Petitioning for Certiorari or Rehearing En Banc: A Practical Guide

Document type: Guide Practice area: Litigation — Appellate Practice Jurisdiction: United States Last reviewed: 5 September 2026

Step one: decide whether to file, and be honest

Before any drafting, answer four questions in writing.

1. Is there a real conflict? Not tension, not different outcomes on different facts — a genuine, outcome-determinative disagreement about the same legal question, ideally acknowledged by one of the courts. If the answer is no, the petition will almost certainly fail, and the client should be told so.

2. Is the question important and recurring? How many cases raise it each year? Can you count them? A question that has produced thirty published decisions in five years is important; one that has produced two is percolating at best.

3. Is this case a clean vehicle? Was the question preserved? Did the court below decide it? Does the judgment rest on that question alone? Is the posture final? Is the record clear? Would the petitioner win if the rule changed?

4. What is the client's actual objective? A grant, a settlement position, a dissent from denial, a GVR, or delay. The last is not a legitimate objective and should be declined.

If you have a portfolio of cases raising the question, the vehicle decides which one to petition in — not the size of the judgment and not the sympathy of the client.

Step two: research the conflict properly

This is the single largest research task and it is done badly more often than not.

Find every decision in every circuit and every state court of last resort that addresses the question. Not the leading cases — all of them. Unpublished decisions matter for showing recurrence even where they are not precedential.

Read them, do not rely on headnotes. A case that a treatise describes as taking one position frequently turns out to have decided something narrower.

Build a chart: court, case, year, holding in the court's own words, facts, whether the decision acknowledges disagreement with another circuit, and whether it has been limited or overruled.

Test the conflict by applying each rule to your facts. If the outcome is the same under both, there is no conflict worth the Court's time, and you have just saved your client the cost of a petition.

Look for the acknowledgment. A court saying "we respectfully disagree with the Ninth Circuit" is the strongest evidence available. Search for it deliberately.

Check whether the conflict has been resolved by statutory amendment, agency rulemaking, or an intervening decision. A petition asserting a conflict that Congress fixed last year is an embarrassment.

Check what the Court has recently denied. If the same question was presented and denied twice in the last three Terms, something is wrong with the way it is being presented, and you should find out what.

Step three: run the vehicle analysis before drafting

List every reason the Court might decline to reach the question, and answer each in writing.

Preservation — with record citations to where the argument was made below.
Whether the court below actually decided it — quoting the opinion.
Alternative grounds supporting the judgment.
Jurisdiction, standing, mootness.
Finality and posture.
Record clarity on the facts the answer would turn on.
Whether the petitioner would win on remand.
Independent and adequate state grounds, for a state court petition.

If a vehicle problem cannot be answered, do not file. File in a different case, or wait for one.

Step four: draft the question presented

Spend more time here than on anything else. Ten drafts is normal.

The rules:

One question. Two only if genuinely inseparable.
Under fifty words; under thirty is better.
Neutral in form. No adjectives, no characterizations of the other side.
Contains the conflict, if there is one: "Whether, as three circuits have held and two have rejected, ..."
Matches what the court below actually decided.
Answerable yes or no.
Comprehensible to someone who has never heard of the case.

Test it. Give it to a colleague who knows nothing about the matter and ask what the case is about. If their answer is not yours, keep drafting.

A useful discipline: write the question you want the Court to answer, then write the question the respondent will say the case actually presents. If those are far apart, you have a vehicle problem you have not yet confronted.

Step five: write the petition in the right order

Structure, in the order it should appear and be written:

Question presented.
Opinions below, jurisdiction, provisions involved. Mechanical, and errors here are unforced.
Statement of the case. Short. Accurate to the point of pedantry. Organized around the legal question, not the equities. Every fact cited to the record or the opinion below.
Reasons for granting:
- A. The circuits are divided. Court by court, holdings quoted, with a chart if it helps. Acknowledge any weakness in the conflict yourself.
- B. The question is important and recurring. Numbers. Consequences. Who else is affected.
- C. This case is an ideal vehicle. Address every problem before the respondent raises it.
- D. The decision below is wrong. Last, and shortest.
Conclusion.
Appendix.

Length. 9,000 words permitted. Aim for 5,000–6,000. A shorter petition that makes one point cleanly beats a longer one that makes six.

Tone. Institutional, not aggrieved. The Court is not being asked to rescue your client; it is being asked to settle the law.

Step six: for the respondent — waive or oppose

Waive when the petition is plainly meritless, the case is unremarkable, and the petitioner is not sophisticated. A waiver signals confidence.

Do not waive when the petitioner is experienced appellate counsel, the conflict claim is superficially plausible, amici have appeared, or the case has attracted attention. A call for a response after a waiver puts the brief under a spotlight.

If a response is called for, treat it as a serious signal. At least one chambers is interested. File the best possible brief, and consider whether the calculus on settlement has changed.

Structure of a brief in opposition:

Counter-statement of the question, if the petitioner's framing is inaccurate.
Statement, correcting material errors precisely and without heat.
Vehicle problems, first. Preservation, alternative grounds, the court below did not decide it, factbound, interlocutory. These are the cheapest grounds for denial.
No conflict. Case by case: distinguishable, superseded, decided on other grounds. Show that the outcome would be the same in every circuit.
Not important. Rare, narrow, superseded, industry adapted.
The decision below is correct. Last, briefly.

Length: aim for half the limit. An opposition that fights everything loses the reader.

Step seven: the reply

A petitioner may file a reply brief, capped at 3,000 words. Most are unnecessary. File one when:

The opposition raises a vehicle problem you did not anticipate and can answer.
The opposition mischaracterizes a case in the conflict, and you can quote the actual holding.
The opposition asserts a factual matter that the record contradicts.

Do not use the reply to repeat the petition, and do not use it to argue the merits at greater length. Two pages that fix one thing is a good reply.

Step eight: coordinate amicus support

Start before the petition is filed. Amicus briefs supporting a petitioner are due within thirty days of the docketing of the petition or the call for a response, and that window is not enough time to recruit from a standing start.

What to look for in amici: genuine independent interest; diversity across industries, regions, or political orientation; the ability to supply information the parties cannot; and counsel who will file on time.

Assign topics. One brief on how often the question arises; one on the practical consequences of the conflict; one on the statutory or historical background; one from states or former officials if available. Duplication wastes the Court's attention and the amici's money.

Practical mechanics: consent from both parties is generally sought and given; the brief must disclose whether party counsel authored it in whole or part and whether any party funded it; and the brief must be filed and served in the required form. Someone must own the tracking.

What not to do: organize a large number of briefs from entities with no independent interest. It is transparent and it is counted against the petitioner.

Step nine: rehearing en banc

Decide first whether you have the intra-circuit conflict ground. FRAP 35 permits en banc rehearing where necessary to secure or maintain uniformity of the court's decisions. If the panel decision cannot be reconciled with an earlier published decision of the same circuit, you have the strongest available argument, because most circuits hold that only the en banc court can overrule prior panel precedent.

Establish it by quotation, not paraphrase. Set the two holdings side by side. If a reasonable reader could reconcile them, the ground fails.

The exceptional importance ground requires showing that the question transcends the case: the number of affected cases, a constitutional question of general application, a circuit conflict, substantial economic or governmental consequences, or the administration of justice in the circuit.

Include the Rule 35 certificate, in the terms the rule specifies. Omitting it is a free error.

Use the dissent. If a panel judge dissented, that dissent is the spine of the petition. Quote it, rely on it, and let it state the problem.

Length: 3,900 words. Use fewer. The petition's audience is the one judge most likely to call for a vote, and that judge will read a tight document.

Consider amicus support, which is permitted at the en banc stage and is most useful when it addresses consequences within the circuit rather than the merits.

Step ten: panel rehearing

FRAP 40 asks for something narrower: what the panel overlooked or misapprehended.

File one when:

The opinion does not address an argument the briefs made.
A controlling authority was not cited.
A material fact is stated incorrectly, with a record citation to prove it.
The holding sweeps beyond the question presented.
The opinion is internally inconsistent.

The narrowing petition is the underrated use. A panel that wrote more broadly than necessary will often revise rather than reverse, and a petition that says "we do not ask the Court to change the result; we ask that the statement at page 14 be limited to the question presented" asks for something the panel can give without conceding error. For a party with other cases, or an industry, that can be worth more than a reversal.

Combine with the en banc petition, but write two genuinely different sections: particular and narrow for the panel; institutional and general for the en banc court.

Step eleven: manage the deadlines

Calendar these the day the opinion issues:

Panel rehearing and en banc: fourteen days in an ordinary civil case; forty-five days where the United States is a party. Extensions must be sought before expiry and are granted sparingly.
Certiorari: ninety days from entry of judgment, or from denial of rehearing if a timely rehearing petition was filed. An untimely rehearing petition does not restart the clock.
Extension application: must be filed at least ten days before the certiorari deadline, addressed to the Circuit Justice, and is not routine.
Amicus deadline: thirty days from docketing of the petition or from a call for a response.
Brief in opposition: thirty days from docketing, unless waived.
Reply: promptly, and in any event to permit distribution.

One person owns the calendar. Missed deadlines in appellate practice are unforgiving and are the most common malpractice claim in the field.

Step twelve: the appendix and the mechanics

Unglamorous, and errors here are noticed.

Contents: the opinion of the court of appeals; the district court opinion or order; any magistrate judge's report adopted; the order denying rehearing; the relevant constitutional, statutory, and regulatory provisions in full; and any other document essential to understanding the question.

Order: reverse chronological is the convention — most recent decision first.

Pagination: sequential, with the appendix cited as "Pet. App. 12a."

Format: the Court's booklet format for paid petitions, with the required cover colours, word count certificate, and proof of service. Use a printer experienced with Supreme Court work; they will catch what you miss.

Electronic filing is required for represented parties, alongside the paper submission.

Step thirteen: read the signals

Call for a response where the respondent waived: at least one chambers is interested.

Relist: something is happening. A single relist materially raises the odds; repeated relists more often precede an opinion respecting the denial.

CVSG: a federal interest is engaged, the case will take many months longer, and the Solicitor General's recommendation is highly predictive.

Hold: a related case is pending; the likely outcome is a GVR.

Distribution without action: rescheduling before a first conference is routine and means little.

What to do with a relist: nothing publicly. Do not file supplemental briefs unless there is genuinely new authority, and if there is, file promptly and briefly under the supplemental brief rule.

Step fourteen: after the denial

Certiorari denied is not a ruling on the merits, and it is worth saying so to the client and, where appropriate, in later litigation. It has no precedential force.

A dissent from denial is an asset. It signals that the question is live and it will be cited in the next petition presenting it.

Preserve the position. The question may return in a better vehicle, and the work done on the conflict chart is reusable.

Tell the client plainly. The odds were always long, the reasons for denial are usually institutional, and the case is over. A candid closing conversation is part of the engagement.

Writing for the actual reader

A petition is read first, and often only, by a law clerk working through a pool of petitions under time pressure, who will summarize it in a memorandum a Justice may read. Everything about the document should be designed for that reader.

What that means concretely:

The first page must contain the case. Question presented on page one; if the reader stops there, they should know what the conflict is and why it matters.
Assume no background knowledge. The reader has not heard of your industry, your statute, or your case.
Make the conflict verifiable in minutes. Quote the holdings. Give pin cites. A reader who has to reconstruct the conflict from your characterizations will conclude there is not one.
Do not bury. If there is a preservation problem, address it where the reader will look for it, not in a footnote on page 24.
Be accurate to a fault. A single mischaracterized case is enough to discount the whole document, and the pool reader will check at least one.
Use headings that state propositions, so that the table of contents alone conveys the argument.
Keep the record light. Facts matter only insofar as they show the question is presented cleanly.
Do not editorialize about the court below. Criticism of the panel reads as grievance and has never persuaded anyone.

A test worth applying. Write the one-paragraph summary you would want the clerk's memorandum to contain. If your petition does not make writing that paragraph easy, restructure it until it does.

The specialist Supreme Court bar, and what it actually does

A small number of lawyers and firms file a disproportionate share of granted petitions, and it is worth being clear about why, because the reasons are learnable rather than mystical.

They select ruthlessly. They decline far more cases than they take, and they take cases with clean vehicles and real conflicts.

They know the landscape. They know what the Court has recently denied, what questions are percolating, which conflicts the Justices have flagged, and which questions the Solicitor General is watching.

They write short. Their petitions are routinely half the permitted length.

They spend disproportionately on the question presented. Days, not hours.

They confront weaknesses. A petition from an experienced practitioner will say "respondent will argue that the conflict is shallow; here is why it is not" rather than hoping nobody notices.

They coordinate amici properly, with topics allocated and deadlines managed.

They are trusted. A petition from counsel known for accuracy is read differently from one that is not, and that credibility is earned over years and lost in a single overstated conflict.

None of that is inaccessible to a lawyer who has not filed before. It is a description of care, honesty, and selection — the same virtues that make any brief good, applied under a much harsher constraint.

Worked example one: choosing the vehicle

Priya Deshmukh's firm has three pending appeals raising the same question about the scope of a federal statutory exemption. All three are likely to be decided against the client's position in circuits that have already rejected it.

Case A. Largest judgment, $80 million. But the district court also found waiver, and the panel affirmed on that ground in the alternative.

Case B. Modest judgment, $2.1 million. The question was preserved cleanly, the panel decided it as a matter of statutory construction, and there is no alternative ground.

Case C. Interlocutory, on an appeal from a certified question. Clean legal issue but the posture is awkward and further proceedings would follow.

Priya petitions in Case B. The client with the $80 million judgment is unhappy, and the conversation is difficult: the case with the most at stake is the case least likely to produce a grant, because the alternative holding means the Court could rule for the petitioner and change nothing.

What she does for Case A instead: files a rehearing petition on the waiver holding, which is the ground she can actually attack; holds the certiorari petition until the Case B petition is resolved; and, if certiorari is granted in Case B, seeks to hold Case A pending the decision.

The outcome. Case B's petition is granted. Case A is held and, after the decision, receives a GVR — the court of appeals reconsiders and, this time, reaches the exemption question. The $80 million judgment is vacated.

The lesson. Vehicle selection is not a technicality. It is the decision that determines whether any of the cases can be helped.

Worked example two: the narrowing petition

Callum Errington's client won the appeal, but the panel opinion contains a paragraph stating a rule far broader than the case required — one that, applied literally, would govern a category of transactions the client's industry engages in constantly.

The problem. The client won. There is nothing to appeal. But the opinion is now circuit precedent and it is worse for the industry than the loss would have been for this client.

What Callum files. A petition for panel rehearing under FRAP 40, stating plainly in its first paragraph that the petitioner does not seek any change in the judgment, and that the panel is asked only to consider whether the statement at pages 17–18 sweeps beyond the question presented.

The body: four pages. It identifies the question the parties briefed, quotes the panel's statement, identifies three categories of transaction that were not before the court and that the statement would govern, and proposes replacement language limited to the facts of the case.

The tone is not adversarial. It reads as what it is: counsel pointing out that an opinion says more than the court likely intended.

The outcome. The panel issues an amended opinion, narrowing the paragraph substantially. No party's rights change.

Why this works. Panels are receptive to being told that an opinion may have unintended reach, because that is a correction rather than a reversal. A petition framed as an attack on the result would have been denied without discussion.

Worked example three: the opposition that ended it

Yusra Al-Hashimi represents a respondent in a well-funded petition supported by six amicus briefs. The petitioner asserts a four-circuit conflict on a question of administrative law.

Her first move is to read the opinion below against the petition, line by line, rather than against her memory of the case. She finds that the panel's holding rests on the agency's interpretation of its own regulation — not on the statutory question the petition presents.

Her brief in opposition is 3,800 words, and it opens with that point: the question presented was not decided below. She quotes the panel's actual holding, twice, and shows that the statutory discussion the petition relies on is a single sentence of dicta.

Second, she addresses the conflict. Two of the four cited decisions involved a different statutory provision; one has been superseded by an amendment; the fourth is consistent with the decision below when read in full. She quotes each.

Third, she notes that the amicus briefs address the importance of the statutory question — which the case does not present.

She does not argue the merits at all, beyond a closing paragraph.

The outcome. Denied at the first conference, without a call for a response and without a relist.

The lesson. The strongest opposition argument is almost always that the case does not present the question, and it is found by reading the opinion below with fresh eyes.

If certiorari is granted

The grant changes everything about the engagement, and the first week matters.

Assemble the team immediately. Merits briefing in the Supreme Court is a different exercise from certiorari-stage practice, and from any other appellate work. If specialist counsel are not already leading, this is the moment.

Reframe. The question presented at the certiorari stage was designed to show why the Court should take the case. The merits brief must show why your client wins. Those are different arguments, and a merits brief that reads like an expanded petition is a common and serious error.

Reconsider the theory. The Court granted for institutional reasons. The rule it adopts will govern thousands of cases, most of them unlike yours. A theory that wins your case but produces absurd results elsewhere will lose, and the strongest merits arguments are those the Court can adopt as a general rule comfortably.

Plan the amicus programme properly. At the merits stage, amici carry substantial weight and the coordination effort is much larger. Allocate: history and text; consequences; alternative rules; the government's position if it is participating.

Prepare for the joint appendix, which has its own rules and deadlines and requires cooperation with the other side.

Moot the argument, repeatedly. Advocates who have argued many times still moot three or four times, with panels of people who will be hostile.

Manage the client's expectations about the remedy. A win in the Supreme Court is frequently a vacatur and remand, not a final judgment. The case may continue for years.

When the answer is not to petition

A guide about how to file should be candid about when not to.

Do not file where the question was not preserved; where the judgment rests on an independent alternative ground; where the asserted conflict is manufactured; where the case is factbound; or where the client's real objective is delay.

Consider instead:

A rehearing petition alone, which is cheaper and, where an intra-circuit conflict exists, has better odds.
Waiting for a better vehicle, particularly for an institutional client or a trade association with a portfolio of cases. The conflict research is reusable and the next case may be clean.
Legislative or regulatory advocacy, where the question is one Congress or an agency could resolve. Some conflicts are better fixed that way, and the Court sometimes says so.
An amicus programme in someone else's case, which is often the most cost-effective way for an industry to influence the law.
Accepting the loss and restructuring around the rule. Businesses do this constantly, and it is frequently the rational answer.

Advising a client not to petition, with reasons, is a service. Filing a petition that had no chance, at substantial cost, because the client was upset and the firm was willing, is not.

Building the conflict chart

The chart is the analytical foundation of the petition and the reusable asset that survives it. Build it properly the first time.

One row per decision, covering every circuit and every state court of last resort that has addressed the question.

Rules for filling it in:

Quote the holding. Not a summary — the court's own sentence, with a pin cite. Every characterization you make will be checked.
Record the facts in one line, because the respondent will argue the cases are distinguishable and you need to know whether that is true.
Search deliberately for acknowledgment. Phrases like "we respectfully disagree," "we decline to follow," "we part company with," and "contrary to the [X] Circuit" are the strongest evidence a petition can have. Run those searches.
Track subsequent history. A decision narrowed by a later panel is not a conflict.
The last column is the test. Apply each rule to your facts. If the result is the same, the conflict is not outcome-determinative and the petition will fail.
Include district court decisions in a separate section — not as evidence of a circuit conflict, which they are not, but as evidence of recurrence and of confusion in the lower courts.
Count. "Thirty-one district courts have addressed the question since 2021, reaching different results depending on the circuit" is a sentence worth having, and it requires the count to be real.

Keep it current. A conflict chart built for one case is the starting point for the next, and for an institutional client with a recurring issue it is a standing asset worth maintaining.

The client conversation, in four parts

Handled well, this conversation takes twenty minutes and prevents a year of misunderstanding.

1. The numbers, plainly. "Fewer than one paid petition in twenty is granted. With a genuine acknowledged conflict on a recurring question in a clean vehicle, the odds are meaningfully better than that, but they are still well short of even. I would rather tell you that now than after."

2. What the document actually argues. "We will not primarily be arguing that the court of appeals was wrong. We will be arguing that the courts have divided, that the question recurs, and that this case presents it cleanly. Whether the decision below was wrong is the shortest section of the brief. That is not because we concede it; it is because that is not what the Court grants on."

3. The outcomes short of a grant, and what they are worth. A dissent from denial that shapes the next case. A relist that improves settlement leverage. A GVR if a related case is decided. A preserved position for a better vehicle. Say which of these is realistically available here.

4. The alternatives. Rehearing alone. Waiting for a better case. An amicus programme in someone else's appeal. Legislative advocacy. Restructuring around the rule. Say which you recommend and why.

On fees. Be specific about scope: petition only, or petition plus reply plus amicus coordination, and what merits briefing would cost if certiorari were granted. Clients who understand the phases are far less likely to be surprised.

And write it down. A short engagement memorandum recording the odds discussed, the vehicle assessment, and the client's decision protects everyone, and it is the document you will be glad to have if the petition is denied and someone asks why it was filed.

En banc in the state systems, and other discretionary doors

The federal model is not the only one, and counsel whose practice crosses systems should note the differences.

State supreme courts generally exercise discretionary review on criteria resembling Rule 10 — conflicts among intermediate appellate districts, questions of first impression, issues of substantial public importance — but the procedures, deadlines, and page limits vary considerably, and several states have distinctive requirements such as a mandatory statement of the basis for jurisdiction or a certification of conflict by the intermediate court itself.

Intermediate appellate rehearing in state systems ranges from a genuine en banc mechanism to a purely panel-level reconsideration. Check whether the court has an en banc procedure at all before drafting a petition premised on one.

Certified questions. Where a federal court of appeals confronts an unsettled question of state law, it may certify the question to the state's highest court, and most states permit it. For a party facing an adverse federal prediction of state law, a certification request is a genuine alternative to rehearing — the state court answers definitively, and the federal court applies the answer. Ask for it in the merits briefing, not for the first time on rehearing.

Interlocutory review. Section 1292(b) certification, Rule 23(f) for class certification orders, the collateral order doctrine, and mandamus are all discretionary doors that arise earlier, and each has its own criteria and short deadlines. The common feature with certiorari and en banc is that all of them are decided on institutional grounds rather than on the merits of the underlying dispute, and all of them reward a petition that understands the difference.

Mandamus deserves a word: it is available where there is no other adequate means to obtain relief, the right to relief is clear and indisputable, and the writ is appropriate under the circumstances. It is not an appeal substitute, and petitions that treat it as one fail. Where it succeeds, it is usually because a district court has done something plainly outside its authority and the harm cannot be repaired on appeal from a final judgment.

A drafting sequence that works

For anyone building a petition from scratch, this order produces a better document than working front to back.

Week 1. Conflict chart. Read every case. Apply each rule to your facts. Decide whether the conflict is real. Vehicle analysis in writing.

Week 2. Draft the question presented — five or six attempts, tested on colleagues who do not know the case. Then draft the reasons-for-granting section, starting with the conflict.

Week 3. Statement of the case, written after the argument so that it contains what the argument needs and nothing else. Vehicle section, written as a rebuttal to the opposition you expect. Merits section, deliberately short.

Week 4. Cut. A first draft that comes in at 8,000 words should leave at 5,500. Read it aloud. Verify every citation and every quotation against the source. Have someone unfamiliar with the case read it and tell you what the question is.

Week 5. Appendix assembly, printer, and filing mechanics. Amicus coordination should have started in week 1 and be concluding now.

Throughout. Keep a running list of every weakness someone raises internally, and make sure each one is either answered in the petition or consciously left alone.

The single most valuable habit in that sequence is the cutting week. Almost every petition improves substantially when a third of it is removed, and almost nobody builds the time for it into the schedule.

The last word

The two doors described in this guide are narrow by design. They are narrow because a system that reviewed everything would review nothing carefully, and because the courts that stand behind them have to spend their limited attention where it does the most good for the law rather than for any particular litigant.

Understanding that is not resignation. It is the whole technique. A lawyer who accepts that the question is "why should this institution spend its scarce capacity here" writes a document aimed at the decision actually being made. A lawyer who cannot let go of "my client was treated unfairly" writes a document aimed at a decision nobody is making.

The petitions that get granted are, almost without exception, the ones where somebody did the unglamorous work: read every case in the conflict rather than the leading three, applied each rule to the facts to check that the conflict mattered, picked the right case out of three, spent ten drafts on fifty words, addressed the vehicle problems before the other side raised them, and then cut a third of what they had written.

None of that requires being one of the two dozen lawyers who do this constantly. It requires being honest about the odds, disciplined about selection, and willing to write something shorter than the rules allow.