Concept
Permitted Purposes Under Rule 408(b)
Compromise offers and negotiations
Permitted Purposes Under Rule 408(b)
Rule 408(a) bars compromise evidence for two uses only: to prove or disprove the validity or amount of a disputed claim, and to impeach by a prior inconsistent statement or a contradiction.1 Rule 408(b) then lets the court admit the same evidence for another purpose and names three: proving a witness's bias or prejudice, negating a contention of undue delay, and proving an effort to obstruct a criminal investigation or prosecution.2 The ban attaches to inferences, not to documents, so one offer or payment can be forbidden proof on one theory and legitimate proof on another. The question to put to any compromise evidence is what the jury must infer from it.
The words “such as” make the list illustrative.3 Any issue that can be proved without reasoning from the offer to the strength or value of the compromised claim avoids the danger the rule guards against, which is that jurors will treat a statement made in negotiations as proof of liability.4 The subsections below take the three named purposes first, because they show the pattern most clearly, then the wider categories courts have admitted, and finally the limits on the whole exception. Before any of that, confirm that the material is compromise evidence at all: a letter that merely encloses photographs of an accident and invites opposing counsel to discuss the case does not offer or accept consideration in compromising a disputed claim, and one court held it outside the rule.5
Two features keep the exception from swallowing the prohibition. The first is discretion: the court “may” admit the evidence, so a genuine other purpose makes the evidence eligible without making it admissible as of right.6 The second is substance: the other purpose must differ from proving the validity or amount of the claim that was being compromised. The exception covers evidence focused on issues different from the elements of the primary claim in dispute,7 so a proponent whose only route to the point runs through the suggestion that the claim was good, weak, or worth a given sum has named a purpose without finding one. A stem that offers compromise evidence for some reason other than the claim therefore calls for two questions in order: what inference must the jury draw, and is that inference distinct from the validity or amount of the compromised claim? If it is, the judge's discretion decides whether the evidence comes in.
Bias or Prejudice
Bias is any reason, unrelated to the truth of what a witness says, that gives the witness a motive to favor one side: a financial stake, a debt of gratitude, a relationship worth protecting. Jurors who learn of the motive can discount the testimony, which is why the first example in Rule 408(b) is proving a witness's bias or prejudice.8 Settlement evidence fits because a witness whom a party has paid, or promised money, in compromise of the witness's own claim has a financial tie to that party, and the tie may be shown to establish bias.9
The payment comes in because the inference it supports runs through the witness's interest in the payer and never touches the merits of any claim. The same $25,000 could be offered for two inferences. One is that the company paid because it knew its van was at fault; that reasons from an offer to the validity of a claim, which subdivision (a) forbids.10 The other is that a passenger who took the company's money may want to keep its goodwill. The second inference holds whether her claim was strong or weak and whether the sum was generous or stingy, because a large payment for a thin claim creates the same tie as a small payment for a solid one. Excluding the payment would let a party buy favorable testimony through a settlement and then keep the arrangement from the jury, so the rule would shelter the very thing cross-examination exists to expose. The settlement policy loses little, because the evidence says something about the witness and nothing about what the payer thought of its own liability.
An objector may reply that a point scored on credibility helps win the case, which brings the evidence back under the prohibition. The Second Circuit rejected that kind of argument in PRL USA Holdings, Inc. v. United States Polo Association, Inc. Every item of evidence a party offers is meant to help establish its own position and defeat its adversary's, so reading the rule to bar evidence offered to defeat the opponent's position would forbid the very examples the rule gives and leave the exception with nothing to do.11 The exception exempts instead evidence focused on issues different from the elements of the primary claim in dispute, so the useful question is which issue the evidence addresses.12
Bias is the permitted route; the two forms of impeachment named in subdivision (a) are the barred ones. A stem that gives a witness both a payment and a remark from the talks therefore calls for two rulings.
Both items come from one negotiation and concern one witness, and the rulings differ because the inferences differ. The payment shows an interest that exists whatever was said at the table. The remark helps only if jurors treat the version at the table as the true account and the testimony as the false one, which is the use the rule now bars expressly: Rule 408(a) was amended effective December 1, 2006 to prohibit using conduct or statements made in settlement negotiations to impeach by a prior inconsistent statement or a contradiction.13 Commentary explains why. Using inconsistent statements from negotiations for general impeachment is fraught with danger that jurors will take them as proof of liability, and it threatens the frank exchange of information that negotiation depends on.14 The rule leaves other forms of impeachment untouched, including a witness's bias or prejudice.15
Admission for bias remains discretionary, because subdivision (b) says the court may admit the evidence and does not say it must.16 Since the payment is admissible for one purpose and barred for another, the judge weighs its probative value and the need for it against the danger of unfair prejudice and of misleading the jury, and considers how well a limiting instruction would work.17 The danger is concrete: jurors who hear that the company paid the passenger may slide from “she has a reason to favor the company” to “the company must have been at fault.” A limiting instruction tells them to use the payment only in judging her credibility, and the judge decides whether they can be expected to follow it. When the witness is central and the tie is direct, the balance leans toward admission. When the payment is remote or the bias can be shown another way, the risk may outweigh the gain. A later subsection develops this discretion in full.
Two opposite errors recur on this topic. One treats the payment as excluded automatically because it is settlement evidence, which ignores the bias example in subdivision (b). The other treats bias as a pass that carries the witness's negotiation statements in with it, which ignores the impeachment bar in subdivision (a).
Negating Undue Delay
Undue delay is the charge that a claimant waited too long to act, to the other side's detriment; a defendant might argue, for example, that the plaintiff forfeited a claim by delaying to bring it.19 Settlement talks bear on that charge because a party who is negotiating is often not suing, so the reason time passed may be that the two sides were talking. Rule 408(b) therefore lets the court admit compromise evidence to negate a contention of undue delay,20 and commentary on the rule agrees that evidence of settlement negotiations is admissible to prove lack of undue delay and that the rule does not call for exclusion when negotiations explain delay in taking action.21
This use sits outside the prohibition because a delay contention puts a different question to the jury. Subdivision (a) forbids reasoning from compromise evidence to the validity or amount of the claim,22 and negotiation history offered on delay does not show whether the claim is good or what it is worth; it shows why time passed. In PRL USA Holdings, Inc. v. United States Polo Association, Inc., the Second Circuit described the delay example as very like estoppel by acquiescence, a defense whose issues differ from the elements of the plaintiff's claim, and it read the exception to reach evidence focused on such issues.23 The court also answered the objection that defeating a delay defense ultimately helps the plaintiff. A plaintiff who offers the defendant's conduct in negotiations to negate a delay contention does seek to establish its claim by defeating an affirmative defense, but every item of trial evidence does something like that, and if that were enough to count as proving validity, the rule's own examples would be forbidden.24
Fairness is clearest when the other side invited the wait. Bankcard America, Inc. v. Universal Bancard Systems, Inc. makes the point in terms of lulling: the Seventh Circuit treated it as an abuse of Rule 408 to let a party lull its opponent during compromise talks and then block proof of what the lulling led the opponent to do. The mechanism is easy to trace. If the rule shielded the negotiation history, a defendant could ask a claimant to hold off, accept the patience requested, and then plead the wait as a defense, with the one fact that answers the defense locked out of the courtroom. Claimants who saw that risk would refuse every request for time, and refusing to talk undermines the settlement policy the exclusion exists to serve.25 Admission costs that policy little, because the proponent needs only the fact that talks were under way and that the other side asked for time.
An exam stem about a late-filed claim and an earlier exchange of offers calls for two checks. First, ask whether anyone is actually contending that the claimant delayed, because without that contention the negotiation history has no work to do except the forbidden work of showing the claim's strength or value. Second, ask which fact the proponent needs from the talks. A request to wait, or an assurance that suit was unnecessary, answers the delay question; the size of an offer or an admission of liability answers only the merits. Even a proper delay use leaves the decision to the judge, since the court “may” admit the evidence and can weigh the need for it against the risk that jurors will treat it as proof of the claim.26
Obstructing a Criminal Investigation
To obstruct, in this setting, means to try to stop or derail the government's case, most often by paying or pressuring the victim or a witness so that the case goes away. Evidence offered to prove that effort shows what the speaker tried to do to the prosecution; it does not ask the jury to decide whether a civil claim was good or what it was worth. Rule 408(b) accordingly lists “proving an effort to obstruct a criminal investigation or prosecution” as its third example,28 and neither inference that subdivision (a) forbids is in play.29 Obstruction is the sharpest case of a theme running through this chapter: one offer of money can be protected on one theory and legitimate proof on another, depending on what the money was meant to accomplish.
Commentary on the rule states the result directly: using evidence of an effort to settle a civil case is not barred when offered to prove an effort to obstruct a criminal prosecution.30 The reason appears when you ask what the exclusion buys. Rule 408 gives up some relevant evidence so that parties will resolve disputes without trial, and a person who pays a victim to withhold cooperation from a prosecutor resolves no dispute. Excluding the payment would purchase none of the candor the policy is after, and a defendant who could stamp “settlement offer” on a payment to a witness would keep the obstruction itself from the jury. Evidence of negotiations is therefore admissible to show that a party sought to obstruct a criminal investigation, for example by buying someone off to avoid criminal liability.31
United States v. Davis marks the other side of the line. Davis asked Hammock what it would take to “make this go away.” Hammock said to pay back what he took; Davis countered with an offer to split $29,000 in checks made out to cash that he thought the fraternity had discovered; Hammock countered that the missing funds totaled more than $100,000.32 That is haggling over a private claim, and Hammock himself understood it as a compromise negotiation. The D.C. Circuit said there could be no doubt that Davis offered to compromise a disputed claim, and it held the exchange inadmissible for the use the government chose, which was consciousness of guilt and not obstruction, because that use proves liability, the very thing Rule 408 keeps from the jury.33
The contrast with a payment aimed at a witness explains the result. Obstruction is an act directed at the prosecution or its witnesses, while consciousness of guilt is an inference about the underlying wrongdoing, and the second is the liability use subdivision (a) forbids.34 The usual error is to treat the buy-off language as an invitation to admit any offer that suggests the speaker wanted trouble to end. If a prosecutor could win admission by saying that an offer reveals a guilty mind and then relabeling that guilt an effort to obstruct, subdivision (b) would erase subdivision (a) in every prosecution that grows out of a dispute with a civil side. The proponent must offer the evidence to prove the effort itself and point to facts that show it.
The first payment went to the person whose testimony the case depended on, no claim existed between them for the money to settle, and the proponent offered it to prove the interference. The size of the offer and the criminal setting decided neither scene. This route also operates independently of the carve-out in Rule 408(a)(2), which removes protection only for statements and conduct in negotiations over a claim by a public office exercising regulatory, investigative, or enforcement authority, and only when offered in a criminal case.35 A private complainant can be the target of a buy-off, and an offer to a private claimant stays protected against the guilt inference. Even a proven obstruction purpose makes the evidence eligible and not admissible as of right, since subdivision (b) says the court may admit it, a point the later discussion of discretion develops.
Independent Claims and Defenses
The three named purposes are examples, and courts have carried their reasoning well beyond them. The decisions share one test: the exception covers evidence focused on issues different from the elements of the primary claim in dispute,37 so the question for any proposed use is whether the issue it serves has elements of its own. A fact that settles such an issue can be proved even though it arose at the bargaining table. A fact that matters only by showing the compromised claim to be good, weak, or worth a certain sum stays barred.38 Commentary lists examples in the same spirit: showing the extent of legal services rendered in conducting negotiations, or showing the terms of a compromise in another dispute that are part of the current case.39
The test takes this shape because of what the exclusion protects. Rule 408 spares a negotiator from seeing concessions turned into admissions about the claim under negotiation. An assurance or a wrong at the table is a different kind of event: a new fact with legal consequences of its own, whose significance does not depend on what the speaker thought of the original claim. If the rule hid such events, a party could assure an opponent that no suit would follow, wait while the opponent spent money in reliance, and then invoke the rule to keep the assurance from the jury. The Second Circuit saw that result as an unfair curtailment of a defendant's ability to rely on estoppel, noting that an owner's assurance that it will not sue may itself be characterized as compromise negotiations.40
PRL USA Holdings, Inc. v. United States Polo Association, Inc. applied that test to an assurance given in settlement talks. The defendant, USPA, presented evidence that it had received assurances from PRL about a version of the double horsemen mark, and it offered them to prove estoppel by acquiescence: a defense that bars a trademark owner from claiming infringement after the owner told a potential defendant it would not assert a claim and the recipient relied on that assurance to its substantial detriment.41 PRL objected under Rule 408, arguing that proving estoppel would establish the “invalidity” of its claim.42 The Second Circuit held that the firm prohibition should not apply to a defense that depended on issues distinct from the elements of the infringement claim, and it found those elements distinct.43 The court applied the version of the rule then in force, which spoke of liability for or invalidity of a claim,44 yet the question it asked is the one subdivision (b) still poses: is the evidence aimed at something other than the merits of the compromised claim?
The court's answers to PRL's objections show the test at work. It rejected the invalidity argument because every item of evidence a party offers is meant to help that party and defeat the adversary, so on PRL's reading nothing could ever be offered for another purpose and even the rule's own examples would be forbidden.45 It acknowledged that a defendant could use estoppel pretextually, as a ploy to suggest impermissible inferences from settlement discussions, but found this was not such a case, because the only way USPA could place its entitlement to estoppel in contention was to offer the evidence.46 And it held that any spillover prejudice to PRL, including from counsel's comments implying that PRL believed the marks distinct, did not require a retrial.47
The limit appears in Trebor Sportswear Co. v. The Limited Stores, Inc., which PRL distinguished. In Trebor the purpose for which settlement evidence was arguably permissible, proving compliance with the statute of frauds, overlapped so heavily with the forbidden purpose that it was in effect proving the contract, a necessary element of the plaintiffs' breach claim, so the evidence lay close to the rule's categorical prohibition.48 In PRL, by contrast, the elements of estoppel were distinct from the elements of infringement.49 Logical overlap existed in both cases, since an assurance of no objection might also suggest the owner saw little risk of confusion, so overlap alone does not keep evidence out. What doomed the Trebor documents was overlap in the elements themselves: the plaintiffs could not prove the writing without proving the contract they were suing on. A defense can also serve as a vehicle for the forbidden inference, a risk the later subsections on discretion and on liability in disguise take up.
A wrong committed at the bargaining table supplies a second kind of independent issue. In Carney v. American University, the plaintiff sued for discrimination and also claimed retaliation, alleging that the university withheld extra severance pay because she had signaled that she intended to sue.50 The university answered that all evidence linking the extra pay to her lawsuit lay in inadmissible settlement correspondence.51 The court disagreed. Carney offered the letters not to prove that the university discriminated against her but to show that it committed an entirely separate wrong by conditioning her benefits on a waiver of her rights, so the letters were admissible.52 The opinion supports that holding with a broader proposition from the treatise literature: Rule 408 is inapplicable when a claim is based on a wrong committed in the course of settlement discussions, such as libel, assault, breach of contract, or an unfair labor practice.53 The reason follows from the rule's structure. The letters did not reveal what the university thought of the discrimination claim; they were the alleged retaliation itself, so the forbidden inference was never needed. Shielding them would let an employer punish an employee for threatening suit and then block the only proof by pointing out that the punishment arrived inside a settlement letter.
A heading on a document changes nothing either way. In McAllister Olivarius v. Mermel, an email marked “for settlement purposes only” was received because it was not offered to prove or disprove the validity or amount of a disputed claim or to impeach.
An exam stem in which one party pleads a defense or a second claim built on something the other side said in settlement talks is asking whether that issue has elements of its own.
The fourth step carries most of the weight, because logical overlap is common and decides nothing by itself; what matters is whether the separate issue can be proved without proving the compromised claim. The fifth step restates that admission under subdivision (b) is discretionary.54
Motive, Knowledge, and Explanation
One cluster of other purposes covers what a party wanted, what it knew, and why it acted as it did. Those are facts a jury sometimes must find, and a negotiation can be the clearest record of them. Subdivision (a) bars the inference from a compromise to the validity or amount of a claim, and it bars impeachment by prior inconsistent statement or contradiction,56 so a use that travels neither road sits outside the bar. The decisions share one condition: the proponent can name a contested issue that the evidence serves and whose answer does not depend on whether the compromised claim was sound.
In ESPN v. Office of the Commissioner of Baseball, the question was whether Baseball withheld approval for preemption because it wanted to renegotiate the 1996 Agreement. The court said that if the jury found that to be Baseball's reason, letting Baseball hide its improper act behind Rule 408 would frustrate both the purpose of the rule and the public policy favoring settlement.57 A party's reason for refusing is a fact about that party, and evidence of the bargaining that reveals it proves it however strong or weak the dispute the bargaining was meant to settle. If the rule shielded such evidence, a party could pursue an improper goal at the bargaining table and keep the goal from the jury by calling the bargaining a settlement effort.
The knowledge cases ask what a party believed when it acted. In Bradshaw v. State Farm Mutual Automobile Insurance, the Bradshaws sued State Farm for malicious prosecution after State Farm had sued them over an accident, and the court held that State Farm's payment for the value of a vehicle was relevant to its belief about whether its insured was responsible for the accident.58 Two claims were in play, and the difference between them carries the result. The accident claim was the one compromised, and nobody in the malicious prosecution trial asked the jury to decide whether it was valid. The jury had to decide what State Farm believed when it sued, and the negotiations were direct evidence of that belief. The same negotiations, offered in the accident case to prove that State Farm's insured was at fault, would have been barred.
The explanation cases involve a party who needs the settlement to make sense of events the jury has already heard about. In United States v. Wilford, the defense's questions put the circumstances of certain refund checks in issue, and the Eighth Circuit held that the government could show that the refunds stemmed from the union's settlement stipulation with the National Labor Relations Board. The stipulation was offered to explain why the refunds were paid and not to show that the union violated the labor statute, and it recited that the union admitted no violation. Without it the jury would have seen money flowing from the defendant's union and been left to guess at the reason, so the evidence completed an account that the defense itself had opened. Today the criminal-case carve-out for negotiations over a public office's claim would also bear on such a stipulation.59
Commentary adds a further explanatory use: negotiations offered to explain a failure to seek employment to mitigate damages.60 Kubista v. Romaine illustrates it. The defense argued that the injured plaintiff should have looked for work instead of attending school, and the Washington Supreme Court held that the plaintiff should have been allowed to prove that the defendant's insurance agent encouraged that choice, because the evidence bore on mitigation and liability had already been established by summary judgment. The court added that excluding the evidence would discourage injured people from accepting advance payments, which cuts against the policy the exclusion serves.
A document tied to a different proceeding can also prove a fact that has nothing to do with the settled claim. In United States v. O'Hagan, a financial statement was offered to show that the defendant could not repay borrowed funds at all times, as he maintained. The court found no error in admitting it, because the statement was not offered to prove the validity of any SEC claim and its relevance did not depend on any inference that the SEC, in an entirely different case, might have a valid civil claim against him.61
What ties these decisions together is one question put to the proponent: would the evidence matter to the named issue whether the compromised claim were sound or baseless? The same question accounts for Baseball's motive, State Farm's belief, the union's reason for the refunds, and the defendant's ability to repay. Each also required that the other issue be live: Baseball's reason was disputed, the defense in Wilford had raised the refunds, and mitigation was argued in Kubista.
The fourth step is where a purpose label fails. A state of mind that leads straight back to the claim's amount does not escape the bar, as the last subsection's discussion of Saint Alphonsus v. MRI Associates shows. A purpose that passes the test is still only eligible for admission: Rule 408(b) says the court may admit the evidence, so a weighing of need against the risk of discouraging settlement is part of every ruling, and the next subsection takes it up.62
Discretion and Balancing
The word “may” controls what a proponent gains by proving a legitimate purpose.64 A purpose such as bias, delay, motive, or explanation makes the evidence eligible for admission, and eligibility leaves the decision with the trial judge. The earlier subsections each ended by noting that the court still decides; this one explains how. The exception must be balanced against the policy of encouraging settlements, and the trial court's judgment on that balance is upset only for abuse of discretion.65
The reason for discretion becomes clear when you test each fixed alternative. If a legitimate purpose guaranteed admission, a proponent could name any purpose with a plausible link to the case and put the evidence before the jury with all the force of an admission, and settling parties could never be sure that a figure or a remark would stay out. If a legitimate purpose never justified admission, the uses the rule itself lists would disappear, and a party could pay a witness, or invite an opponent to wait, and then hide the arrangement behind the settlement label. A flexible standard lets the judge look at what a label cannot show: how badly the proponent needs the evidence, whether the same point can be proved another way, how likely jurors are to treat the evidence as proof of the claim, and how far admission would discourage the next round of talks.
The text has long carried that allocation. Before 2006 the rule said that it did not “require” exclusion when evidence was offered for another purpose,66 and the Second Circuit has stressed that although the rule speaks categorically about evidence offered to prove the validity or invalidity of a claim, the terms of the exception are less categorical.67 The present wording, that the court may admit the evidence, keeps the decision in the same place. The same opinion corrected a misreading of Trebor: the earlier court held only that exclusion was not an abuse of discretion, not that exclusion was mandated.68 Discretion therefore cuts both ways, and a judge who admits the evidence and a judge who excludes it may each be right on similar facts.
Ramada Development Co. v. Rauch shows the weighing when the proponent's need is thin. The Fifth Circuit upheld the exclusion of the Goldsmith Report, offered to prove notice, because notice could be effectively proved by means less in conflict with the policy behind Rule 408.69 The record showed what those means were: letters between counsel referring to a list of complaints could have proved the notice issue without prejudice to the opposing party.70 The need was small because a substitute existed, and a real other purpose could not carry the report past that fact.
Fiberglass Insulators v. Dupuy shows the pull of the other side of the scale. The trial judge there concluded that the case arose out of the same transaction, the breakup of a business association, as earlier lawsuits between the same parties.71 The plaintiffs argued that statements made during settlement negotiations were admissible for other purposes even if otherwise excludable. The Fourth Circuit answered that the trial court has discretion on that question, that its decision is reversed only for an abuse amounting to manifest error, and that it found none in holding the statements inadmissible.72 Federal courts generally hold inadmissible the attempted use of a completed compromise of a claim from the same transaction between a third person and a party to the suit,73 and the court saw even more reason to exclude when the settlement or negotiations ran between the same parties to the suit at hand.74 The strong public policy favoring exclusion had been correctly applied.75
PRL shows the same discretion running toward admission: the Second Circuit held that the district court did not abuse its discretion in admitting the settlement evidence and that any spillover prejudice was insufficient to require a retrial.76 The Second Circuit has also sustained a trial judge's discretion to exclude settlement evidence, even when offered for a purpose Rule 408(b) permits, where the issue the evidence addressed was closely intertwined with the validity of the plaintiff's claims.77
The outcomes track the factors a judge weighs. In Ramada an adequate substitute existed, so the report added little that justified its cost. In Fiberglass the negotiations ran between the same adversaries, where admission does the most damage to future talks. In PRL the assurance was the only way to put the estoppel defense in contention, and the elements were distinct, so the balance favored admission.78 The same claim of another purpose therefore produced different results, and the appellate decisions deferred to each trial court's reasonable judgment.
The fifth step draws on rules outside Rule 408. Compromise evidence offered for another purpose is admissible for one use and forbidden for another, so Rule 403 requires the court to compare the evidence's probative value and the need for it with the danger of unfair prejudice and of misleading the jury, and Rule 105 directs attention to how well a limiting instruction would work.79 A limiting instruction tells jurors what the evidence is for, and by implication what it is not for. In Molinos Valle Del Cibao, C. por A. v. Lama, the court told jurors that the evidence was not admitted to establish that the Lamas owed the debt, but that they could use it to decide with whom Molinos had its agreement and whether the Lamas were legally distinct from the corporations involved.80 A proponent who wins admission should argue only the permitted purpose. In PRL, counsel's remarks implying that PRL believed the marks distinct were probably improper, though too minor to warrant a retrial.81
When an instruction may not be enough, a judge can change the structure of the trial. A defense such as estoppel can be pleaded pretextually, as a way to put a forbidden inference before the jury,82 and the PRL opinion observes that a party who feared that evidence of acquiescence might improperly influence the jury on the main issue could have asked the court to consider bifurcation.83 In a split trial the jury decides the main issue without hearing the settlement evidence, and the evidence on the other-purpose issue is presented only if that issue still matters.
The split works because the waiver question arises only after the main question is resolved, so the judge can postpone the risky evidence until it matters. It fits less well when the other-purpose issue must be decided together with the merits, and then the judge falls back on the weighing and the limiting instruction.
Because the weighing belongs to the trial judge, review is deferential in both directions. The Fifth Circuit upset the balance only for abuse of discretion in Ramada, the Fourth Circuit required an abuse amounting to manifest error in Fiberglass, and the Second Circuit found no abuse in admitting the settlement evidence over PRL's objection.84 On an exam, two answers lose points. One says that once the proponent names a purpose listed in subdivision (b) the evidence comes in, which ignores the word “may.” The other says that a ruling to exclude is automatically correct because settlement evidence is barred, which ignores the exception. A strong answer names the purpose, weighs need against risk, notes the available instruction or split trial, and concludes that either ruling is within the judge's discretion.
Liability in Disguise
Rule 408(a) and Rule 408(b) both describe evidence by what it is offered to prove, and neither asks what the proponent chooses to call the purpose.86 A label therefore decides nothing until it is tested against the inference the jury must draw. The test is a single question: does the evidence help the proponent without the jury first accepting that the compromised claim is valid, invalid, or worth a particular sum? If it does not, the purpose is the forbidden one under another name.
The test takes this form because of how the two subdivisions fit together. Subdivision (b) introduces its examples with “such as,” so no closed catalogue of approved labels exists against which a court could check a proponent.87 Words like knowledge, state of mind, good faith, and motive are always available, and a proponent who needed only to recite one could put any offer or admission in front of the jury. Subdivision (a) would then bind only the lawyer who was careless about wording, and every negotiator would know that a figure or a remark could reach the jury under a suitable heading, which restores the very hesitation the rule exists to remove. Looking through the label to the inference closes that gap, because the forbidden inference is the same whatever name it travels under: that a party who offered, paid, or demanded a sum must have thought the claim was good, weak, or worth that amount.
Consciousness of guilt is the plainest disguise. In Davis, the government offered the defendant's proposal to split the missing money to show that he knew he was guilty, and the D.C. Circuit treated the exchange as an offer to compromise a disputed claim and held it inadmissible for that use.88 A proposal to split the missing money reveals a guilty mind only if the jury assumes that people offer to pay when they know they owe, and that assumption is the inference that an offer concedes the claim. Guilt is validity restated for a prosecution. The obstruction purpose differs because of what the evidence is aimed at: payment directed at a witness or prosecutor proves an act of interference whatever any civil claim was worth, while a proposal made to a claimant proves only what the proposer must have believed about that claim.
Belief about value is a second disguise. In Saint Alphonsus Diversified Care, Inc. v. MRI Associates, LLP, the Idaho Supreme Court addressed a ruling that an offer of settlement was admissible not to show value but to show that the other partners really believed their opinion as to value. It called that “a distinction without a difference” and said the rule should be given a broad, not narrow, interpretation to encourage settlement negotiations.89 The partners' belief mattered to the case only as a stand-in for the worth in dispute, so recasting the offer as evidence of belief changed nothing about the inference the jury would draw.
A third disguise runs the evidence through a different claim tied to the compromised one. In Sanders v. Madison Square Garden, L.P., the plaintiff had two analytically distinct claims, sexual harassment and retaliation, and her settlement offer was an offer to settle the harassment claim.90 Rule 408(a) therefore clearly barred any use of her demands to disprove that claim. The defendants argued that they offered the demands on the retaliation claim only, but the court said this cut too fine a line: the demands mattered to retaliation because, according to the defendants, they showed that the harassment claim was made in bad faith, so even limited to retaliation the offer was made for the forbidden purpose of showing that the underlying claim was frivolous.91 Any instruction would have to tell jurors to use the demand against retaliation precisely because it suggests the harassment claim lacks merit, while ignoring that suggestion on the harassment claim itself, a line jurors cannot realistically hold. The Second Circuit has sustained a trial judge's discretion to exclude settlement evidence, even when offered for a purpose Rule 408(b) permits, where the issue the evidence addresses is closely intertwined with the validity of the plaintiff's claims.92
Impeachment supplies a fourth disguise. In a contract dispute between Digh and Fesmire, a master used redacted letters exchanged in compromise negotiations under the guise of impeaching Digh's credibility, in a way that bolstered Fesmire's version of the contract. The reviewing court held that this misused supposed prior inconsistent statements to prove liability and that admitting the letters violated Rule 408.93
The other pole is a purpose with elements of its own, and PRL USA Holdings, Inc. v. United States Polo Association, Inc. explains where the line falls by comparing the case with Trebor. Proving compliance with the statute of frauds by presenting a signed writing was also, in effect, proving the contract, while the elements of estoppel, an assurance and substantial detrimental reliance, were distinct from the elements of infringement.94 Logical overlap with the merits therefore does not turn a purpose into liability in disguise. What matters is whether the proponent can prove an element of the separate issue without first persuading the jury that the compromised claim has merit.
A student who credits a stated purpose because it resembles one of the listed examples has skipped the first two steps. The stronger answer names the inference, shows whether it runs through the merit of the settled claim, and treats an independent issue such as estoppel as admissible subject to the court's discretion. It treats consciousness of guilt, belief about value, and the bad faith of the settled claim as liability in disguise.