Newly Discovered Evidence

Newly Discovered Evidence

Rule 33 names no particular ground for granting a new trial; it asks only whether the interest of justice requires one.¹ The ground courts see most often is evidence that surfaces after the trial has ended. A defendant who invokes it faces two questions that are easy to run together. One is whether the motion succeeds, which depends on the multi-part showing taught in this section. The other is whether the motion was filed in time, and that question also turns on the label, because Rule 33(b)(1) gives any motion grounded on newly discovered evidence its own filing period.² The label therefore selects both the clock and the test, and the next section takes up the clock.

Courts phrase and order these elements a little differently, but the content is stable. In United States v. Slatten the D.C. Circuit listed the same five requirements, placing the cumulative-or-impeaching limit before materiality, and added that trial courts have broad discretion over these motions and that a denial is reviewed only for abuse of discretion.⁴ Because the trial judge applies the test, the movant has to frame the argument there. In United States v. Gibson the trial court denied a Rule 33 motion, filed six days after the verdict, because the defendant had not met the requirements the Ninth Circuit set out in United States v. Brashier.⁵ On appeal the defendant urged a more lenient standard for prompt motions, and the Ninth Circuit declined to consider it because the argument had never been made in the trial court.⁶

The five elements serve one purpose, which is protecting the finality of a verdict reached after a full trial. A conviction carries a presumption of validity, and the defendant bears the burden of overcoming it.⁷ Evidence found after the verdict threatens finality more than a claim of trial error does, because it invites a defendant to reopen a case that was completely tried, and so the movant carries a high burden.⁸ The elements work as a series of screens. The first two ask whether the jury could have heard the evidence if the defense had acted reasonably, the next two ask whether the evidence adds anything the jury lacked, and the last asks whether it would probably change the result. Evidence that passes all of them is evidence the jury could not have heard through no fault of the defendant and that would probably alter the outcome, which is the only kind of post-verdict discovery worth the cost of a second trial. Without the screens, a new trial would be available whenever a defendant could find something more to say. The next three parts teach the elements in groups, and the last part teaches the exception that replaces the whole test when the prosecution held the evidence.

Discovery and Diligence

The first two elements ask one question from two angles: could the jury have heard this evidence if the defense had acted reasonably? The first element requires that the evidence actually was discovered after trial, and the second requires that the failure to find it sooner was not caused by the defendant's lack of diligence, meaning a lack of reasonable effort.⁹ A civil formulation quoted by the Supreme Court in a footnote to Taylor v. Illinois contains the same screen, since it asks for facts discovered since the trial that could not have been found earlier through due diligence.¹⁰ The authorities gathered in that footnote add that a losing party earns no new trial merely by coming upon a witness or information, after the verdict, that the party and counsel did not previously know about.¹¹ Ignorance alone does not satisfy the elements, because the question is why the evidence was not found.

Each element does its own work. The first is a threshold of fact, since a motion cannot rest on material the defense already had, and the word new has to be earned. In United States v. Salameh a defendant moved for a new trial on allegedly newly discovered evidence, and at a conference on the post-trial motions the district judge voiced skepticism and remarked that he did not see anything new.¹² That reaction shows the first element at work: the judge looks at what the motion actually contains and does not accept the label at face value. The second element does subtler work. Without it, a defendant would have an incentive to hold back an available witness or document at trial and keep it in reserve, because an acquittal would end the case and a conviction would supply a second trial in which the held-back evidence could be used after the defense had previewed the government's whole case. The diligence requirement removes that incentive by making the defendant answer for whatever reasonable effort would have uncovered.

The trap on an exam is treating a witness or document the defense knew about but never pursued as new. A fact pattern that has counsel mention the source before trial is signaling lack of diligence, and the motion fails at the second element however strong the evidence looks.

Materiality and Cumulative Evidence

The third and fourth elements ask what the evidence adds to the case the jury already decided. The evidence must be material, and it must not be cumulative or merely impeaching.¹⁴ Material evidence is evidence that bears on an issue in the case. A fact that bears on nothing the jury had to decide cannot change a verdict, so a retrial would reach the same result at the cost of finality, and this element mostly weeds out the irrelevant. The fourth element does more demanding work, because it excludes evidence that is relevant but redundant.

Cumulative evidence is more proof of a point the jury already had before it. A fact can be material and still cumulative: the point matters, but the jury has already heard evidence on it, and one more voice is unlikely to change how the jury weighed it. Impeaching evidence is evidence whose function is to attack a witness's credibility and not to prove or disprove a fact in issue. The reason for excluding it is structural. Nearly every witness can be attacked on something, and the jury already weighed this witness with the tools of cross-examination, so a rule admitting post-verdict impeachment would let almost any conviction resting on testimony be reopened. The word merely marks the limit of the exclusion: evidence that does nothing except repeat or attack is barred, while evidence that supplies a new fact is not. Some jurisdictions state the showing in four parts, folding materiality into the likelihood of a different verdict and excluding evidence that is merely corroborative or collateral as well as cumulative, but the concern behind that wording is the same.¹⁵

In United States v. Reyes a jury convicted the defendant of conspiring to distribute cocaine on testimony that included a DEA informant's account of a street meeting and surveilling agents' observations of lookout activity.¹⁶ After his conviction was affirmed, Reyes moved for a new trial on the strength of developments in later cases that had seriously challenged the credibility of James Hunt, one of the surveilling agents who testified that he saw Reyes acting as a lookout.¹⁷ The Second Circuit held that denying the motion on the merits was within the district court's discretion.¹⁸ The new material attacked the reliability of one witness without supplying any fact placing Reyes elsewhere, which is the kind of showing the fourth element is designed to screen.

Probable Acquittal

The fifth element is usually the one that decides the motion. The defendant must show that the new evidence, if produced at a new trial, would probably lead to an acquittal.²⁰ The D.C. Circuit calls this a high bar and states that trial courts have broad discretion on these motions, which are reviewed on appeal only for abuse of discretion.²¹ The word probably asks for more than a possibility. If a showing that new evidence might raise some doubt were enough, any new evidence on a contested point would qualify, and the element would screen out nothing. The discretion belongs to the trial judge because that judge presided over the trial, saw the witnesses, and is best placed to compare the new evidence with the whole record.

The recantation illustration shows how much room that discretion leaves. A trial witness who implicated the defendant may later say the defendant is innocent. Courts can, and very often do, choose to believe the trial testimony over the recantation, and that decision is routinely upheld on appeal.²² The reason lies in what a recantation asks the judge to do. The question is how a new jury would weigh it, and the judge can reasonably find that a sworn account tested by cross-examination at trial is more reliable than one offered after the verdict by a person who may have new reasons to change the story.

United States v. Slatten applies the element to a recantation-type claim. After the verdict, the prosecution's trial witness Monem gave a victim impact statement for sentencing that appeared to conflict with his trial testimony.²³ One defendant argued that the statement showed the person he was convicted of murdering was alive after the shooting began, and the others argued that it showed Monem had committed perjury and undercut facts vital to the government's case.²⁴ The court stated that when perjury by a prosecution witness is discovered after trial and the prosecution did not know of it until then, the defendant is entitled to a new trial only if the defendant would probably be acquitted on retrial.²⁵ The qualifier about what the prosecution knew foreshadows the next part, where the prosecution's possession of the evidence changes the test. The court was troubled by the government's handling of the witness, which had been an unrecorded call made without defense counsel, yet it held that even if the statements were read most favorably to the defendants as an admission of perjury and a recantation, the district court did not abuse its discretion in declining a new trial.²⁶

The contrast between the two versions isolates the variable that matters. Independent evidence lets the judge predict the result of a new trial without choosing whom to believe, while a recantation leaves that choice, which belongs to the trial judge, at the center of the motion.

Brady Violations

The five-part test governs evidence the defense came across on its own, and a different inquiry governs evidence the prosecution held. If the new evidence was in the prosecutor's possession or control, including with law enforcement working with the prosecutor, the five-part test is not applied, and the failure to reveal the evidence is judged under the Brady test.²⁸ That test, which enforces the prosecution's duty to disclose evidence helpful to the accused, asks two questions: whether the evidence was favorable to the accused, and whether it was material, meaning that if it had been disclosed there was a reasonable probability the outcome would have been different.²⁹ The Brady inquiry has no diligence element, and it asks how disclosure would have affected the original trial.

The exception follows from what the diligence element presupposes. The five-part test assumes the defense could have found the evidence and holds the defense responsible for the gap. When the evidence sat in the government's file, the gap is the prosecution's nondisclosure, and asking whether the defense looked hard enough puts the responsibility on the wrong party. The reach of the rule to law enforcement working with the prosecutor serves a related purpose. Otherwise a prosecutor could avoid the disclosure duty by leaving what investigators learned in the hands of the investigators, so the rule counts the whole prosecution team's possession as the prosecutor's.

The fact that sorts the two versions is who held the evidence. Evidence with the same content is tested under one standard if the prosecution team had it and under a different one if it did not.

A motion that clears these steps has shown that the evidence supports a new trial. Whether the motion was filed in time is a separate question, and the next section teaches it.

Discovery and Diligence

Rule 33(b)(1) gives a defendant three years from the verdict to move for a new trial grounded on newly discovered evidence, but the rule does not say what makes evidence newly discovered or what such evidence must show.³³ Courts have filled that gap with a five-part showing, and because the conviction starts with a presumption of validity, the defendant carries the burden on every part.³⁴ Courts view claims of this kind with caution, so the showing is demanding.³⁵ The first two elements, taught here, ask where the evidence came from and whether the defense could have found it sooner. The remaining three, taken up in the following sections, ask what the evidence is worth.

The first element requires that the evidence was in fact discovered after the trial, and the second requires that the failure to find it sooner was not caused by the defendant's lack of diligence.³⁶ The D.C. Circuit puts the second element as a requirement that the movant acted diligently in attempting to procure the evidence.³⁷ Another common formulation asks whether the evidence could not, by the exercise of diligence, have been discovered before the trial ended.³⁸ The wordings differ, but each tests the same thing: whether the evidence was out of the defense's reach at trial despite reasonable effort, and not merely left unused.

The first element fixes a moment, the end of the trial, and asks whether the defense learned of the evidence only after it. Evidence that the defense held, or knew about, before the verdict could have been offered to the jury, so a motion built on it really challenges choices made at trial, since the record had no gap that anyone was powerless to fill. That is why evidence the defense simply never used cannot be new. A defense that decides at trial not to call a witness, or that never opens a file it already has, has discovered nothing afterward. The movant must show discovery in fact, which in practice means being able to say when and how the evidence surfaced.

The second element looks backward from the discovery to what the defense did before and during trial. Diligence means the reasonable investigative effort a careful defense would make given what it knew and what the case suggested. A defense that investigated sensibly and still missed the evidence satisfies the element, because the question is why the evidence was missed. Since the movant must prove diligence, the motion is strongest when the defense can describe the efforts it made, and weakest when it can only say that it did not know.

The two elements serve different purposes, and the second does the subtler work. The first element ties the motion to evidence the jury could not have seen, so a retrial answers a real gap in the first trial. The second element removes a temptation. A defendant who knows of helpful evidence has a reason to keep it in reserve: if the trial goes well, it was never needed, and if the verdict is guilty, it becomes the basis for a second trial held after the defense has watched the government present its entire case. A thin investigation produces the same result by a quieter route, because skipping the work before trial and doing it after the verdict gives the defense a second chance. The diligence requirement makes the defendant answer for whatever reasonable effort would have uncovered, which keeps the trial as the one full opportunity to present evidence.

In both versions the ledger is identical and surfaces after the verdict. The variable that moves the answer is what the defense knew, or could reasonably have learned, before trial, which is why the date of discovery alone never settles the question. On an exam, the usual mistake is to treat a document or witness as new because it first reached the defense after the verdict. The fact pattern will often include a pretrial detail, such as a note, a tip, or a client's remark, that pointed the defense toward the source. That detail signals lack of diligence, and the motion fails at the second element however persuasive the evidence looks.

Clearing these two elements gets the motion only as far as the next one. In United States v. Slatten, the D.C. Circuit applied the five-part showing to a victim impact statement that a prosecution witness gave while the government prepared for sentencing, after the guilty verdicts, and the defendants offered it as evidence that the witness had committed perjury at trial.³⁹ The court held that the district court did not abuse its discretion in denying a new trial, even viewing the statements in the light most favorable to the defendants, because they had not shown that the statement would probably produce an acquittal, which the court described as a high bar.⁴⁰ The later elements, and especially probable acquittal, are where such motions are commonly lost.

Materiality and Cumulative Evidence

The first two elements of the newly-discovered-evidence showing ask where the evidence came from and whether the defense could have found it sooner. The next two ask what the evidence adds to the case, and they work as screens that sort out weak motions before the court reaches the demanding question of whether the result would probably change. The third element requires that the evidence be material, which the D.C. Circuit states as evidence "material to the issues involved" in the case.⁴² The fourth requires that the evidence not be merely cumulative or impeaching.⁴³ Courts do not always number the elements in this order, and some formulations fold a likely-different-verdict requirement into materiality or add that merely corroborative or collateral evidence does not qualify, but the screening function is the same.⁴⁴

Material evidence is evidence that connects the motion to the verdict. A jury's verdict rests on the facts that the elements of the charged offense make decisive, so a second trial can matter only if the new evidence could move one of those facts. Evidence about a matter no one had to decide, such as a side dispute over a witness's unrelated conduct or a collateral detail in someone's story, cannot change what a jury would find even if it is entirely true. Reopening a verdict for evidence of that sort would produce a second trial that ends where the first one did, at the cost of the finality the conviction carries. The inquiry under this element therefore asks whether the evidence bears on an issue on which guilt depends, whatever its truth or interest.

Evidence is cumulative when it supplies more proof of a point the jury has already heard proved, and corroborative evidence is a close relative, since it only supports proof already in the record.⁴⁵ The exclusion follows from what a retrial is for. A second witness who says what an earlier witness said gives the jury no new fact to weigh, because the jury has already considered that point and either accepted or rejected it. The incentive concern is sharper: after a loss, a defense can nearly always find another voice to repeat favorable testimony, and if repetition sufficed, a verdict could be reopened as often as the defense could recruit another witness. A footnote in Taylor v. Illinois collects decisions stating, among other limits, that a motion for a new trial on newly discovered evidence may not be granted unless the facts will probably change the result and are not merely cumulative or impeaching, and that courts indulge every presumption in favor of the validity of a verdict.⁴⁶

Impeaching evidence is evidence whose only function is to attack a witness's credibility, such as proof of a bias, a prior inconsistent statement, or a record of dishonesty. Courts regard it skeptically because of what it can and cannot do. Impeachment gives the jury a reason to doubt a witness but supplies no affirmative proof that the defendant did not commit the offense, so a verdict supported by other evidence survives even if the witness is entirely discredited. Impeachment material is also easy to come by, since nearly every witness has some history an adversary can use, and a rule that let it reopen verdicts would leave most convictions open to a second trial. The word merely marks the boundary of the exclusion. Evidence that discredits a witness and also proves a fact inconsistent with guilt does more than impeach, so the impeaching label alone does not defeat the motion.

Two decisions show where the line falls. In United States v. Reyes, the defendant sought a new trial on the strength of developments in later cases that, he said, seriously challenged the credibility of a DEA agent who had testified to seeing Reyes act as a lookout, and the Second Circuit held that the denial of the motion on the merits was within the district court's discretion.⁴⁷ The opinion's principal purpose was to settle the date from which the filing period runs, a subject taken up later in this chapter.⁴⁸ The merits ruling still shows the posture of a motion built on new material aimed at a government witness's credibility: the trial judge's denial stands unless it was an abuse of discretion.

United States v. Slatten, discussed above, presented two theories built on one piece of evidence. One defendant offered the victim impact statement as direct proof of innocence, because it indicated that the man he was convicted of killing was alive after the shooting began, while the other defendants offered it to show that a prosecution witness had committed perjury.⁴⁹ The first theory goes to the substance of the government's case. The second works only by discrediting a witness, so it reaches an acquittal indirectly, through the jury's reassessment of that witness. The court resolved that theory at the fifth element, holding that even if the statements were read most favorably to the defendants as an admission of perjury and a recantation, they had not shown that a new trial would probably end in acquittal.⁵⁰

All three versions share the same central issue, the cause of the fire, and the variable that moves the answer is whether the new evidence adds something on that issue that the jury lacked. A report that fills a gap in the proof counts. A repetition adds nothing new, and a credibility attack adds a reason to doubt without adding a fact. Passing these two elements does not win the motion, because the fifth element still asks whether the result would probably change, but failing either one usually ends it. The trap lies in treating any post-verdict material that helps the defense as sufficient. When the facts describe a witness who would say what an earlier witness said, or a document showing only that a government witness has a poor record for honesty, the evidence repeats or only discredits and the motion ordinarily fails at the fourth element.

Probable Acquittal

The fifth element is where most newly-discovered-evidence motions are lost. Once the first four elements establish that the evidence is new, was diligently sought, is material, and is more than cumulative or impeaching, the movant must still prove that the evidence is of such a nature that in a new trial it would probably produce an acquittal.⁵³ The question looks forward and compares: the court asks what a second jury, hearing the old trial proof together with the new, would probably do. A showing that a different verdict is conceivable falls short, because almost any helpful item of evidence could conceivably move some jury. The burden stays on the defendant throughout, and the conviction enters the analysis with a presumption of validity that the motion must overcome.⁵⁴

The standard is demanding for reasons that follow from what a retrial costs and who is best placed to predict its result. A retrial reopens a case the system has already decided, so it is worth ordering only when the new evidence is likely to change the outcome, and the word "probably" ties the remedy to that likelihood. The prediction falls to the trial judge, who watched the trial and can set the new evidence beside the live testimony it would contradict. In United States v. Chinasa, a district court deciding a Rule 33 motion stated that the court may consider all the evidence introduced at trial and may evaluate the credibility of the trial witnesses.⁵⁵ The motion in that case did not rest on new evidence, but the statement describes the court's general power on a new-trial motion, and that power gives the fifth element its practical force. The judge need not ask only whether the new material, taken at face value, would help the defense. The judge may ask how a jury would likely receive it against everything the jury already heard, and because the ruling is discretionary, an appellate court reviews a denial only for abuse of discretion.⁵⁶

The recanting witness shows the consequence most clearly. A witness who implicated the defendant at trial and later says the defendant is innocent has produced the new evidence that sounds strongest, yet courts can, and very often do, choose to believe the trial testimony over the recantation, and that choice is routinely upheld on appeal.⁵⁷ The reason is that two sworn accounts now exist and they cannot both be true. The first was given in open court, under cross-examination, in front of the judge who is now ruling. The second usually arrives after the verdict, outside that setting, so its reliability has never been tested in the same way. A judge who credits the trial account has concluded that a new jury would probably credit it too, which means the recantation does not make acquittal probable, and nothing in the rule requires the judge to treat the later statement as the truer one.

United States v. Slatten shows how far that discretion reaches even when the recantation is assumed to be genuine. The court declined to treat the government's notes of an unrecorded call with its witness as a repudiation of his trial testimony, because nothing allowed anyone to verify what was said, and it voiced concern about the government's decision to hold that call without defense counsel.⁵⁸ It then gave the defendants the benefit of the best reading, treating the statements as an admission of perjury and a recantation, and still found no abuse of discretion in denying a new trial, because the defendants had not shown that the statements would probably produce an acquittal, which the court called "a high bar to cross."⁵⁹ The court also stated that when a prosecution witness's perjury comes to light after trial and the prosecution did not know of it at trial, the defendant must prove he would probably be acquitted on retrial.⁶⁰ Perjury therefore leads to the same fifth-element showing as any other new evidence, so long as the government was unaware of it. What changes when the government itself held the evidence is the subject of the next section.

The two versions differ only in how much of the government's case survives the new evidence, which is why the fifth element has to be measured against the whole trial record. The recording and the recantation also differ in kind. A recording is an objective item the judge can assess without deciding whom to believe, while a recantation forces the judge to choose between two accounts from the same witness, and that choice rests with the judge who heard the first. The usual error on an exam is to argue that the new evidence could change the verdict. The element asks whether it probably would, with the judge entitled to doubt the new witness and to weigh the new material against everything the jury already heard.

Brady Violations

The five-part test applies to evidence that arrived from outside the government: a witness who surfaced late, a laboratory that retested a sample, a recording held by a stranger. A different analysis governs when the evidence was in the government's own hands before trial. If the new evidence was in the prosecutor's possession or control, which includes evidence held by law enforcement working with the prosecutor, the five-part test is not applied.⁶² The failure to reveal that information is judged instead by the Brady test, which asks whether the evidence was favorable to the accused and whether it was material, meaning that, had it been disclosed, there was a reasonable probability that the outcome would have been different.⁶³ The movant therefore stops asking whether the defense searched hard enough or whether acquittal is probable, and answers a question about what the suppression did to the result.

The reason for the switch follows from what each test is built to measure. The diligence element asks whether the defendant's failure to find the evidence earlier came from a lack of effort, and that question makes sense only for evidence a careful defense could have located in neutral hands. A defendant cannot search a government file the defendant cannot see, so when the prosecution team held the evidence, the reason it surfaced late is a disclosure failure by the government rather than a lapse by the defense. There is also an incentive reason. If a withheld item faced the same burden as a stranger's belated statement, a prosecutor who kept favorable material back would risk little, because the defendant would still have to prove that a new jury would probably acquit. In Agurs the Court reasoned that applying the ordinary newly-discovered-evidence standard to evidence the state possessed would leave the prosecutor's obligation to serve the cause of justice with no special significance.⁶⁴

The resulting standard sits between two others. For nondisclosure without a specific request, the materiality standard is stricter than harmless error but more lenient to the defense than the newly-discovered-evidence standard.⁶⁵ The Court later settled on one formulation for the no-request, general-request, and specific-request cases: evidence is material only if there is a reasonable probability that, had it been disclosed, the result of the proceeding would have been different, and a reasonable probability is one sufficient to undermine confidence in the outcome.⁶⁶ That wording looks to the reliability of the verdict that was actually returned. The fifth element of the ordinary test, by contrast, requires the movant to show that a second jury would probably acquit.

The two versions contain the same evidence, and the variable that moves the answer is who held it before trial. That variable controls which test the judge applies, and it often decides the motion, because the Brady showing is the more forgiving one.

The State of Franklin v. Hale performance-test problem shows how this arises in practice. A defendant convicted of attempted murder filed a Rule 33 motion alleging Brady violations, and the court held an evidentiary hearing on them.⁶⁷ The hearing included testimony about a statement a witness had made to a detective, saying the defendant was not the shooter, which the detective testified he assumed had been included in the file sent to the prosecutor's office when he returned from medical leave.⁶⁸ The live issue in that setting is whether information held by police is treated as held by the prosecution, and the prosecutor's response must also ask whether any violation entitles the defendant to a new trial.⁶⁹ The rule text in that exercise is Franklin's invented version, which ties relief to a prejudicial error rather than the federal interest-of-justice standard, so the exercise is useful for its structure while the governing federal standard comes from this chapter.⁷⁰

The exception has a limit. When the claim is that the government lost or destroyed evidence, the usual Brady remedy of a new trial at which the evidence is now available cannot be given, because the evidence no longer exists.⁷¹ The Court therefore imposes two rigorous requirements for those claims: the defendant cannot obtain relief if comparable evidence is reasonably available by other means, and the defense must show bad faith by the police or prosecution in failing to preserve it.⁷² A missing item thus raises a preliminary question of whether it was disclosed late or is gone for good, and the answer decides which standard the defendant must meet.

The usual error is to run every post-trial discovery through the five-part test, or to call every late discovery a Brady violation. The exam facts almost always signal the answer by describing who held the evidence before trial, so that detail should be located first.