Summary. More claims die on the calendar than on the merits, and the reason is rarely that a lawyer forgot a date. It is that accrual is harder than it looks, the applicable period is often not the obvious one, and the doctrines that stop the clock are narrower than practitioners assume. This article works through when a claim accrues under the injury and discovery rules, the difference between a limitations period and a repose period and why that distinction defeats tolling arguments, the continuing violation doctrine and its limits after Morgan, and the tolling mechanisms that actually work. It closes with choice-of-law problems, contractual limitations periods, and a diary discipline for a portfolio of claims.
The client's story is clear, the documents are good, and the damages are real. Then someone asks when the contract was breached, and the answer is "sometime in 2019, we think," and the case is over.
That happens more often than any other single cause of loss in civil practice, and it usually happens for one of four reasons:
- The wrong period was applied. The claim was characterized as breach of contract, six years, when the state characterizes claims against professionals as malpractice regardless of the label — two years.
- Accrual was assumed rather than analyzed. The lawyer counted from discovery in a jurisdiction that counts from injury.
- A statute of repose was in play, and every tolling argument that would have worked against a limitations period failed.
- The claim was filed in the right period but in the wrong court, and by the time it was refiled the period had run with no savings statute to catch it.
None of these is exotic. All of them are avoidable with a disciplined intake analysis that takes about an hour.
Two clocks, and why the difference matters more than anything else
A statute of limitations runs from accrual — from when the plaintiff has a complete and present cause of action, or in discovery jurisdictions from when the plaintiff knew or should have known of the injury and its cause. It is subject to tolling, estoppel, and the equitable doctrines described below. Its purpose is to bar stale claims and to protect defendants from having to defend against evidence that has degraded.
A statute of repose runs from a defendant-side event — the sale of a product, substantial completion of construction, the date of the offering, the last culpable act — regardless of whether the plaintiff has been injured yet or knows anything. It creates a substantive right to be free of liability after a fixed period, and it can extinguish a claim before it accrues.
CTS Corp. v. Waldburger, 573 U.S. 1 (2014), is the clearest statement of the distinction. The Court held that CERCLA's provision preempting state limitations periods for environmental claims did not preempt statutes of repose, reasoning that the two are "distinct legal concepts" — limitations periods encourage plaintiffs to be diligent, while repose periods reflect a legislative judgment that a defendant should be free from liability after a defined time, "however unfair" the result may be in an individual case.
Practical consequences of that distinction:
- Equitable tolling generally does not apply to repose periods. Neither does the discovery rule, class action tolling, or fraudulent concealment, unless the statute expressly says otherwise.
- California Public Employees' Retirement System v. ANZ Securities, Inc., 582 U.S. 497 (2017), held that American Pipe class action tolling does not extend the three-year repose period in Securities Act § 13. Class members who wanted to opt out and sue individually had to file within three years of the offering, full stop.
- Where they live: product liability (many states, typically 8–15 years from sale), construction and improvements to real property (typically 6–12 years from substantial completion), medical malpractice, securities (Securities Act § 13, Exchange Act § 1658(b)(2)), and legal malpractice in some states.
Always ask, at intake: is there a repose period, and when did it start? It is the single question most often skipped and the one with the least forgiving answer.
When does a claim accrue?
The default: injury or occurrence
At common law, a claim accrues when the wrong is complete and the plaintiff could first sue. For contract, that is the breach, not the damage. For tort, it is generally the injury. The plaintiff's ignorance is irrelevant.
Federal courts start from this baseline. Rotkiske v. Klemm, 589 U.S. 8 (2019), held that the FDCPA's one-year period runs from the date of the violation, as the text says, and that courts may not graft a general discovery rule onto a statute whose text sets a different trigger. TRW Inc. v. Andrews, 534 U.S. 19 (2001), similarly rejected a general presumption in favor of the discovery rule where Congress had specified an accrual trigger.
The discovery rule
Where it applies — by statute or by state common law — the period runs from when the plaintiff knew or, in the exercise of reasonable diligence, should have known of the injury and, in most formulations, of its cause.
Variations that matter:
- Injury only, versus injury plus causation, versus injury plus causation plus the identity of the wrongdoer. States differ, and the difference can be years.
- Inquiry notice. Once a plaintiff has enough information to prompt a reasonable person to investigate, the clock generally starts, whether or not the plaintiff investigated. "Storm warnings" is the securities-law phrasing.
- Merck & Co. v. Reynolds, 559 U.S. 633 (2010), held that for securities fraud under § 1658(b)(1), the two-year period begins when the plaintiff did discover or a reasonably diligent plaintiff would have discovered the facts constituting the violation, including scienter — a meaningfully later trigger than mere notice of a misstatement.
- Gabelli v. SEC, 568 U.S. 442 (2013), refused to apply the discovery rule to the government's civil penalty claims under 28 U.S.C. § 2462, holding that the five-year period runs from when the fraud occurred. The Court emphasized that the discovery rule exists to protect injured victims, not enforcement agencies whose mission is to root out wrongdoing.
Where the discovery rule is most commonly available: latent injury and disease, professional malpractice, fraud (nearly universally), and claims where the injury is inherently unknowable — undiscovered defects, misappropriated trade secrets, concealed defalcation.
Special accrual rules worth knowing
- Installment obligations. Each missed payment accrues separately, so a plaintiff can recover the payments within the period even if the first default was long ago — unless the obligation was accelerated, in which case the entire debt accrues at acceleration.
- Continuing contracts and open accounts. Accrual may run from the last item or from termination.
- Indemnity and contribution. Generally accrue on payment or on judgment against the indemnitee, not on the underlying loss. This is why an indemnity claim can be timely long after the primary claim is dead.
- Constructive discharge. Green v. Brennan, 578 U.S. 547 (2016), held that the limitations period runs from the date the employee gives notice of resignation, not from the employer's last discriminatory act.
- Legal malpractice. Many states apply a "continuous representation" rule tolling the period while the lawyer keeps representing the client on the matter, and many require actual injury, which for litigation malpractice may not occur until the underlying case concludes.
- Trade secret misappropriation. Both the DTSA and the UTSA treat a continuing misappropriation as a single claim accruing on discovery — expressly rejecting a continuing-violation approach.
Continuing violations
The doctrine is narrower than plaintiffs argue and broader than defendants concede.
National Railroad Passenger Corp. v. Morgan, 536 U.S. 101 (2002), drew the operative line for employment cases and its logic has spread:
- Discrete acts — termination, failure to promote, denial of transfer, refusal to hire — each constitute a separate actionable unlawful employment practice and each starts its own clock. Prior discrete acts outside the period are not actionable, though they remain admissible as background evidence.
- Hostile work environment claims are different in kind, because they are composed of a series of acts that collectively constitute one unlawful practice. If one act contributing to the claim falls within the period, the court may consider the entire scope of the claim, including conduct outside it.
Elsewhere the doctrine appears as:
- Separate accrual for repeated wrongs — each act of infringement, each nuisance, each antitrust overcharge starts a new period as to that act, with recovery limited to the lookback window.
- True continuing violations for an ongoing unlawful condition, such as a continuing trespass or an ongoing failure to accommodate, where the period runs from when the condition ends.
- Rejected for the continuing effects of a single past act. A demotion in 2020 that continues to depress pay in 2026 is a single act with continuing consequences, not a continuing violation.
A statutory correction worth noting. After Ledbetter v. Goodyear Tire & Rubber Co., Congress enacted the Lilly Ledbetter Fair Pay Act, providing that an unlawful compensation practice occurs each time compensation is paid pursuant to a discriminatory decision — codified at 42 U.S.C. § 2000e-5(e)(3). Pay discrimination claims therefore reaccrue with each paycheck, though back pay is limited to two years preceding the charge.
Tolling: the doctrines that actually stop the clock
Equitable tolling
Available where the plaintiff has been pursuing rights diligently and some extraordinary circumstance stood in the way. Irwin v. Department of Veterans Affairs, 498 U.S. 89 (1990), established a rebuttable presumption that equitable tolling applies to suits against the United States as it does to private suits, and Menominee Indian Tribe of Wisconsin v. United States, 577 U.S. 250 (2016), confirmed the two elements are distinct requirements: diligence and an external obstacle beyond the litigant's control.
What courts have accepted: the defendant's affirmative misconduct; a court's misleading order; mental incapacity in some jurisdictions; a timely but defective filing in the wrong forum; and, in narrow circumstances, attorney abandonment.
What courts reject, reliably: attorney negligence (the client is bound by counsel's error), ignorance of the law, ordinary illness, pro se status, and being busy. Equitable tolling is described as available "only sparingly," and courts mean it.
Equitable estoppel and fraudulent concealment
Estoppel bars a defendant from asserting the defense where its own conduct induced the plaintiff to delay — a promise not to plead the statute, settlement negotiations conducted with assurances that suit was unnecessary, or a representation that the claim would be paid.
Fraudulent concealment tolls where the defendant actively concealed the existence of the claim. Most formulations require an affirmative act of concealment beyond the wrong itself, though where the underlying claim sounds in fraud or involves a fiduciary relationship, the concealment element may be satisfied by the failure to disclose.
The distinction from equitable tolling matters: estoppel and concealment focus on the defendant's conduct, and courts are considerably more receptive to them.
Class action tolling
American Pipe & Construction Co. v. Utah, 414 U.S. 538 (1974), held that the filing of a class action tolls the limitations period for all putative class members until certification is denied, so that class members need not file protective individual suits.
Two decisions define its limits:
- China Agritech, Inc. v. Resh, 584 U.S. 732 (2018), held that American Pipe tolling does not permit a putative class member to file a successive class action after certification is denied. Tolling preserves individual claims, not the ability to keep trying to certify.
- ANZ Securities, above, held that it does not apply to statutes of repose.
Whether state courts apply cross-jurisdictional tolling — tolling a state period based on a federal class action, or vice versa — varies significantly. Several states have declined. This is a real trap for a class member relying on a federal case to preserve a state claim.
Statutory tolling
- 28 U.S.C. § 1367(d) tolls the limitations period for supplemental state claims while they are pending in federal court and for thirty days after dismissal. Artis v. District of Columbia, 583 U.S. 71 (2018), resolved a split by holding that "tolled" means the clock is stopped during the federal case, not merely that a thirty-day grace period is added. That is a substantial difference for a claim filed near the end of its period.
- Minority and incapacity. Nearly every state tolls during minority and, usually, during adjudicated incompetency.
- Defendant's absence from the state. Traditional tolling statutes remain on the books, though their application to defendants amenable to long-arm jurisdiction raises dormant Commerce Clause problems after Bendix Autolite Corp. v. Midwesco Enterprises.
- Bankruptcy. The automatic stay tolls actions against the debtor, and 11 U.S.C. § 108 extends deadlines for the trustee — two years for commencing actions the debtor could have brought.
- Military service. The Servicemembers Civil Relief Act tolls periods during active duty.
- Administrative exhaustion. Many statutory schemes toll while a required charge or claim is pending. Others do not, which is why a Title VII plaintiff must file suit within ninety days of the right-to-sue letter and an FTCA plaintiff must present an administrative claim within two years and sue within six months of denial.
Savings statutes
Most states have one: a plaintiff whose timely action was dismissed for reasons other than the merits may refile within a fixed period, often six months to a year, even if the original period has expired. Terms vary sharply — some cover only involuntary dismissals, some exclude dismissals for want of prosecution, and some apply only once. Check the savings statute before agreeing to a voluntary dismissal.
Relation back
An amendment adding a claim or a party after the period has run may be timely if it relates back under Rule 15(c).
New claims relate back if they arise out of the conduct, transaction, or occurrence set out or attempted to be set out in the original pleading. This is applied generously.
New parties are harder. Rule 15(c)(1)(C) requires that, within the Rule 4(m) service period, the new party (i) received notice of the action such that it will not be prejudiced, and (ii) knew or should have known that the action would have been brought against it but for a mistake concerning the proper party's identity.
Krupski v. Costa Crociere S.p.A., 560 U.S. 538 (2010), is the essential case and it corrected a widespread misunderstanding. The inquiry is what the newly named defendant knew or should have known, not what the plaintiff knew or why the plaintiff delayed. A plaintiff who knew of the correct party's existence but misunderstood its role has made a "mistake" within the rule. The Court also emphasized that relation back does not depend on the plaintiff's diligence.
"John Doe" defendants. Most circuits hold that suing a fictitious defendant and later substituting a real one is not a mistake under Rule 15(c)(1)(C) — it is a lack of knowledge — so relation back is unavailable. Some states are more permissive, and Rule 15(c)(1)(A) permits relation back where the applicable state law allows it, which is the route to check in a diversity case.
Choice of law
Limitations periods are traditionally procedural, so the forum's period applies. Two significant qualifications:
- Borrowing statutes. Most states have one, directing that a claim arising in another state is barred if it would be barred there. They vary in whether they look to where the claim "arose," where the plaintiff resided, and whether they apply to resident plaintiffs.
- Substantive characterization. Where a limitations period is built into the statute creating the right — a "built-in" period — it is treated as substantive and travels with the claim. Statutes of repose are generally treated as substantive.
In federal court: Guaranty Trust Co. v. York requires a federal court sitting in diversity to apply the forum state's limitations law, including its borrowing and tolling rules, because applying a different period would be outcome-determinative. For federal claims without an express period, 28 U.S.C. § 1658 supplies a four-year catch-all for claims arising under statutes enacted after December 1, 1990 — and Jones v. R.R. Donnelley & Sons Co., 541 U.S. 369 (2004), held that this includes claims made possible by post-1990 amendments to older statutes. For older statutes with no period, courts borrow the most analogous state period, or occasionally a federal one.
Section 1983 borrows the forum state's personal injury period, with federal law governing accrual — a two-step rule that produces periods ranging from one to six years depending on the state.
Contractual limitations periods and tolling agreements
Shortening by contract. Parties may agree to a shorter period than the statute provides, and courts enforce such provisions if the period is reasonable and the provision is conspicuous. One year is commonly enforced; ninety days is frequently not. Limits:
- Statutes forbid it in some contexts — many insurance codes prescribe a minimum, and consumer protection statutes often void shortened periods.
- Unconscionability and adhesion analysis apply, particularly in consumer and employment contracts, and courts scrutinize shortened periods for claims that are inherently late-discovered.
- UCC § 2-725 sets a four-year period for sale-of-goods claims running from tender of delivery regardless of the buyer's knowledge, except where a warranty explicitly extends to future performance, and permits the parties to reduce it to not less than one year but not to extend it.
- Lengthening is generally not permitted by contract in advance in many states, on the theory that limitations periods serve public interests — though a waiver after the claim accrues is usually enforceable.
Tolling agreements are the standard tool for preserving a claim during negotiation, investigation, or a related proceeding. Draft carefully:
- Define the claims covered. "All claims arising from the transaction" is better than a list that omits something.
- Define the parties covered, including affiliates, successors, and individuals.
- Specify the effect precisely: does the period stop and resume, or is the defense waived for a fixed window? Ambiguity here produces litigation.
- State that the agreement does not revive claims already time-barred, and that neither party admits anything.
- Provide a termination mechanism with notice — typically thirty days — so the plaintiff has time to file.
- Get signatures from every entity that might be sued. A tolling agreement with the parent does not toll against the subsidiary.
Laches and the equitable analogue
Laches — unreasonable delay causing prejudice — is the equitable counterpart, and its scope has narrowed.
Petrella v. Metro-Goldwyn-Mayer, Inc., 572 U.S. 663 (2014), held that laches cannot bar a claim for damages brought within the Copyright Act's three-year window; Congress set the period and courts may not shorten it. SCA Hygiene Products v. First Quality Baby Products, 580 U.S. 328 (2017), extended that reasoning to patent damages under § 286. Laches survives as a defense to equitable relief, and estoppel remains available in both contexts.
Where a federal statute has no limitations period and the claim is equitable — as with some Lanham Act claims for injunctive relief — laches continues to play a substantial role, usually measured against the analogous state period as a presumptive benchmark.
A working diary discipline
For a firm handling a portfolio of matters, the answer to limitations risk is a system, not vigilance.
At intake, in writing, for every matter:
- Every potential claim, listed separately, including alternatives you may not plead.
- The applicable period for each, with a citation to the statute — not to a memory or a chart.
- The accrual rule for each, with a note on whether the jurisdiction applies the discovery rule to that claim type.
- The earliest plausible accrual date and the latest defensible one, with the facts supporting each.
- Any repose period, and the date it started.
- Any conditions precedent: administrative exhaustion, notice of claim to a public entity (frequently 60 to 180 days and jurisdictional), pre-suit expert affidavits in malpractice cases, contractual notice and cure.
- The filing deadline calculated from the earliest plausible accrual date — not the latest. Build in margin.
- Three calendar entries: 180 days, 90 days, and 30 days before that deadline, assigned to a named person.
Rules for the system:
- Never rely on a tolling theory to set the deadline. File within the period and argue tolling only if you must.
- Never rely on relation back to set the deadline. Name the right defendants the first time; if you are unsure, name all plausible ones and dismiss later.
- Update the diary when facts change, particularly when discovery reveals an earlier date of knowledge than the client reported.
- Treat notice-of-claim deadlines against government entities as the real deadline. They are shorter than any limitations period, they are often jurisdictional, and missing one ends the claim before it starts.
- When declining a matter, tell the prospective client the deadline in writing. A non-engagement letter that identifies the approximate limitations date is the single most valuable risk-management document a firm produces, and its absence is a recurring source of malpractice claims.
From the defense side
Limitations is an affirmative defense under Rule 8(c) and is waived if not pleaded. Plead it in the answer, always, even when the dates look fine — facts change.
Then develop it:
- Interrogatories and requests for admission directed at when the plaintiff first learned of the injury, its cause, and the defendant's role.
- Documents showing earlier knowledge: complaints to the company, internal memoranda, communications with other advisors, prior consultations with counsel, insurance claims.
- Deposition testimony locking down the timeline before the plaintiff appreciates its significance. Ask about the sequence in a neutral, chronological way early in the deposition.
- Third-party records — medical, financial, regulatory — that establish notice.
Move early. Where the dates appear on the face of the complaint, a Rule 12(b)(6) motion is available, since limitations is one of the few affirmative defenses that can be resolved on the pleadings when the facts establishing it are alleged. Where they do not, an early, targeted summary judgment motion on limitations alone is often the most efficient motion in the case — narrow, document-driven, and dispositive.
The uncomfortable but correct policy
Limitations periods bar meritorious claims. That is not a malfunction; it is the design. Evidence degrades, witnesses die and forget, businesses reorganize and lose records, and at some point the cost of defending against a decades-old allegation exceeds the social value of adjudicating it. Repose statutes go further and say that after a fixed time, a manufacturer or a builder should be able to close the file entirely.
Practitioners cannot change that, but they can stop being surprised by it. The discipline is unglamorous: at intake, write down every claim, every period, every accrual rule, and every repose date, and calendar backwards from the earliest defensible deadline. An hour spent on that at the beginning of a matter is worth more than any brief written later about why the clock should not have run.
A field guide to common periods
The following are typical rather than universal, and every one of them must be checked against the specific state's code. They are offered to show the range, and to make the point that the characterization of a claim frequently matters more than the facts.
| Claim | Typical period | Usual trigger |
|---|---|---|
| Written contract | 4–6 years (up to 10 in a few states) | Breach |
| Oral contract | 2–4 years | Breach |
| Sale of goods (UCC) | 4 years, reducible to 1 by agreement | Tender of delivery |
| Personal injury | 1–3 years | Injury, often with discovery |
| Property damage | 2–6 years | Injury |
| Fraud | 2–6 years | Discovery, nearly universally |
| Professional malpractice | 1–3 years, often with a repose backstop | Varies: occurrence, discovery, or termination of representation |
| Defamation | 1–2 years | Publication, with a single-publication rule |
| Statutory claims | As specified | As specified |
| Copyright infringement | 3 years | Each act, with a discovery rule in most circuits |
| Patent damages | 6-year lookback, not a limitations period | Per act |
| Lanham Act | No federal period; laches by analogy | — |
| Securities fraud (10b-5) | 2 years / 5-year repose | Discovery of facts including scienter |
| Securities Act § 11 | 1 year / 3-year repose | Discovery / offering |
| Title VII | 180 or 300 days to charge, then 90 days to sue | Discrete act |
| FLSA | 2 years, 3 if willful | Each paycheck |
| ERISA fiduciary breach | 6 years / 3 years from actual knowledge | Breach or knowledge |
| FTCA | 2 years to present, 6 months to sue after denial | Accrual / denial |
| Section 1983 | Forum's personal injury period | Federal accrual rules |
The characterization problem, concretely. A client sues an accountant for a bad tax position. Is that breach of contract (four to six years), professional negligence (two to three years, possibly with a repose period), or fraud (discovery rule)? Many states hold that the gravamen of the claim controls regardless of how it is pleaded, and will apply the malpractice period to a contract claim against a professional. Others allow the plaintiff's choice of theory to control. That single question can decide the case, and it should be researched at intake rather than discovered in a reply brief.
Multiple defendants, multiple clocks. In a construction defect case the owner may have a contract claim against the general contractor, a negligence claim against the design professional subject to a shorter period and a repose statute, a warranty claim against a product manufacturer under § 2-725 running from delivery, and an indemnity claim against a subcontractor that has not accrued at all. Treating these as one deadline is how a case gets partially dismissed.
Revival statutes and the constitutional limits
Occasionally a legislature decides that a category of claims should be revived after the period has run — most visibly in the child sexual abuse "lookback window" statutes enacted in many states, and in a scattering of environmental and consumer measures.
Whether that works depends on the nature of the claim and the state's constitution. The federal Constitution imposes few limits on the revival of civil claims: the Supreme Court held in Chase Securities Corp. v. Donaldson that a state may revive a time-barred civil action without violating the Due Process Clause, because a limitations defense is not a vested property right. The Ex Post Facto Clause blocks revival of criminal prosecutions, as Stogner v. California confirmed, but it does not reach civil liability.
State constitutions are the real constraint, and they diverge sharply. A number of states hold that once a limitations period has run, the defendant acquires a vested right that the legislature cannot disturb; others permit revival freely; and several distinguish between reviving a limitations bar (sometimes permitted) and a repose bar (usually not, since repose creates a substantive right). The result is that identical lookback statutes have been upheld in some states and struck down in others.
The practical points for anyone advising an institution with historical exposure: revival windows are usually announced well in advance, they generate concentrated filings at the open and close of the window, and the insurance question — which policy year responds to a claim revived decades later, and whether those policies can even be located — is frequently harder than the liability question. Institutions in affected sectors should locate and index historical policies before a window opens, because reconstructing coverage from the 1970s after suit is filed is close to impossible.
One last habit. When a matter closes, record the limitations analysis in the closing memo rather than deleting the diary entries. Related claims surface years later — an indemnity demand, a contribution action, a successor's inquiry — and the single most useful document at that moment is the dated intake sheet showing what the firm knew about accrual and when it knew it.
Related articles
- Res Judicata, Collateral Estoppel, and the Preclusive Effect of Judgments — the other doctrine that ends cases before the merits.
- Evaluating a New Civil Case — the intake analysis this fits inside.
- Civil Procedure Toolkit: Pleadings, Jurisdiction, Preclusion, and Deadlines — the procedural roadmap.
- Drafting a Complaint That Survives a Motion to Dismiss — pleading around a limitations problem.
- Choice of Law, Forum Selection, and Where Your Dispute Will Be Decided — borrowing statutes and which clock you get.
- Understanding Equitable Defenses — laches, acquiescence, and estoppel in detail.
- The UCC Article 2 Sale of Goods — § 2-725 and warranty timing.
- Sovereign Immunity and Suing the Government — notice-of-claim deadlines that come first.
- Statute of Limitations Diary Checklist — the operational version of the discipline above.
- Professional Malpractice: Standards of Care, Expert Proof, and Defenses — continuous representation and occurrence rules.
This article is provided for general informational purposes and does not constitute legal advice. Limitations periods, accrual rules, tolling doctrines, savings statutes, and notice-of-claim requirements vary substantially by jurisdiction and by claim type. Consult qualified counsel promptly; a delay of days can be dispositive.