Abstention and When a Federal Court Steps Back

Having established that a federal court has jurisdiction, and having removed a case into it, a party can still find the court declining to decide. Abstention is the collection of doctrines that permit that, and understanding how narrow they are is more useful than memorizing their names.

The starting point

Jurisdiction is to be exercised. The obligation is described as virtually unflagging, and declining requires justification.

Not a general discretion. A federal court cannot decline a case because it would rather a state court heard it.

The doctrines are exceptions. Each addressing a particular circumstance rather than expressing a preference.

And each is narrow. Arguments to extend them are more common than successful.

Which distinguishes abstention from convenience. A transfer for convenience keeps the case in the federal system, as described in forum non conveniens between two states.

The recognized grounds

Unsettled state law bearing on a constitutional question. Where a state court's construction of its own statute might avoid the need for a federal constitutional ruling.

Complex state regulatory schemes. Where a state has established a coherent system with its own review process, and federal interference would disrupt it.

Ongoing state enforcement proceedings. Where a federal ruling would interfere with a pending state prosecution or comparable proceeding.

Exceptional circumstances of duplication. Where parallel state litigation makes a federal case clearly wasteful, on demanding factors examined in two courts hearing related cases.

And certain domestic relations and probate matters. Which federal courts decline as a matter of long-standing practice.

What is not a ground

That state law governs. Federal courts apply state law in diversity cases routinely, as noted in diversity of citizenship explained.

That the state law question is difficult. Difficulty is not the same as the kind of unsettledness the doctrines address.

That a state court is more familiar with the area. Which is true in many cases and is not a reason to decline.

That a parallel state case exists. Mere duplication is not enough without exceptional circumstances.

And that the case is inconvenient. Which is addressed by transfer rather than by abstention.

Certification

An alternative to abstaining. The federal court asks the state's highest court to answer an unsettled question of that state's law.

Available in most states. Under rules or statutes providing for the procedure.

The federal case pauses and resumes. Which preserves the forum while obtaining an authoritative answer.

Reserved for determinative questions. Since the delay can be considerable and is not worth incurring on a peripheral point.

And it removes the need to guess. Which is otherwise what a federal court does when applying unsettled state law, alongside the process in proving another state's law in court.

In practice

Abstention arguments usually fail. Which should inform how much is invested in making one.

Identify the specific doctrine. A general appeal to comity without fitting the circumstances into a recognized category rarely succeeds.

Ask what relief is sought. A stay preserves the federal case; a dismissal does not, and the difference matters for limitation.

Consider certification instead. Where the real problem is an unsettled state law question rather than a need to be in state court.

And watch the timing. Abstention arguments raised late, after substantial federal proceedings, are weaker than those made at the outset.

The doctrines fit together with the rest of this subject as the final answer to a recurring question: when does one court give way to another. Personal jurisdiction asks whether a court may act at all. Convenience asks whether another forum is better suited. Abstention asks whether, everything else being satisfied, the case should nonetheless be left to a different court system.

That the answer is almost always no reflects a deliberate structural choice. Federal courts exist to provide a forum that is available regardless of local conditions, and a broad power to decline would undermine that. The exceptions cluster around situations where federal involvement would actively disrupt something a state is entitled to run for itself: its own enforcement proceedings, its own regulatory schemes, its own domestic relations and probate systems.

For a practitioner, the useful takeaway is to be skeptical of abstention as a strategy. Where the objective is to get a case into state court, the productive arguments are jurisdictional and procedural — attacking removal, identifying a defect, relying on the forum defendant rule — rather than asking a federal court to decline a case it plainly has power to hear.

Where the objective is genuinely to get an authoritative answer on state law, certification is the tool designed for that purpose, and it achieves what abstention arguments are usually reaching for without asking any court to give anything up. It is underused, largely because it is slower, and in cases where a single unsettled state question decides everything it is frequently the most sensible route available.

The practical mechanics of certification are worth knowing because they are simpler than the doctrine surrounding abstention. The federal court frames a question, sets out the facts on which it arises, and transmits it to the state's highest court, which may accept or decline it. If accepted, the parties brief and sometimes argue the question there, an answer issues, and the federal case resumes with the state law settled. The answer binds not only that case but every future one, which is part of the reason state supreme courts generally welcome the procedure.

Timing is the main constraint. A certified question can take many months to answer, and in a case where the parties need a resolution quickly, that delay may be worse than a federal court predicting the state's likely position and moving on. That prediction is what federal courts do routinely and competently in diversity cases, drawing on intermediate appellate authority, the direction of the state's case law and any comparable decisions elsewhere.

Where a party thinks the prediction is likely to go against them, certification becomes attractive for reasons that have little to do with jurisprudence, and courts are alert to that. A certification request framed as a neutral desire for authoritative guidance, made by the side that would lose under the federal court's apparent inclination, is read for what it is.

Both tools — abstention and certification — are ultimately about the same underlying tension. Federal courts regularly decide questions of state law that state courts would decide differently, and there is no clean solution to that. Abstention hands the case over; certification hands over the question; prediction accepts the risk. Each is right in some circumstances, and only the third of them is available quickly enough to matter in most cases.