Concept
Threshold Requirements
Compromise offers and negotiations
Threshold Requirements
Rule 408(a) does not exclude every remark that sounds like settlement. It bars evidence offered to prove or disprove the validity or amount of a disputed claim, and it protects two kinds of material: furnishing, promising, or offering valuable consideration (or accepting, promising to accept, or offering to accept it) in compromising or attempting to compromise the claim, and conduct or statements made during compromise negotiations about the claim.1 Each description presupposes two facts. A claim must exist that someone disputes, and the offer or statement must belong to a genuine attempt to settle it. Unless both facts are present, the prohibition never operates.
These conditions follow from the reason for the rule. Exclusion is the price the law pays to keep negotiation going, so the rule has work to do only where a negotiation exists to protect. A conversation in which nothing is contested, or in which no one is trying to compromise anything, gives it none. The conditions also close a loophole: if a damaging admission could be sheltered by mentioning a dollar figure in the same breath, every embarrassing statement would arrive wearing a settlement label, and the rule would hide ordinary evidence instead of protecting bargaining. When both conditions are met, courts enforce the exclusion firmly. One court reasoned that a strong public policy favors exclusion, and that the case for it is even stronger for negotiations between the parties to the suit than for a completed compromise between one party and a third person.2
The subsections below take the conditions in order, then ask who decides whether they are met. The answer to the last question is short: the judge, because a jury would have to hear the evidence in order to classify it.
A Disputed Claim
Rule 408(a) protects compromise evidence only when it is offered on the validity or amount of a disputed claim, and that word states the first condition.3 A claim is disputed as to validity when the parties disagree about whether the defendant owes anything, and as to amount when they agree that something is owed but not how much. Either disagreement suffices, so a defendant who concedes fault while contesting the size of the damages is still negotiating a disputed claim. The dispute must be actual: the parties must hold at least an apparent difference of view about validity or amount, and negotiations are the usual sign of it.4 The plain case is an offer to split a contested sum. One court found 'no doubt' that Davis offered to compromise a disputed claim when he proposed to split $29,000 in checks to cash that he thought the fraternity had discovered.5
The reason lies in what the exclusion buys. Rule 408 gives up some relevant evidence to keep parties talking, and that bargain makes sense only when the parties have positions to trade. A debtor who admits both the debt and its amount, then asks for a discount or more time, trades nothing. No policy favors compromising a debt that should be paid in full, so an offer to pay an admitted claim is unprotected, and an offer to pay a lesser sum or to pay in installments is admissible when both validity and amount are undisputed.6 Protecting such offers would also let a debtor confess liability and shelter the confession inside a request for leniency.
A driver who says that he made a mistake, that the damage to the other car comes to about $500, and that he will pay only $200 has admitted responsibility and conceded the amount. A witness may testify to what he said, because the $200 figure asks for a discount on a loss he has already acknowledged.7 In Molinos Valle Del Cibao, C. por A. v. Lama, the debtors acknowledged both the validity and the amount of the debt at discussions they later described as settlement negotiations.8 The district court found that they did not dispute the debt and admitted what was said. The Eleventh Circuit affirmed, treating that finding as a preliminary factual determination reviewable only for clear error.9 The label the debtors chose could not supply a dispute that the substance of the discussions lacked.
The remark about the step is protected because subdivision (a)(2) covers statements made during compromise negotiations, and the negotiation here concerned a claim still disputed as to amount.10 The rule's text names validity or amount, not liability alone. The driver in the earlier example left nothing open on either point.
A claim can also exist without yet being disputed. S.A. Healy Co. v. Milwaukee Metropolitan Sewerage District rejected the argument that a claim is disputed from the moment it is made: 'A dispute arises only when a claim is rejected at the initial or some subsequent level.' Had the sewage authority accepted Healy's claim for a price adjustment, no dispute would have arisen.11 The timing follows from the policy. Until someone rejects a claim, nobody has taken a position that could be compromised, so an early reaction to the claim falls outside any negotiation the exclusion exists to protect.
Because the question is whether a contest exists, a lawsuit is not required. Under Rule 408(a), such evidence is inadmissible provided either liability or amount was disputed, and litigation need not have commenced or even been threatened.12 A filed or threatened suit nonetheless makes the characterization easy. When one party has sued or threatened to sue, later discussions are readily treated as compromise negotiations, while conversations in which litigation was never mentioned tend to be treated as ordinary business give-and-take outside the rule.13
A common error is to treat every statement about money after a mishap as a settlement offer. A dispute, not the mention of money, starts the analysis, and the substance of the exchange decides whether one existed. Calling a conversation a settlement discussion does not create a dispute, as the Molinos debtors learned.14 Omitting words such as 'without prejudice' or 'for purposes of settlement only' costs a genuinely disputed negotiation nothing, because the federal rule requires no such label.15
A Genuine Compromise Effort
Rule 408(a) shields the furnishing or offering of valuable consideration (or its acceptance) in compromising or attempting to compromise the claim, and it shields conduct or statements made during compromise negotiations about the claim.17 Both clauses turn on an effort to compromise, so the second threshold condition asks whether a communication belonged to one. Adversaries talk constantly, and the rule needs a way to pick out the exchanges that count. The federal test asks whether the statements or conduct 'were intended to be part of the negotiations toward compromise.'18
Ramada Development Co. v. Rauch applies that test to a report and not to an offer. The court observed that the rule does not require a pre-trial understanding or agreement between the parties about the nature of the report, and that Rauch's contrary position would partly revive the common-law rule that a party assert that a statement is made 'without prejudice.'19 The present rule, the court explained, fosters free discussion and eliminates the need to decide whether an unqualified statement 'falls within or without the protected area of compromise.'20
Intent decides because a test built on labels fails in both directions. The older approach, which protected hypothetically phrased statements, preferred the sophisticated and trapped the unwary.21 A label rule also invites abuse: if a heading could settle the question, any party could shield an ordinary statement, or an inconvenient admission, by calling a letter a settlement proposal. A heading is evidence of what the writer intended, but nothing in the rule makes it conclusive, and the court looks to the substance and circumstances of the exchange.
The failing side appears in Davidson v. Prince. A man and his wife wrote to the owner of a truck from which a steer had escaped. Their letter recited the facts of the accident, said the writers would not let the owner or the trucking company off with a letter denying responsibility, and told the owner to speak with them directly or leave the matter to lawyers and the courts. When the owner offered the writer's statement from the letter, the court admitted it, reasoning that the letter was an attempt to inform the recipient of the facts and a demand for payment in full, in a tone showing that the writers would not compromise one bit. A recital of facts, a demand at full value, and a threat of suit are the ordinary materials of asserting a claim, and none of them reaches toward the other side.
The claim of protection in United States v. Peed failed for a different reason, even though the speaker there did propose a trade. When Stasko turned to law enforcement over dolls that had gone missing in the mail, Peed told her on a recorded call, 'will you drop the charges?' and said she wanted Stasko's word that everything would be dropped if Stasko wanted the dolls back. The court agreed with the trial judge, who had called the exchange a 'halfhearted suggestion of a settlement,' that Peed was trying to avoid criminal prosecution and not to resolve a civil claim. It pointed to three features: no civil suit was pending, the phrase 'drop the charges' shows concern over criminal prosecution, and nothing followed in the way of an attempt to obtain money or resources for a settlement.
The two cases fail for different reasons, and the difference is the useful part. The Davidson letter contained no movement at all, since the writers offered nothing and asked for everything. Peed's call contained a trade, the dolls for the charges, but the trade aimed at a criminal prosecution and not at a civil claim, so the reason for protecting negotiation was absent. Subdivision (b) separately lists proving an effort to obstruct a criminal investigation or prosecution as a purpose for which the court may admit compromise evidence, and a later section takes that up.22
The two letters differ in one variable, whether the writer reached toward the other side's position. The heading sat on the letter that moved nowhere and was missing from the one that moved. A bare invitation to talk shows the same absence of movement in quieter form. In Hanson v. Waller, counsel's letter enclosed photographs of an accident, stated that the photographs showed the truck driver could not have seen the deceased had she been standing directly in front of the truck, and asked opposing counsel to discuss the case.23 The court agreed with the district court that the letter was not evidence of furnishing or offering, or of accepting or offering to accept, valuable consideration in compromising the claim.24 The letter was no offer of compromise, so it was properly admitted as a statement by a person the party had authorized to speak on the subject.25
Meeting both threshold conditions makes Rule 408 applicable, but it does not end the analysis, because the rule bars only particular uses of the evidence, and renaming the use does not escape it. A master who used redacted settlement letters to 'impeach' a party's credibility, and so bolstered the other side's version of a contract, misused alleged prior inconsistent statements to prove liability, and admitting the letters violated the state's counterpart to Rule 408.26 In the same way, defendants could not offer settlement talks over a harassment claim as relevant only to the plaintiff's retaliation claim when their relevance there depended on showing that the harassment claim was made in bad faith. That is 'precisely the forbidden purpose' of showing that the underlying claim was frivolous.27 The purposes for which a court may admit compromise evidence, such as bias or prejudice, undue delay, and obstruction of a criminal investigation, come later.28
The Judge Decides
The two preceding subsections identified the facts on which protection depends: a claim disputed as to validity or amount, and a communication intended as part of an effort to compromise it. Someone must find those facts before the statement is either kept from the jury or shown to it, and that someone is the judge. A determination whether a statement falls within the protection of Rule 408 is exclusively for the court under Rule 104(a), the provision that assigns preliminary questions about admissibility to the judge.30 The allocation covers both inquiries, so the judge decides whether a dispute existed and whether the statement was a genuine compromise effort, and the jury has no part in either decision.
The jury cannot be given that job because it would have to hear the statement in order to classify it. Under a procedure in which jurors decided, they would listen to the challenged words and the circumstances in which they were spoken, the judge would explain Rule 408, and the jurors would be told to ignore the statement if they concluded it came out of settlement talks.31 Jurors who reached that conclusion would still have heard the statement, and significant words are hard to forget on command. The rule exists to keep the jury from reasoning on the strength of such words, so a procedure that delivers them first would cause the very harm it was meant to prevent. A party who asks that the jury hear the remark first and disregard it if the jurors find it was made in settlement talks has therefore misplaced the decision.
A natural question is why the jury does not decide the settlement issue as it decides other preliminary facts. The practice behind that question is conditional relevance, in which an item matters to the case only if some further fact is true, so the jury hears the item and decides whether that fact exists.32
Rule 408 presents the opposite situation, because the words stay significant whichever way the preliminary fact comes out. A doctor's admission that an operation went wrong keeps its force even if jurors conclude it was spoken during settlement talks, and nothing in human nature stops a juror from relying on significant words after being told to ignore them for reasons of policy.33 The note in the example loses all weight when its condition fails, so the jury can safely find the condition. The Rule 408 exclusion rests on policy and not on any lack of logical force, so the risk that jurors will use the evidence survives a finding in favor of exclusion, and the finding must come before the jury hears the words.
The ruling protects the rule's boundary. Rule 408 requires that a dispute be involved, and compromise negotiations must take place before its protections can be invoked.34 If a speaker's own prediction of litigation could turn words into protected statements, nearly anything said to a person who later sued would be excluded.
The judge also looks at the whole record before ruling. Rix v. General Motors Corp. turned on whether statements by Mr. Ramboldt could be used against GMC, which depended on whether he was GMC's agent acting within the scope of that agency, and on whether compromise negotiations had taken place between him and Mr. Olson. The record was too incomplete for the appellate court to decide, so it directed that, if the question arose at retrial, the district court consider all of the facts before ruling.35 The direction treats the threshold facts, including whether negotiations occurred at all, as matters for the trial court to find, and it shows that a party's bare assertion that the talks were settlement talks decides nothing.
Because the trial judge finds these facts, an appellate court reviews the findings with deference. In Molinos, the district court found that the debtors did not dispute the debt, since they acknowledged that it existed and that they were liable to pay it. The Eleventh Circuit treated that determination as a subsidiary finding of fact under Rule 104(a), which it could overturn only for clear error, and finding none, it held that the district court did not abuse its discretion in admitting the purported settlement negotiations.36 Fiberglass Insulators shows the same deference on the compromise-effort side: after evidence at four hearings, including live testimony, the trial court found that the statements were spoken during settlement discussions, and the reviewing court held those findings not clearly erroneous.37 Even a close call stands. In Platis v. Stockwell, the court called it a close question whether a secretly recorded conversation was a settlement discussion, one in which a speaker described a proposed distribution of a deceased relative's property and debts while the other said his brothers would not accept it.38 It upheld the trial court's exclusion anyway, because evidentiary rulings are presumed correct and reviewed for abuse of discretion.39 A party who loses the classification question at trial therefore faces a demanding standard on appeal, which makes the record built before the trial judge the place where the question is really decided.