece of evidence: Penny supplied an affirmative account of the red light, and the first plaintiff did not. That fact moves the issue from the judge to the jury, because a direct conflict in testimony is a credibility question and a judge may not weigh the evidence or make credibility determinations on this motion.⁷⁰ The jury may well believe the defense witnesses and find for Desmond, and that is a verdict the court has no basis to preempt.

The same structure explains why the motion is usually a defendant’s weapon. A plaintiff must produce evidence on every element, so the defendant needs to show a gap on only one, while a plaintiff seeking judgment must show that the record leaves no reasonable alternative on each element.⁷¹ Some states refuse a directed verdict for the party with the burden unless the facts are wholly uncontroverted or the parties agree to them.⁷² How much a court makes of the movant’s burden can also depend on whether the opponent has already presented all of its evidence or may still add to it, which returns the analysis to the fully-heard requirement.⁷³

Relief the Court May Grant

Once the court finds that a reasonable jury would lack a legally sufficient evidentiary basis on an issue, Rule 50(a)(1) says what it may do, in two clauses best read in order. Clause (A) lets the court “resolve the issue against the party.”⁷⁵ Clause (B) supplies the consequence for the lawsuit: the court may grant judgment as a matter of law against the party “on a claim or defense that, under the controlling law, can be maintained or defeated only with a favorable finding on that issue.”⁷⁶ The first clause operates on a single issue, and the second asks what the loss of that issue does to a whole claim or defense. A court can take the first step without taking the second, and the gap between them is where most of the analysis on this motion takes place.

The relief is issue-based, so the court works from the bottom up. It first identifies the issues that the substantive law makes essential to the claim or defense, and then it asks whether the party’s failure of proof on one of them ends the claim or defense. The substantive law does the sorting, because a hole in the proof matters only when the missing finding is necessary to the result. If the claim can be won without the issue, resolving that issue against the party leaves the claim standing for the jury. If the claim cannot be won without it, no verdict for the party could rest on the record. A ruling on this motion therefore carries the elements of the underlying cause of action inside it, because the judge is laying those elements against the evidence.⁷⁷

The same evidence can produce opposite results in different states. Suppose a plaintiff is injured diving into an above-ground pool with aluminum sides and submits evidence of the dangers of the design, of causation, and of damages, but nothing showing that a safer alternative design was feasible. In a state whose law holds a design defective only if a safer alternative was feasible, the missing proof goes to an essential finding, and the court grants the defendant’s motion. In a state that does not require that proof, the identical record leaves the claim intact, and the court denies the motion.⁷⁸ Nothing about the evidence changed between the two courtrooms. What changed was the legal rule that decides which findings the plaintiff must win. Clause (B) reaches defenses as well: a defendant that must prove a release and has offered nothing on it lacks the finding the defense requires, and the defense can be defeated as a matter of law once the defendant has been fully heard.

The motorist’s case shows clause (A) operating by itself. The brake issue is resolved against the plaintiff, and the claim survives because negligence can be proved through any one careless act that caused the harm. Remove the speeding proof, and clause (B) applies to the whole claim, since no remaining finding could support a verdict. A weak issue is therefore not enough to end a claim, and the court asks whether the controlling law makes the issue necessary.

The rule’s wording is permissive as well. When the court finds the proof insufficient, it may resolve the issue and grant judgment, and the rule does not say it must.⁷⁹ A judge who grants the motion and is later reversed has sent the case back for a complete retrial. A judge who defers, submits the case to the jury, and rules on sufficiency after the verdict risks nothing in the same way, because an appellate court that disagrees can remand with instructions to enter judgment on the jury’s original verdict.⁸⁰ For that reason most courts prefer to deny the pre-verdict motion and revisit the question if the jury returns an unsupportable verdict. When the post-verdict motion is joined with a motion for a new trial, Rule 50(c)(1) requires the judge to rule on the judgment motion and also to rule conditionally on the new-trial motion, so that the court of appeals can address both at once.⁸¹ The sufficiency standard is the same before and after the verdict, but the renewed motion runs on its own clock: it must be made no later than 28 days after entry of judgment, or, if it addresses a jury issue not decided by a verdict, no later than 28 days after the jury was discharged.⁸²

Both clauses are open to either side. The defendant is the usual movant, and a verdict may be directed for the plaintiff, though that happens much less often.⁸³