A Rule About Timing That Decides Cases Without Deciding Them

The Purcell principle is the doctrine that federal courts should ordinarily refuse to change state election rules close to an election because late changes confuse voters and burden administrators, so a court can leave a challenged rule in force for an election without holding that the rule is lawful. The same restraint shaped Republican National Committee v. Democratic National Committee, 589 U.S. 423 (2020), where the Supreme Court stated that a federal court “should not alter rules on the eve of an election” and stayed an extension of the Wisconsin absentee ballot receipt deadline. Timing, in each dispute, operated as a gate placed before the merits.

The governing claim defended across the sections that follow is that the doctrine exchanges one democratic injury for another. Leaving a rule in force can mean that an election runs under a provision that a fuller hearing might have held unlawful, with votes cast and counted under terms that a later judgment displaces. Ordering a change at a late stage can mean that an election runs under instructions voters had no fair chance to learn, administered by officials who had no fair chance to train staff, print accurate material, or test altered systems. The doctrine ranks the second injury ahead of the first once an election draws near, and that ranking, rather than any view about the wisdom of identification rules, ballot deadlines, or district lines, explains the outcomes in case after case.

What the doctrine governs is a defined set of federal court actions in United States election litigation. It governs applications to stay a lower court order, requests for a preliminary injunction against a state rule, appeals from injunctions already entered, and emergency applications that reach the Supreme Court through a circuit justice. It governs the conduct of federal judges toward state election codes, state constitutions as applied through federal claims, and orders issued by state officials under state law. It does not govern a state acting for itself, a distinction that People Not Politicians v. Onder, decided on September 25, 2026, stated in holding that states remain free to make late changes for themselves even though federal courts should not impose them.

Because the restraint operates before the merits, its effects look paradoxical from outside the procedure. A plaintiff can hold the stronger legal argument and still watch the challenged rule govern the election that prompted the suit. A state can lose on liability in a later merits judgment, as Alabama did in Merrill v. Milligan in June 2023, after an election cycle had already run under the map that the judgment rejected. Each result follows from separating two questions that public argument tends to merge: whether a rule is lawful, and whether a federal court should displace that rule at the moment the request arrives.

The puzzle that gives the doctrine its difficulty, and that the remaining sections work on, is a restraint with no fixed boundary in days. No Purcell decision sets a number of days before an election at which federal intervention closes. Months separated the stay in Merrill v. Evans in February 2022 from the Alabama primaries, while days separated the Wisconsin stay in 2020 from the primary it governed, and both orders rested on the same stated concern. A doctrine that decides real elections, expressed without the one measure that timing would seem to require, places unusual weight on the judgment of the court that happens to hear the emergency application.

What Purcell v. Gonzalez Actually Decided in 2006

Arizona voters approved Proposition 200 in 2004, and the measure required proof of citizenship when a person registered to vote and required identification at the polls. In October 2006, weeks before the midterm election, the Ninth Circuit enjoined parts of the regime without a written explanation that set out the reasoning behind the order. An emergency application then reached the Supreme Court, which decided the matter in Purcell v. Gonzalez, 549 U.S. 1 (2006), as a per curiam opinion, meaning an opinion issued in the name of the Court rather than under the name of an individual justice.

The holding was narrow and procedural. The Supreme Court vacated the Ninth Circuit injunction, so the Arizona rules governed the 2006 election, and the Court faulted the appellate court for intervening without explaining itself at a moment when clarity mattered most. An unexplained order leaves a registrar guessing about which registrations to accept, leaves a poll worker guessing about which documents satisfy the rule, and leaves a reviewing court guessing about which findings the order rested on. Vacatur rested on that failure of explanation joined to the lateness of the intervention, and on the risk that competing instructions would reach voters through county offices, party organisations, and press reports that described different versions of the governing rule.

What the opinion said about voters reached beyond the Arizona dispute. Court orders affecting elections, the per curiam reasoned, can themselves produce the harm that election law exists to prevent, because a voter who learns one identification rule and meets another at the polling place may turn away, may cast a ballot that officials later reject, or may lose confidence that the count treats each ballot alike. Confidence entered the reasoning as an independent concern rather than as decoration. A late judicial change, even a change that corrects a genuine defect, can weaken that appearance by making the rules look improvised under pressure from whichever side filed the stronger emergency papers.

What Purcell did not decide matters as much as what it decided, because later argument often treats the case as a ruling on identification law. The per curiam did not hold that proof of citizenship at registration complied with the National Voter Registration Act, did not hold that identification at the polls complied with the Constitution, and did not resolve the claims that the challengers had pleaded about burden and discrimination. The opinion did not weigh evidence about how many Arizona voters lacked the required documents, did not assess the state interest in the form the rules took, and did not apply the balancing test of Anderson v. Celebrezze, 460 U.S. 780 (1983), or Burdick v. Takushi, 504 U.S. 428 (1992), to the merits.

The mechanism that explains the result is the collision of two rule sets inside the same election. Once an appellate court enjoins a registration or identification regime after officials have published guidance, trained workers, and told voters what to bring, the injunction does not replace one clear rule with another clear rule. It creates a period in which the statute says one thing, the court order says a second thing, the training material says whichever version was printed first, and public discussion repeats both versions with unequal reach. Purcell treated that interval of duplication, rather than either rule standing alone, as the injury that justified vacating the order and restoring the statutory regime for the election at hand.

The Problem the Principle Was Built to Solve

Confusion counts as a legal harm in election cases because voting is an act performed once, under time pressure, by millions of people who receive instructions from sources of uneven quality. A rule change issued by a court reaches a voter through a chain of intermediaries, and each link in that chain can distort, delay, or contradict the message. A county clerk updates a website, a local newspaper summarises the order in a headline, a campaign repeats the version that favours its own supporters, a neighbour repeats an older version learned weeks earlier, and a poll worker applies training received before the order existed.

Voter confusion forms the first of three distinct harms that the doctrine bundles together, and its mechanism runs through learning rather than through administration. Most voters encounter an election rule once or twice in a cycle, absorb a simplified version, and act on that version without checking for revisions. A change in the identification required at the polls, the deadline for a mail ballot to arrive, or the district in which a voter casts a ballot alters a fact that the voter believed settled. Others respond by staying home, by arriving with the wrong document, by mailing a ballot under a deadline that has moved, or by voting in a manner that officials must later set aside.

Administrative burden forms the second harm, and its mechanism runs through systems rather than through individual learning. An election office that receives a late order must revise forms, reprint or discard material already prepared, retrain staff who have already completed training, reprogramme ballot counting equipment, retest that equipment after any change, and answer a surge of public questions with a staff sized for ordinary demand. Each task competes with the others for the same small group of officials in the same compressed days. Purcell reasoning treats that uneven capacity as part of the disruption, not as an excuse that a state must overcome before deserving restraint.

Loss of confidence forms the third harm, and its mechanism runs through perception of the result rather than through any single ballot. An election decided under rules that changed during the contest invites the losing side to attribute defeat to the change itself, whatever the evidence about how many votes the change affected. The concern associated with Bush v. Gore, 531 U.S. 98 (2000), voiced in the concurrence era of that litigation rather than established as a holding of the per curiam, warned against judicial intervention once voting was under way for related reasons about courts displacing the ordinary processes of a count.

The three harms explain why the doctrine speaks in terms of risk rather than proof. A plaintiff asking for a late change can seldom prove in advance how many voters a change would confuse, and a state resisting the change can seldom prove how many voters the existing rule would burden, because both effects occur inside an election that has not yet run. Courts therefore weigh plausible disruption against plausible burden using declarations from officials, the experience of prior cycles, and the visible state of preparation on the ground. The principle resolves ties in that contest of forecasts in favour of the rule already in force, on the stated ground that a known rule, even a contested one, produces a more legible election than a replacement announced while the machinery is already moving.

Election Administration Runs on a Calendar Courts Cannot Stretch

An election office works through a fixed sequence that begins long before any voter sees a ballot, and the sequence has no spare weeks hidden inside it. Candidate filing opens and closes on statutory dates, and the close of filing fixes the names that ballot design must accommodate. Design follows, with contests arranged in a statutory order, district boundaries matched to precinct assignments, and ballot styles multiplied until each combination of districts that a group of voters shares has its own accurate version. A court order altering a district line, a candidate qualification, or a ballot instruction arrives inside that sequence as a command to reopen a stage that the calendar had already closed.

Printing and mailing extend the sequence into a period where delay becomes irreversible. Federal timing law then imposes a hard point in the sequence through the 45 day transmittal rule under the Uniformed and Overseas Citizens Absentee Voting Act as amended, which requires ballots to reach military and overseas voters 45 days before a federal election. A change ordered after that transmittal goes out cannot reach those voters in a corrected form with the same time to return it. The ballot already mailed carries the older district, the older deadline, or the older instruction, and an official choosing between honouring the mailed ballot and enforcing the changed rule faces a dilemma that an earlier order would never have created.

Recruitment and training occupy the middle of the sequence and depend on stability in exactly the rules that litigation most often challenges. A county must recruit enough poll workers for each polling place, assign workers to roles, and train each worker on the identification to request, the circumstances that call for a provisional ballot, the handling of a voter whose name is missing from the register, and the closing procedure for the count. Purcell v. Gonzalez, 549 U.S. 1 (2006), addressed identification rules in Arizona, and the training problem in that setting is concrete: a worker trained to demand one document, then told to accept another set, must apply the second instruction uniformly across a long day or produce unequal treatment between morning and afternoon voters.

Machine preparation adds a technical stage that resists compression for reasons unrelated to law. Voting systems require logic and accuracy testing, in which officials run a known set of marked ballots through the counting equipment and confirm that each contest totals correctly for each ballot style. Early voting then begins while later stages continue, so ballots are being cast under the tested configuration at the same time that a court may be considering an order to alter that configuration. Once early ballots exist, a change no longer affects only future conduct. It creates two populations of ballots cast under different rules inside one election, and canvass officials must later decide whether both populations count on equal terms.

That is why the same order carries such different costs at different points in the sequence. An order issued six months before an election lands before ballot design begins, before printing contracts run, before training material is written, and before any ballot has been mailed or cast, so officials absorb it as one more input into a process that has room for inputs. An order issued six weeks before an election lands after design has closed, after the 45 day transmittal point approaches or passes, after training has started, and after testing has consumed the available days, so each correction displaces another task that the calendar cannot reschedule. The legal content of the two orders can be identical word for word.

A Doctrine With No Number: How Close Is Too Close

No decision in the Purcell line fixes the boundary that the doctrine enforces. Purcell v. Gonzalez, 549 U.S. 1 (2006), involved an injunction entered weeks before a midterm election, and Republican National Committee v. Democratic National Committee, 589 U.S. 423 (2020), involved a stay entered days before a primary with voting under way, yet neither opinion announced a number of days that would govern the next dispute. Merrill v. Evans, the February 2022 Alabama map stay, applied the restraint months before the primaries in that cycle, which extended the practical reach of the doctrine far beyond the eve of an election. At the other end of the range, federal courts have entered orders very close to voting in disputes where the change restored an earlier rule or corrected an official error, and those orders show that proximity alone has never operated as an absolute bar.

In place of a number, courts weigh a set of variables that describe the actual state of the election in front of them. Whether voting has begun carries heavy weight, because ballots already cast under one rule cannot be recalled and recast under another without destroying equality between early and later voters. Whether ballots are printed carries similar weight, since a printed ballot fixes district assignments and instructions in a physical form that correction notices can only partly cure. Restoration matters because voters and officials return to instructions some had already learned, while a different rule requires learning across the whole electorate.

How many days before an election does the Purcell principle start to apply?

No fixed number exists in any Purcell decision. The restraint has no set starting day, and its force grows as voting approaches, reaching its greatest strength after ballots have been cast. Courts weigh disruption, administrative burden, and voter confusion in each dispute rather than counting calendar days alone in isolation.

Responsibility for lateness forms a further variable, and it cuts in both directions depending on which party created the emergency. A plaintiff who held a claim for months and filed weeks before an election weakens the request for urgent relief, because the court can ask why the merits could not have been tested while change was still cheap. A state that enacted a rule, defended it slowly, or produced a map after prolonged delay can face the opposite argument, that official delay manufactured the very proximity the state then invokes as a shield. Justice Ketanji Brown Jackson, dissenting in Scanlan v. Coalition for Open Democracy on October 22, 2026, pressed a version of that objection against a stay entered under two weeks before a midterm election, in a dispute over New Hampshire student identification guidance issued under House Bill 323.

The disruptiveness of the specific change completes the set of variables, and it explains outcomes that a calendar rule would predict badly. Extending a ballot receipt deadline by a stated number of days alters one instruction that officials communicate through notices and that voters apply once, while replacing a congressional map alters the district of every voter in the affected territory, the ballot style for each precinct, and the contest each voter decides. A court can therefore treat a deadline adjustment as absorbable at a stage where a map replacement is not, even with the same number of days remaining.

The absence of a number has a cost that the doctrine’s defenders and critics both recognise, though each side prices the cost differently. The boundary exists in each case only after a court has weighed voting status, printing status, responsibility, and disruption together, and that case by case method is the settled form of a doctrine whose central term the Supreme Court has left undefined across two decades of use.

Richard Hasen Names the Principle and Gives It a Critique

Doctrines in election law often acquire names from the cases that state them, and the Purcell principle followed that path with a scholarly intermediary. The per curiam in Purcell v. Gonzalez, 549 U.S. 1 (2006), stated a warning about late federal court intervention without giving the warning a label that later courts could cite as a unit. Richard L. Hasen supplied the label in Reining in the Purcell Principle, published in the Florida State University Law Review in 2016. Hasen, an election law scholar who had tracked the emergency orders accumulating after 2006, coined the name Purcell principle for the practice of refusing or staying late changes to election rules.

Naming converts scattered orders into a doctrine through a recognisable mechanism. Emergency decisions are short, often unsigned, and issued under time pressure, so each decision standing alone carries limited reasoning that later courts can extend. A label gathers those decisions into a line, signals that the line states a position a court is expected to address, and shifts the burden of argument onto the party asking a court to depart from the gathered practice. A brief filed after 2016 could argue that an injunction would violate the Purcell principle and expect the deciding judge to understand the claim, the supporting cases, and the counterarguments without extended exposition. That shift in argumentative burden is the real work a name performs in law, and it operates whether or not the Supreme Court ever formally adopted the label as its own.

Hasen’s early position, set out in the 2016 article, accepted part of the practice that the name described. Late changes to election rules carry real costs in voter confusion and administrative strain, and Hasen’s argument did not treat those costs as inventions offered by states to defeat meritorious claims. The critical move in the article lay elsewhere. Hasen proposed to discipline the principle with conditions rather than abolish it, so that the restraint would operate as a structured inquiry into closeness, disruption, and justification instead of as a general reluctance that expanded to fill whatever space an emergency application left open. A disciplined version would ask a court to say how the timing harmed administration in the specific jurisdiction, rather than accept lateness alone as a sufficient answer to a claim of unlawful exclusion.

That proposal identifies the fault line that later litigation kept reopening. An undisciplined restraint protects a rule because of when a challenge reached decision, even if the rule would fail on a full record, and each cycle run under such a rule gives the rule a practical legitimacy that its legal status does not support. Merrill v. Evans in February 2022 illustrated the stakes, because the Alabama map stay rested on timing months before the primaries, and Justice Brett Kavanaugh’s concurrence attempted to state conditions under which a late change could still be ordered. The distance between Hasen’s disciplined version and the broader practice visible in the stay orders is the distance between a principle that asks questions and a principle that ends them.

The conversion from scholarly description to working doctrine completed itself when justices used the label in their own opinions. Once the phrase Purcell principle appeared in concurrences and dissents at the Supreme Court, the label no longer belonged to commentary about the Court, since it had become vocabulary inside the Court’s own reasoning about stays and injunctions. Dissents could invoke the principle to describe what a majority had done, concurrences could invoke it to explain what a stay required, and lower courts could cite the justices’ usage as confirmation that the label named a genuine doctrine rather than a critic’s shorthand.

The Emergency Docket Is Where the Principle Lives

The Purcell principle operates inside a procedural route that most election disputes travel at speed, and understanding the route explains much of what the doctrine does. A challenge to an election rule typically begins in a federal district court, where the plaintiff files a complaint together with a request for a preliminary injunction or a temporary restraining order against state officials. A preliminary injunction is an order entered before final judgment that directs a party to act or refrain from acting while the case proceeds, and it governs conduct during the very election that the plaintiff says is unlawful.

The test for that relief sets the frame within which Purcell reasoning works. A party asking for a preliminary injunction or for a stay of a lower court order must ordinarily show a likelihood of success on the merits, a likelihood of irreparable harm without relief, and that the balance of equities and the public interest favour intervention. Purcell operates as a weight placed on the equities and public interest side of that test as an election approaches. Voter confusion and administrative burden enter the analysis as public interest harms that count against intervention, and their weight grows as voting draws nearer, until at some point the equities can defeat a claim that the same court might have found likely to succeed on the merits. The doctrine therefore does not replace the ordinary test. It tilts one part of the test, and the tilt strengthens with proximity in a way the other elements do not.

Why do so many election cases reach the Supreme Court as emergency applications?

Election disputes ripen late because legislatures, officials and map commissions often finish their work close to the cycle the rules will govern, and a challenger cannot sue over a rule that does not yet exist. Once an election is weeks away, only emergency relief can arrive in time, so the emergency application becomes the case.

Emergency applications that reach the Supreme Court travel a defined path from the lower courts. A party dissatisfied with a district court order seeks a stay or vacatur from the court of appeals, and a party dissatisfied there files an emergency application with the circuit justice assigned to the region, who may decide the application alone or refer it to the full Court. Republican National Committee v. Democratic National Committee, 589 U.S. 423 (2020), arrived by that path during a primary conducted in a public health emergency, and Purcell v. Gonzalez, 549 U.S. 1 (2006), arrived by the same path from the Ninth Circuit.

Short reasoning follows from that route as a structural consequence rather than as a choice made case by case. Per curiam and unsigned orders allow the Court to act within the window the election allows, at the price of leaving lower courts, officials, and future litigants to infer the reasoning from the result and from any separate concurrences or dissents that individual justices file. Scanlan v. Coalition for Open Democracy, the October 22, 2026 emergency stay concerning New Hampshire student identification guidance, followed the pattern, with Justice Ketanji Brown Jackson dissenting from a stay whose majority reasoning had to be reconstructed from the posture of the dispute and from the Purcell line of cases behind it.

A stay entered on that route is not a merits judgment, and the distinction carries consequences that public discussion regularly collapses. A stay of an injunction preserves the state rule for the election at hand while leaving the underlying claims alive for a later trial, an appeal after judgment, and a possible merits ruling that bars the rule from future cycles. Merrill v. Milligan, decided in June 2023, proved the point in sequence, since the merits judgment for the Alabama plaintiffs arrived after the 2022 elections had already run under the map that the February 2022 stay had preserved. Holding the two apart is the only way to read a Purcell order accurately, because the order states a judgment about timing and disruption on an emergency record, expressed through the equities test, rather than a judgment about the legality that a full record would test.

What a Challenger Must Show to Win a Late Change

Justice Brett Kavanaugh’s concurrence in Merrill v. Evans, the February 2022 Alabama map stay, restated the working test that governs a challenger seeking a late change to an election rule. Justice Kavanaugh wrote that the Purcell restraint is not absolute, and that a plaintiff can overcome it, but that the showing required grows heavier as an election nears. Three demands structure that showing in his formulation. The claim must be strong on the merits, the threatened harm from leaving the rule in place must be severe, and the change ordered must be feasible to implement without heavy cost, confusion, or trouble for the officials who must carry it out.

Strength on the merits operates as the first mechanism, and it functions as a sorting device between kinds of illegality. A rule that is plainly unconstitutional on controlling precedent presents the easiest case for intervention, because a court asked to leave such a rule in place for an election is asked to knowingly run a contest under a provision that governing law already condemns. A rule whose illegality depends on resolving a contested question, weighing competing evidence about effects, or extending precedent to different facts presents the hardest case, because the emergency court must predict a merits outcome that reasonable judges dispute while also pricing the disruption of acting on that prediction. It asks how visibly good the claim is to a court reviewing at speed.

Severity of harm operates as the second mechanism, and it directs attention to what the challenged rule does to identifiable voters during the single election at stake. A rule that prevents a defined group from casting a ballot that counts inflicts a harm that no later judgment can repair, since an election conducted without those ballots cannot be rerun on equal terms after certification. A rule that imposes an inconvenience curable by an alternative procedure, such as casting a provisional ballot subject to later verification, presents a lesser harm in the terms the test uses, whatever arguments exist about the dignity or equality of the alternative route.

Feasibility operates as the third mechanism, and it converts an abstract legal request into an engineering question put to election officials. A court considering a late map replacement, a deadline extension, or an identification change must know whether ballots can be reprinted, whether counting equipment can be altered and retested, whether poll workers can be retrained, and whether voters can be told of the change through channels that reach them in time. Election officials answer those questions through declarations and testimony that describe printing status, testing schedules, staffing, and the dates on which early voting and mail voting begin. A challenger who wants intervention must engage that evidence directly, by showing that the requested order uses material already prepared, restores a procedure officials used in a prior cycle, or confines its effects to a jurisdiction with capacity to absorb it.

The rising burden ties the three mechanisms together into a single sliding structure. As ballots are printed, as the 45 day transmittal point for military and overseas voters passes, and as early voting opens, each element must be established more convincingly than the element before it had to be at an earlier stage. The same claim that would have justified intervention in spring can fail in autumn without any judge changing a view about the underlying right, because feasibility has collapsed and the confusion cost has risen. Anderson v. Celebrezze, 460 U.S. 780 (1983), and Burdick v. Takushi, 504 U.S. 428 (1992), supply the merits balancing that the claim must survive, while the Kavanaugh formulation in Merrill v. Evans supplies the timing screen placed in front of that balancing once an election is near.

The Principle Protects Rules, Not Outcomes

Formal neutrality is the doctrine’s stated character, and the record across two decades gives that character support. The restraint has frozen rules that restricted voting and rules that expanded access, rules written by legislatures of one political tendency and rules written by officials of another, because the question a Purcell court asks concerns the timing of a federal order rather than the direction of the rule under challenge. A deadline that a court refuses to extend stays in force, and an identification requirement that a court refuses to enjoin also stays in force, and the same stated concern about confusion and administration explains both refusals. That symmetry is the strongest evidence that the principle protects rules in force rather than outcomes for a side.

The Arizona and Wisconsin disputes show the symmetry in concrete form. In Purcell v. Gonzalez, 549 U.S. 1 (2006), the rule preserved was Arizona’s Proposition 200 regime of proof of citizenship at registration and identification at the polls, a restrictive regime whose challengers argued that it burdened lawful voters. In Republican National Committee v. Democratic National Committee, 589 U.S. 423 (2020), the rule preserved was Wisconsin’s statutory deadline for receipt of absentee ballots, and the order stayed a district court extension that would have counted ballots arriving after election day during a public health emergency. A court in both settings declined to let a federal order displace the rule that officials had prepared under, and in both settings the election ran under the pre litigation terms.

The genuine difficulty sits inside that symmetry rather than outside it. The rule in force at the moment litigation ripens is itself a choice, made by a legislature that passed a statute, a governor who signed it, an official who issued guidance, or a mapmaker who drew district lines, and that choice often embodies exactly the contested judgment that the lawsuit challenges. Formal neutrality about timing therefore allocates a real advantage to whoever set the status quo, because the doctrine preserves the chooser’s work for at least one election cycle whenever a challenge matures late. The advantage is structural rather than conspiratorial, arising from the interaction of legislative timing and judicial restraint, but its distribution is not random.

Scanlan v. Coalition for Open Democracy, the October 22, 2026 stay concerning New Hampshire student identification guidance, illustrates the difficulty in a guidance setting rather than a statute setting. Secretary of State David Scanlan’s guidance operated under House Bill 323, District Judge Joseph Laplante entered orders affecting that guidance, and the Supreme Court stayed those orders under two weeks before the midterm election, leaving the state guidance to govern. Justice Ketanji Brown Jackson dissented, and the disagreement between the stay and the dissent reflects the unresolved question of how much weight an official’s own timing should carry when the official’s act created the proximity that the doctrine then treats as a reason for restraint.

Both halves of the account stand without either cancelling the other. The doctrine genuinely applies across political tendencies, and any description that treats it as a device for one side fails against the Arizona and Wisconsin pair, against deadline cases and map cases alike, and against cycles in which each tendency has lost a favourable order to a stay. The doctrine also genuinely rewards control of the status quo, and any description that treats timing neutrality as outcome neutrality fails against the plain fact that preserving a contested rule for an election is a substantive result for the voters who live under that rule. Within the doctrine as it operates, the tension remains exactly where the cases leave it, between a restraint that is neutral in form and a preservation that is never neutral in effect.

Two Decades of Timing Rulings at a Glance

The table settles which kinds of late change courts have frozen, restored, or permitted across two decades, so a pattern becomes visible instead of a sequence of single cases memorised in isolation.

Case and year	Rule change at issue	Stage of the election	Timing outcome
Purcell v. Gonzalez, 2006	Arizona identification and registration proof rules	Weeks before the midterm election, ballots and procedures set	Ninth Circuit injunction vacated and state rules governed the election
Republican National Committee v. Democratic National Committee, 2020	Wisconsin absentee ballot receipt deadline	Days before the primary, with voting under way during a public health emergency	Lower court extension stayed and the statutory deadline governed
Merrill v. Evans, 2022	Alabama congressional map under Section 2 of the Voting Rights Act	Months before primaries, after a three judge court ordered a redraw	Redraw stayed and the challenged map governed the cycle
Merrill v. Milligan, 2023	The same Alabama dispute on the merits	After the 2022 elections had run under the challenged map	Liability decided for the plaintiffs while remedy was left to a fresh remedial process
Moore v. Harper, 2023	State court review of state election rules under a state constitution	Decided outside the emergency posture, on full briefing	State judicial review preserved and timing restraint left as a federal court discipline
People Not Politicians v. Onder, 2026	Missouri congressional map and referendum sequence	After a primary had run and ballots for the general election were prepared	State’s own late course left standing and a federal order imposing a different map rejected
Scanlan v. Coalition for Open Democracy, 2026	New Hampshire student identification guidance	Under two weeks before the midterm election	Lower court orders stayed and the state guidance governed the election

Two patterns stand out from the rows taken together. Liability and timing are decided separately in every dispute that reached both stages, so a plaintiff can establish a violation and still have contested an election under the violating rule, while a state can preserve a rule for a cycle and still lose the authority to use that rule again. The Alabama pair states the separation most sharply, since the February 2022 stay preserved the challenged map for the 2022 elections and the June 2023 judgment in Merrill v. Milligan held for the plaintiffs on liability under Section 2 of the Voting Rights Act.

The second pattern concerns the weight that proximity carries once the rows are compared by stage. Days before a primary, with voting under way, the Wisconsin extension failed. Weeks before a midterm, with ballots and procedures set, the Arizona injunction failed. Months before primaries, after a remedial order had issued, the Alabama redraw still failed, which shows that the restraint reaches well beyond the final fortnight when the requested change is large enough. At the far end of the scale, the Missouri order in People Not Politicians v. Onder on September 25, 2026, left a state’s own late course standing after a primary had run and general election ballots were prepared, and the October 22, 2026 Scanlan stay left state guidance standing under two weeks before a midterm. Across every stage, the closer voting stood and the larger the change requested, the heavier the burden a challenger had to carry, and no row shows a large federal change imposed on a state over objection once ballots were prepared and voting had begun.

Wisconsin in 2020 and the Sentence Later Orders Repeat

Republican National Committee v. Democratic National Committee, 589 U.S. 423 (2020), turned a dispute over a receipt deadline into the order that later courts reach for first. Wisconsin law required absentee ballots to arrive by election day. A federal district court, confronting a primary held during a public health emergency and a surge in requests for ballots by mail, extended the receipt deadline for ballots postmarked by election day. The extension rested on a practical claim about administration under strain. Voters had requested ballots that had not arrived in time to be returned by the statutory date, clerks faced processing volumes far beyond a typical primary, and the district court treated the added days as a remedy for a breakdown in delivery rather than a preference for one rule over another. Five days before the primary, the Supreme Court stayed the extension in a per curiam order, and ballots again had to arrive by election day as the statute provided.

The majority reasoning treated timing as the decisive fact, separate from sympathy for voters caught in the surge. In that reading, a federal court order that alters the rules governing mail ballots and their receipt deadlines on the eve of voting creates the very confusion Purcell v. Gonzalez, 549 U.S. 1 (2006), had warned against. Voters who had heard about the extension might rely on days that no longer existed. Clerks who had prepared instructions reflecting the extension would have to reverse those instructions while voting was already underway by mail. The stay therefore presented itself as protection for orderly administration and for voters who had complied with the original deadline, not as a judgment that the deadline was wise policy or that the emergency was unreal.

Justice Ruth Bader Ginsburg, in dissent, gave the opposite reading its strongest form. The dissent treated the emergency as the reason intervention was justified rather than a reason to withhold it. If ballots were requested in good faith and delivered late through no fault of the voters asking for them, then enforcing the original receipt date without adjustment risked disenfranchising voters who had followed every step available to them. From that vantage point, the district court order preserved the ability to vote instead of disrupting a settled expectation, because the settled expectation itself had been overtaken by postal and administrative failure. The disagreement was therefore not about whether confusion matters. It was about which course produced more confusion and more lost votes under the conditions that actually existed in Wisconsin in April 2020.

The order acquired influence disproportionate to its length because it compressed the doctrine into portable form. Purcell itself had discussed voter confusion, administrative burden, and incentives for late litigation across several paragraphs tied to Arizona identification rules. The Wisconsin order reduced that discussion to a single proposition about lower federal courts altering election rules on the eve of an election, stated at a level of generality that later litigants could invoke in cases involving maps, deadlines, and identification alike. Counsel in later disputes could cite a short per curiam order from a closely watched primary rather than reconstruct an argument from the older Arizona decision. Judges ruling on emergency applications could adopt the same sentence as a complete explanation for a stay, which explains why the formulation traveled through later orders even when the underlying rules differed sharply from the Wisconsin receipt deadline.

Maps Freeze Earlier Than Other Rules

Redistricting tests the Purcell principle at its widest distance from election day because a map is not one rule applied at one moment. A congressional or legislative map assigns every address in a state to a district, and that assignment drives almost every other administrative act that follows. County and municipal officials translate district lines into precinct assignments, and precinct assignments determine ballot styles, the particular combination of federal, state, and local contests printed for a given voter. In a large state the number of distinct ballot styles can reach into the hundreds or thousands, because school board, judicial, municipal, and special district boundaries overlap legislative boundaries in irregular ways. A change to one district line can therefore require renumbering ballot styles, reassigning voters to polling places, and reissuing notices to households whose district, representative, or polling location has shifted.

Candidate administration moves on the same lines. Residency turns on district boundaries, filing offices must confirm that a candidate lives inside the district named on filing papers, and opponents and journalists test those filings against the map in force. If a court replaces a map after filing has opened, every residency determination made under the earlier map becomes open to challenge, and candidates who organized campaigns around one electorate face a different electorate with little time to adapt. Campaigns also allocate staff, advertising, and voter contact according to district composition, so a late map change discards work that cannot be rebuilt on a compressed calendar. Courts weighing those effects treat map litigation as Purcell-sensitive at distances where a single deadline dispute might still be heard, because the number of dependent systems is larger and the time needed to rebuild each system is longer.

The Voting Rights Act framework that much map litigation enforces adds a second calendar inside the first. Liability litigation asks whether a map dilutes minority voting strength or otherwise violates federal law, and that phase alone involves expert testimony on demographics, voting patterns, and alternative maps. A plaintiff victory on liability does not produce a lawful map by itself. The remedial phase operates as a separate proceeding with its own schedule. A court ordinarily gives the legislature an opportunity to enact a replacement map first, respecting the state role in districting. If the legislature fails to act or produces a map the court finds unlawful, the court may appoint a special master to draw lines, receive proposals from parties and intervenors, hold hearings on competing plans, and test each proposal against population equality and federal statutory requirements.

Each step consumes weeks that the election calendar does not pause to provide. Candidate filing deadlines, ballot printing dates, and the federal deadline for transmitting ballots to military and overseas voters continue to approach while the remedial proceeding runs. Election officials must also program tabulation equipment for each ballot style, conduct logic and accuracy testing, train poll workers on precinct assignments, and mail sample ballots that match the final map. A map settled after those processes have begun forces officials to repeat them or to run an election with materials built for lines that a court has already rejected.

That sequence explains a recurring outcome in redistricting disputes. Plaintiffs who establish liability in one election cycle often watch the next election run under the map they defeated, not because a court approved the map on the merits, but because the remedial calendar could not be completed without displacing deadlines that protect ballot delivery and accurate counting. The victory instead governs the cycle after that, once legislatures, special masters, and administrators have had a full interval to translate a judgment into precincts, ballot styles, and voter notices that hold together on election day.

Alabama Splits Liability From Remedy

Alabama supplied the clearest demonstration that liability and remedy can travel in opposite directions. After the 2020 census, the Alabama Legislature enacted a congressional map in 2021 with one district in which Black voters formed a majority. Challengers argued that the map violated Section 2 of the Voting Rights Act because Black voters, who form a substantial share of the state population, could elect a candidate of choice in only one of seven districts. A three judge district court, hearing the dispute in Merrill v. Evans, found that the challengers were likely to succeed and ordered a second district where Black voters could elect a candidate of choice, in time for the 2022 elections. The order rested on the test from Thornburg v. Gingles, 478 U.S. 30 (1986), which asks whether a minority group is large and compact enough to form a district majority, whether the group votes cohesively, and whether majority bloc voting ordinarily defeats the minority preferred candidate.

Does the Purcell principle decide who wins a redistricting case?

The principle governs timing alone. A court may find a Section 2 violation on liability and still postpone the ordered remedy until after an election. Alabama illustrates the split: the state lost the liability ruling yet conducted the 2022 election under the map the district court had condemned.

In February 2022 the Supreme Court stayed the district court order in Merrill v. Evans, so the 2022 elections proceeded under the 2021 map. Justice Brett Kavanaugh, in concurrence, gave the Purcell argument its most systematic statement in that dispute, stressing that a federal order requiring a redraw close to an election imposes heavy costs on candidates, officials, and voters, and that plaintiffs seeking such relief bear a demanding burden. Justice Elena Kagan, in dissent, argued at full strength that the stay weakened Section 2 without a merits decision saying so, by accepting a map the district court found likely unlawful after full proceedings for an entire election cycle.

The contrast with the older Section 5 preclearance regime sharpens the point about route and remedy. Under preclearance, covered states had to obtain federal approval before a map took effect, so the federal check operated before administration began. Section 2 litigation operates after enactment, through lawsuits that must fit liability and remedy inside an election calendar already running.

Full merits review followed a different path. The Supreme Court heard argument in the case, captioned Merrill v. Milligan at the stay stage and decided as Allen v. Milligan, and in June 2023 the Court affirmed the district court liability analysis, keeping the Gingles preconditions in place. Chief Justice John Roberts, writing for the Court, declined to recast Section 2 into the broader restructuring Alabama had proposed. Alabama therefore lost on liability after full briefing and argument, even though the state had won the stay that determined which map governed the intervening election.

Remedial proceedings after the merits judgment exposed the cost of the split. Alabama lawmakers were given an opportunity to enact a compliant map, the legislature produced a map that did not create the second district the district court had required, and the three judge court rejected that effort and turned to a court supervised process with a special master to supply a map for the 2024 elections. The Alabama sequence teaches that timing operates as a separate axis from right and wrong. A Section 2 violation can be real, proven, and affirmed, and the election most directly affected can still be conducted under the unlawful lines because the remedy arrived on a calendar the doctrine treats as too compressed to use.

Missouri Tests Whether the Restraint Binds States Themselves

People Not Politicians v. Onder, decided by order on September 25, 2026, posed a question the earlier Purcell cases had not squarely answered: whether a principle aimed at federal courts also freezes the state whose election is at stake. Missouri had drawn a congressional map through its legislature, and litigation over a referendum effort to challenge that map produced a ruling from the Supreme Court of Missouri that the referendum kept the legislatively drawn map from taking effect. The sequence carried unusual administrative weight. A primary election had already been conducted under one map, and general election ballots had been prepared on the understanding that particular district lines would govern, so any shift in the governing map arrived after substantial reliance by candidates, officials, and voters who had participated in the primary.

Federal litigation followed the state ruling, and parties resisting the state court outcome invoked Purcell as a reason for federal intervention to preserve the legislatively drawn map. The Supreme Court rejected that application of Purcell against the state course of action. As reported, the order stated that federal courts should not impose changes close to an election while states remain free to decide for themselves whether late changes serve their interests, subject to federal law. That formulation drew a federalism line inside the doctrine. Purcell disciplines federal judicial intervention because a federal order displaces the officials charged with administering the election and substitutes a rule those officials did not choose at the point of lowest capacity to absorb disruption. A state acting through its own courts and its own constitutional processes stands in a different position, even when the practical timing is equally tight.

The distinction matters because converting Purcell into a general freeze on late change would invert the purpose attributed to the principle in earlier cases. If the doctrine exists to protect state administration from ill timed federal displacement, then using the doctrine to block a state from governing its own map through its own institutions protects the opposite interest. The state officials who must print ballots and assign precincts are answerable for the choice the state system makes, and a state court applying a state constitutional provision on referenda acts inside the allocation of authority that federal timing doctrine claims to respect. Federal law still supplies an outer boundary: a state decision that violated the United States Constitution or a federal statute would remain subject to federal review on the merits.

Critics of the distinction can state a strong counterargument without misstating the order. Voter confusion and administrative burden do not change character depending on which sovereign causes the late shift, and a ballot prepared for one map misleads voters in the same way whether a state court or a federal court displaces that map afterward. On that account, a timing rule concerned with confusion should care about the calendar and the reliance interests built on the primary, not about the identity of the institution moving last. Defenders of the Missouri line answer that Purcell has never been a free standing confusion test detached from institutional role. The principle grew out of federal stays of federal injunctions, and its repeated formulation speaks about what federal courts should refrain from doing, so extending the restraint to state self government would require a different doctrine with a different justification.

The order therefore leaves a defined boundary rather than a general permission for late change. States retain authority over the timing of their own election decisions, federal courts retain authority over genuine federal violations, and Purcell occupies the space between those propositions as a rule of federal judicial restraint. Litigants after September 2026 must identify which side of that boundary a requested order falls on before the calendar argument can carry any weight at all.

New Hampshire in 2026 and an Order That Decided No Merits

Scanlan v. Coalition for Open Democracy, resolved by Supreme Court order on October 22, 2026, concerned identification at the polls and a dispute in which no court in the emergency sequence decided whether the contested guidance was lawful. New Hampshire enacted House Bill 323, which removed student identification cards from the documents expressly accepted for obtaining a ballot. Guidance issued by Secretary of State David Scanlan went further in practical effect, barring election officials from treating a student card as evidence of identity alongside other proof under fallback procedures available when a voter lacked a listed document. Challengers contested the guidance rather than the basic structure of the statute, arguing that the fallback route existed precisely for voters whose proof did not match the listed categories in a straightforward way.

District Judge Joseph Laplante drew a line between the statute and the guidance. The district court left House Bill 323 standing and blocked the guidance, so the statute as enacted remained the governing rule while officials retained the ability to consider a student card together with other evidence under the fallback procedures. The United States Court of Appeals for the First Circuit declined to pause that district court order, leaving the block on the guidance in place as the general election approached. The Supreme Court then stayed the lower court orders in an unsigned one paragraph decision, restoring the guidance for the election. The stay rested on the eve of election formulation associated with the Purcell line of orders, stating that the district court had pointed to nothing justifying a departure from the principle against altering election rules close to voting.

Supporters of the stay could present that outcome as a straightforward application of a settled timing rule. A district court order issued close to an election had changed the instructions officials would follow at check in tables across the state, training materials and voter information would have to reflect the block on the guidance, and poll workers would have to apply fallback procedures in a manner the secretary of state had instructed against. On that reading, the identity of the underlying document mattered less than the operational fact of a federal order changing practice after preparation had settled, and the proper course was to preserve the rules administrators had organized the election around until full appellate review could occur after voting ended.

Justice Ketanji Brown Jackson dissented alone in public and gave the opposing reading its full force. In her analysis, the district court order changed no state rule at all, because House Bill 323 itself had not barred consideration of a student card as supporting evidence under fallback procedures, and only the secretary guidance had added that bar. Blocking the guidance therefore restored the statute as written rather than displacing it, and the order was designed to keep accurate information before voters about what proof could support a ballot. A stay that reinstated the guidance, on that account, produced the confusion Purcell claims to prevent, by telling voters and officials that a document the statute allowed as supporting evidence would be disregarded after a federal court had held otherwise.

The structural point survives either reading, and the section ends on that point deliberately. The merits of the guidance were never reached in the emergency sequence. No appellate court in that sequence held that the guidance correctly interpreted House Bill 323, and no court held that the guidance violated federal law. The stay preserved a set of identification practices for one election by resolving a timing question, and litigation over the guidance continued after the election under the rules the stay had preserved, with the underlying statutory interpretation still open for decision in ordinary proceedings.

The Elections Clause Decides Who Holds the Pen

Article I, Section 4 of the United States Constitution, known as the Elections Clause, allocates the first move in federal election regulation to state legislatures. The clause provides that the times, places, and manner of holding elections for Senators and Representatives shall be prescribed in each state by the legislature thereof, subject to the power of Congress to make or alter such regulations. States also conduct elections for state office under their own constitutions and statutes, with state officials administering both state and federal contests in a combined system of polling places, ballots, and counting procedures. That combined administration means a single district court order about a federal contest can alter the ballot that carries state contests as well, and a change to a shared registration or identification rule reaches every office on the ballot at once.

Purcell sits downstream of that allocation, and the location explains part of the force the principle carries in federal court. Because the primary rule maker designated by the Elections Clause is the state, a federal court displacing a state rule days before voting does more than act in haste. The federal court overrides the constitutionally designated administrator at the moment of least capacity to absorb the change, after the administrator has printed ballots, assigned precincts, trained workers, and instructed voters under the displaced rule. Timing doctrine operationalises the allocation by translating a standing assignment of authority into a practical presumption: as voting approaches, the burden on a party asking a federal court to substitute a different rule grows heavier, because the institutional cost of substitution rises even when the legal argument for substitution has merit.

The mechanism operates through emergency procedure rather than through a holding on the Elections Clause itself. A litigant seeking a late change ordinarily asks for a preliminary injunction or asks an appellate court to stay a lower court order, and judges deciding such applications weigh likelihood of success, irreparable harm, and the balance of equities. Purcell enters that weighing as a thumb on the scale against late federal alteration, grounded in the observation that the state system, not the federal court, must execute whatever rule emerges from the emergency order. An appellate judge who doubts that county clerks can retrain thousands of poll workers in a weekend is applying the Elections Clause allocation in practical form, even if the order cites only timing cases and equitable factors.

The same constitutional design supplies the counterweight that prevents deference from becoming total. Congress received express power to alter state regulations for federal elections because state rules can burden federal rights or distort federal representation, and federal courts enforce constitutional protections for voting, including equal protection and prohibitions on racial discrimination, against state election codes. A state legislature acting under the Elections Clause does not acquire immunity from the Fourteenth Amendment, the Fifteenth Amendment, or the Voting Rights Act. Deference that ignored those limits would convert an allocation of administrative authority into a license to violate federal rights without a remedy, a result the structure of the clause, with its congressional override, contradicts on its face.

Purcell therefore functions as a calibration device inside a federal system rather than as a declaration that states always govern. Early in a cycle, federal courts can hear challenges, order remedies, and expect states to implement lawful rules with time to spare. Close to voting, the same courts demand a stronger showing before displacing the state rule, not because the state rule has become lawful, but because the federal court respects both the administrative assignment the Elections Clause makes and the practical limits of the officials who must carry out any substitute rule under deadline pressure.

Moore v. Harper Draws the Boundary the Principle Respects

Moore v. Harper, 600 U.S. 1 (2023), decided who may check a state legislature in election matters, a question adjacent to Purcell without being a Purcell case. North Carolina lawmakers had defended a congressional map against review in state court by advancing the independent state legislature theory in a strong form, arguing that the Elections Clause gives state legislatures authority over federal elections free from limits in state constitutions as enforced by state courts. Chief Justice John Roberts, writing for the Supreme Court, rejected that maximal form. The Elections Clause, the Court held, does not exempt state legislatures from the ordinary constraints of state constitutional law, and state courts retain authority to apply state constitutional provisions to election regulations for federal office.

The judgment preserved a federal backstop at the same time. The Court stated that state courts do not have free rein in federal election disputes, and that federal review remains available where a state court transgresses the ordinary bounds of judicial review in a manner that intrudes on the legislature role under the Elections Clause. Roberts framed that backstop in restrained terms, without supplying a detailed test for identifying a transgression, and dissenting justices warned about the uncertainty such a standard creates. Litigants dissatisfied with a state court ruling can invoke the backstop as a reason for federal intervention, while parties defending the ruling can invoke the core holding that state constitutional review is legitimate and expected.

Connecting Moore to Purcell requires care about the distinct questions each doctrine answers. Moore allocates review power, determining that a state court may enforce a state constitution against a legislature and that a federal court may police the outer boundary of that state judicial role. Purcell allocates timing restraint, determining when a federal court should withhold or stay relief close to an election, even when the federal court possesses authority to grant relief on the merits. A litigant can therefore win the Moore question and lose the Purcell question in the same dispute: a state court ruling can be legitimate in origin, and a federal court asked to displace state arrangements close to voting can still face the calendar argument against intervention.

The Missouri sequence in section 14 shows the interlock in practice. A restraint aimed at federal courts loses footing when pointed at a state court performing state constitutional work, and Moore supplies part of the explanation for that weakness. If state courts legitimately enforce state constitutional limits and procedures, including referendum provisions that determine whether a statute takes effect, then a federal timing doctrine cannot easily treat that state judicial work as the kind of external disruption Purcell was built to prevent. The federal court in such a posture would not be protecting state administration from federal interference. The federal court would be interfering with the state process that Moore recognizes as part of the ordinary structure of election governance.

No part of that connection should be overstated. Moore decided no Purcell question, granted no stay, and announced no rule about how close to an election a federal court may act. The decision fixed the allocation of review power within which later timing disputes play out, by confirming that state legislatures act inside state constitutional limits and that state courts act inside federal outer bounds. Purcell disputes arising after June 2023 occur inside that settled allocation, and advocates who blur the two doctrines, by treating a Moore holding about review power as a timing rule or a Purcell stay as a ruling on state court authority, misdescribe both lines of law.

The Merits Test That Timing Can Push Aside

Before timing doctrine can displace a merits analysis, the merits analysis itself needs a clear statement. Anderson v. Celebrezze, 460 U.S. 780 (1983), and Burdick v. Takushi, 504 U.S. 428 (1992), supply the balancing approach federal courts use for many challenges to state election regulations. Under the Anderson-Burdick approach, a court first assesses the character and magnitude of the burden a challenged rule places on voting or associational rights, and then weighs that burden against the precise interests the state puts forward as justifications for the rule and the extent to which those interests make the burden necessary. A rule imposing a severe burden faces strict scrutiny and must be narrowly drawn to advance a compelling state interest. A reasonable and nondiscriminatory rule imposing a modest burden faces a more deferential inquiry, in which important regulatory interests ordinarily suffice.

The approach is fact intensive by design. A filing deadline, a signature requirement, an identification rule, or a ballot receipt deadline can impose different burdens depending on how the rule interacts with postal delivery, work schedules, disability, transportation, and the information voters receive from officials. Evidence therefore includes testimony from affected voters, data on rejection rates, expert analysis of alternative procedures, and statements from administrators about fraud prevention, order, and counting accuracy.

Purcell asks a prior question that can prevent that analysis from governing the election immediately ahead. A case can be strong under Anderson-Burdick, supported by a district court finding of a severe burden and a weak state justification, and still lose the only election that matters for the voters in that cycle because an appellate court concludes that relief arrived too close to voting. The displacement does not refute the balancing analysis. The appellate court may express no view on burden or justification at all, resting entirely on the risk that changing the rule after ballots have been printed or mailed will confuse voters and burden officials. For readers following only the emergency order, the practical result looks identical to a merits defeat, because the challenged rule governs the election either way.

Doctrinal hygiene demands that the two outcomes be kept separate in language and in later litigation. A denial of late relief is not a holding that a rule survives Anderson-Burdick, and a stay issued on timing grounds should not be cited in a later cycle as proof that the rule imposes only a modest burden or serves a compelling interest. Litigants who treat a stay as a merits precedent overread an order that may contain a single paragraph of reasoning. Careful orders state the limited basis for relief, and careful advocates concede that limit when the dispute returns in a posture with time for full review.

Later merits litigation after the election remains possible in many disputes, and that possibility is the main safeguard against permanent displacement. A challenge to a rule that will govern future elections can proceed after voting ends, with a full record and without the calendar pressure that triggered Purcell, and a court can invalidate the rule prospectively for the next cycle. The safeguard has a hard edge, however. The election conducted under the challenged rule stands, and relief after the fact almost never reruns that election on a corrected rule except in extraordinary circumstances involving defects so pervasive that the result itself cannot be trusted. Timing can therefore push the merits aside for the contest where the violation mattered most, even when the legal claim eventually prevails in full.

The Election That Ran Under the Wrong Rule Is Not Rerun

The remedy gap gives the Purcell principle its real cost, measured in elections rather than in doctrine. When a rule is challenged before an election, stayed or left in place on timing grounds, and later held unlawful after full review, the election conducted under that rule almost never gets unwound. Ballots cast by voters who complied with the published rule were counted under the standards officials announced, candidates organized campaigns around the districts and deadlines in force, and officials certified results on which offices were filled and public business proceeded. Unwinding a result after certification would disenfranchise voters who followed the instructions given to them, punish reliance that the system itself encouraged, and substitute a reconstruction of voter intent for ballots actually cast under known rules.

Courts and legislatures therefore direct relief forward in time. The most common form after a successful post election challenge is prospective invalidation, under which the unlawful rule is struck or modified for the next cycle and officials receive a full administrative interval to print accurate materials, revise training, and notify voters of the lawful standard. A legislature may also revise the statute in response to a judgment, codifying a replacement procedure through ordinary lawmaking rather than emergency litigation. Damages appear in rare settings, generally where a separate claim for monetary relief exists against officials for a constitutional violation, and such awards compensate an individual claimant without altering the certified outcome.

That pattern creates an asymmetry that any honest accounting must state plainly. A voter burdened by an unlawful rule in the affected election receives, at best, assurance that a later election will use a lawful rule. The office decided under the unlawful rule is held by the candidate elected under that rule for the full term, the legislation enacted by officials chosen under a defective map carries the same legal force as legislation from a lawful map, and the policy consequences of the tainted election run their course without a corrective vote. Purcell knowingly accepts a defined class of such elections, conducted under rules that should not have governed them, as the price of avoiding a different class of harm involving confusion, administrative failure, and voters misled by rules changed after reliance had formed.

Defenders of the principle argue that the price is set correctly because the alternative harm falls on voters as well, and falls in a manner that cannot be repaired by prospective relief either. A voter who mails a ballot under an extension that a stay later removes loses that vote entirely, and an election administered under contradictory instructions in successive weeks can produce error rates that undermine confidence in the count itself. Critics answer at equal strength that the accepted price is distributed unevenly, because an unlawful restriction tends to burden identifiable groups of voters cycle after cycle when timing doctrine repeatedly postpones the remedy, while administrative confusion is a transitional cost that officials can reduce with earlier litigation deadlines and faster merits review. Each position treats the other harm as real. The disagreement concerns which harm a legal system should risk when both cannot be avoided in a given cycle.

This accounting hands the argument to the complication that follows without deciding that complication in advance. If the price is the election itself, then the absence of a defined threshold for how close is too close becomes more than a drafting gap in a judicial doctrine. The threshold determines how often the system will pay the price, which rules will benefit from postponement, and whether plaintiffs can protect a right in the cycle where the right is exercised or only in a later contest that different voters and candidates will decide under different conditions.

The Strongest Case Against the Principle

The strongest case against the Purcell principle begins with a definition that was never supplied. Purcell v. Gonzalez, 549 U.S. 1 (2006), warned federal courts against orders that change election rules close to voting, yet the per curiam decision stated no number of days, no list of conditions, and no test a lower court could apply in the same way to the next dispute. Richard L. Hasen, who gave the principle its name in Reining in the Purcell Principle, published in the Florida State University Law Review in 2016, argued that an undefined restraint of that kind invites uneven application. A judge who finds a challenged rule congenial can describe any order against it as disruptive, while a judge who finds the same kind of rule objectionable can describe the identical order as modest or restorative. Because the label carries no stated window, neither judge has departed from any announced standard, and no observer can demonstrate inconsistency from the orders alone.

Hasen pressed the point further by examining the setting in which the principle operates. Emergency applications reach the Supreme Court on compressed briefing, and the resulting orders are often short, sometimes unsigned, and sometimes issued without a disclosed division among the justices. A doctrine applied in that setting escapes the discipline that reasoned opinions impose, because a court that gives no reasons cannot be shown to have weighed the same factors in two comparable disputes. The charge, stated with its proper evidence discipline, is not that any justice acted from a proven improper motive. The charge is that the structure creates conditions in which suspicion can neither be confirmed nor dispelled, and a restraint on democratic litigation that cannot be audited carries a legitimacy cost of its own.

The structural version of the critique reaches deeper than judicial choice. Any timing restraint privileges the rule already in force, because the party defending that rule needs only delay, while the party challenging it must win early enough for the victory to matter. A legislature that enacts a restrictive rule late in an election cycle therefore gains a double advantage. The enactment itself changes the governing arrangement, and the lateness of the enactment then becomes the reason a federal court should refuse to disturb it. Lateness, on this account, can be manufactured rather than merely encountered, and a doctrine that treats every late posture as the challenger’s problem rewards the actor who created the posture. Election administrators following the enacted rule inherit the same protection, even when the rule departed from arrangements voters had used in earlier cycles.

Critics add a distributional observation to the structural one. The burden of a timing refusal does not fall evenly across the electorate. Voters who already possess the required identification, who vote by a familiar method, or who live where administration is stable may feel no effect at all, while voters at the margin of the challenged rule bear the entire cost of the court’s restraint. Because the refusal is framed as neutrality between change and stability, that uneven burden can disappear from the stated reasoning even though it decides who votes under which conditions. The critique therefore asks why stability counts as the baseline when stability itself distributes advantages. The defence begins from the opposite premise, that court-ordered change distributes its own burdens, and that argument receives its full hearing after the objection raised from inside the Court in the following section.

Justice Jackson’s Objection: The Order That Prevents Confusion

Justice Ketanji Brown Jackson, dissenting in Scanlan v. Coalition for Open Democracy in 2026, disputed the starting point on which a Purcell analysis depends. The case concerned New Hampshire, where House Bill 323 supplied the governing state law on student identification and where guidance associated with Secretary of State David Scanlan described what that law allowed. District Judge Joseph Laplante entered an order restraining the guidance, and the Supreme Court stayed that order on October 22, 2026. Justice Jackson argued that the district court’s order did not alter a state election rule at all. In the challengers’ account, which her dissent credited, the guidance misdescribed the statute, so an order restraining the guidance left the enacted rule exactly where the legislature had placed it and prevented officials from applying a description the statute did not support.

That distinction exposes a category problem inside the doctrine. Purcell reasoning assumes a stable baseline and a disruptive intervention: the state rule is the baseline, and the court order is the change that threatens confusion. Litigation posture can reverse those roles. Where officials issue guidance that departs from the enacted text, the guidance is the departure and the court order is the restoration, so blocking enforcement preserves the older and more familiar arrangement rather than disturbing it. A court that asks only which document arrived last will misidentify the source of disruption, because familiarity attaches to the arrangement voters and local officials have actually practised, not to the paper filed last. Justice Jackson’s dissent treated the order in Scanlan as an instrument for keeping accurate information in front of voters, not as an instrument for changing the rules under which voters cast ballots.

The analysis extends beyond one New Hampshire dispute to a class of orders the doctrine fits awkwardly. Purcell grew out of ballot mechanics, identification requirements, registration rules, and deadlines, where a late order forces administrators to reprint, retrain, and reprogram under time pressure. Orders about information operate differently. A ruling that governs what officials may say in a notice, print in a mailing, or post at a polling place can reduce confusion rather than create it, particularly where the restrained statement pointed voters toward a requirement the statute did not impose. Treating every order touching an election as equally disruptive collapses that difference. The relevant inquiry, on the logic of the dissent, is directional: whether the order moves practice toward or away from the enacted rule and toward or away from the arrangement already understood by the people administering and using it.

Justice Jackson’s objection also carries a burden-of-proof implication for emergency practice. If the state must show only that an election is near and that an order exists, then guidance of doubtful accuracy gains protection simply because someone challenged it late. If instead the state must show that the order changes the operative rule, the doctrine retains its confusion rationale while losing its capacity to shield a misdescription. The dissent did not command a majority, and its category distinction remains a proposed refinement rather than a settled limit. Its force lies in showing that the central Purcell question cannot be answered until a court identifies what counts as the rule, what counts as the change, and whose action moved the arrangement away from what voters had reason to expect.

The Defence at Its Strongest

Justice Brett Kavanaugh supplied the defence in its most systematic form in his concurrence in Merrill v. Evans, the February 2022 stay concerning an Alabama congressional map. His starting point was institutional. State legislatures, accountable through elections of their own, hold the primary authority to set the rules for state and federal contests, and federal judicial intervention displaces that authority. Intervention close to voting therefore requires more than a plausible claim. In Justice Kavanaugh’s formulation, a plaintiff seeking late relief must show a strong likelihood of success on a plainly meritorious claim, and the remedy sought must remain feasible for officials to implement without serious disruption. Those conditions convert Purcell from a mood of caution into a gateway with stated elements, and the defence insists that the gateway, properly applied, answers the charge of open-ended discretion.

The asymmetry of error does much of the remaining work. A court that wrongly refuses late relief injures identified plaintiffs, whose claims survive the election and can be heard with full briefing, a developed record, and a remedy applied to the next cycle. A court that wrongly grants late relief injures a different and larger set of people: voters who never sued, who receive conflicting instructions, who appear at the wrong place or without a newly required document, and whose ballots may be lost to confusion that no later judgment can repair. The election itself cannot be rerun as a routine remedy, so the second error is less correctable than the first. On that reasoning, restraint near voting is not indifference to rights but a judgment about which mistake the legal system can still fix after fuller consideration.

Election administrators supply the defence with its evidence class. Across case after case, officials responsible for running elections, drawn from both major parties and speaking through professional associations and sworn declarations, have told federal courts that late changes produce concrete harms: ballots already printed, poll worker training already completed, public guidance already distributed, and systems already tested against a fixed set of rules. A court weighing a late order hears from the people who must execute it, and the defence argues that disregarding those accounts would not protect voters. It would transfer the cost of judicial confidence onto registration clerks, county administrators, and voters standing in line, none of whom chose the litigation timetable. Feasibility, in this account, is not an excuse layered over the merits but a fact about whether an ordered change can exist in practice before voting ends.

The verdict that follows from granting the defence its full strength can be stated plainly. The defence justifies a timing restraint. It does not justify an undefined one. A doctrine grounded in confusion, feasibility, and asymmetric error can state the conditions under which those concerns control, can require the party invoking them to prove disruption rather than assert it, and can release its grip where an order restores the enacted rule instead of replacing it. The thesis therefore narrows rather than collapses: Purcell survives as a principle about timing, and its weakness remains the absence of the stated conditions that its own rationale makes possible.

The Exceptions Courts Still Honour

The restraint described in the preceding sections has never operated as an absolute bar, and the situations in which courts intervene near an election share a common structure. Each exception works by neutralising the rationale that gives Purcell its force. Where an order creates no confusion, disrupts no settled reliance, and restores rather than replaces the arrangement in use, the timing objection loses the harm it exists to prevent. Understanding the exceptions as mechanisms rather than as moods of judicial sympathy explains both their reach and their limits, and it explains why litigants argue so intensely over how a dispute is characterised before any court reaches the underlying right.

The first exception concerns a rule that plainly violates a clear constitutional command. Justice Kavanaugh’s conditions in Merrill v. Evans built this exception into the gateway itself, because a strong showing on a plainly meritorious claim is one of the stated requirements for late relief rather than a plea made outside the doctrine. The mechanism is comparative. Where the constitutional defect is clear and the violation will govern the casting and counting of ballots, the harm of leaving the rule in place is concrete and large, while the confusion rationale depends on disruption that the state must still establish. Clarity does the limiting work: a claim that requires extended analysis to evaluate does not qualify, which is why this exception is described as settled in structure even though its application to particular claims remains contested.

The second exception arises where the challenged action is itself a late change made by officials departing from the enacted rule. Justice Jackson’s dissent in Scanlan v. Coalition for Open Democracy in 2026 supplied the logic, arguing that an order restraining guidance that misdescribed New Hampshire’s House Bill 323 preserved the legislature’s arrangement instead of altering it. The September 25, 2026 order in People Not Politicians v. Onder points in a complementary direction from the state’s side, stating that states remain free to make late changes for themselves even though federal courts should not impose them. Together the two positions mark the boundary of the exception. Official action that departs from enacted law may be restrained without offending Purcell’s rationale, while the identity of the actor matters, since a change a state chooses for itself stands on different footing from a change a federal court orders.

The third exception covers orders that restore the arrangement voters already understood. Reliance interests, which the defence invokes to protect stability, attach to the rule voters learned, the identification they gathered, the deadlines they planned around, and the instructions local officials had given across earlier cycles. An order returning practice to that arrangement serves reliance rather than defeating it, even when the order arrives close to voting. The contested character of this exception should be stated in those words: it is contested rather than settled, because the parties will usually dispute which arrangement counts as the familiar one, and the dispute over characterisation can reproduce the entire merits argument in miniature. Courts honouring these exceptions therefore do more than make room for sympathetic claims. Each exception asks the Purcell question with greater precision, identifying where confusion would actually arise, whose reliance is actually at stake, and whether the order under review is the change or the correction.

Reading a Late Election Order Without a Law Degree

How can a reader tell whether an election ruling rested on timing or on the merits?

Timing rulings speak of confusion, administrative burden, and closeness to voting, and they grant or deny a stay or injunction without settling the claim. Merits rulings analyse the right asserted and the justification offered for the burden, and they dispose of the claim itself by judgment.

A disciplined reading of a late election order begins with the disposition, the formal act the court performed. A stay suspends a lower court order while review proceeds. An injunction commands or forbids conduct. A judgment resolves the claim itself. The disposition tells a reader how much the court actually decided, because a stay granted on timing grounds leaves the underlying dispute alive, while a judgment against a plaintiff ends that dispute subject to appeal. Headlines and summaries often convert the first kind of act into the second, describing a court as having upheld a rule when the court held only that the rule would remain in place for one election. The disposition is the guard against that misreading, and it appears in the opening lines of the order itself.

The second element is the standard the court recited. A timing analysis names the ingredients Purcell made familiar: the closeness of the election, the risk of voter confusion, the burden on officials, and the strength of the claim as one factor among several. A merits analysis works through a different apparatus, weighing the burden a rule imposes on voting against the justification the state offers, in the manner associated with Anderson v. Celebrezze, 460 U.S. 780 (1983), and Burdick v. Takushi, 504 U.S. 428 (1992). A reader who finds the right analysed element by element is reading a merits ruling. A reader who finds the calendar doing the decisive work, with the right discussed briefly or assumed for argument, is reading a timing ruling, whatever the surrounding commentary claims.

Two further checks complete the method. The first asks whether the court analysed the underlying right at all, or instead recorded that the claim might be strong while declining relief on timing grounds alone. The second asks what happens after the election: where litigation continues toward a judgment governing the next cycle, the earlier order was provisional by design. The Wisconsin litigation resolved in Republican National Committee v. Democratic National Committee, 589 U.S. 423 (2020), illustrates the pattern, because the Supreme Court stayed an extension of the absentee ballot receipt deadline while the merits of ballot receipt rules remained a separate question. Scanlan v. Coalition for Open Democracy in 2026 offers the contrasting study in how characterisation controls outcome, since the majority’s stay treated the district court’s order as a late change while Justice Jackson’s dissent treated the same order as a restraint on guidance that had departed from the statute. Reading both documents side by side shows the method plainly: disposition first, standard second, the treatment of the right third, and the afterlife of the case last.

How Election Officials Plan Around the Freeze

Election administration absorbs the Purcell principle into routine planning long before any dispute reaches a courtroom. Administrators who have watched late orders freeze one arrangement or another learn to move their own decisions earlier in the cycle. Rulemaking, guidance documents, and interpretive memoranda are issued with wide margins before voting begins, because guidance released close to an election invites two risks at once: litigation challenging its content, and a timing objection that locks in whichever posture the challenge creates. An office that publishes its interpretation early gives challengers time to sue while a court can still grant effective relief, and gives itself time to comply with an adverse ruling while compliance remains administratively possible. If you run such an office, the calendar on the wall is a litigation document as much as a production schedule.

Version control supplies a second defensive practice. Training manuals for poll workers, scripts for telephone inquiries, and instructions printed for polling places are locked to a single approved version at a fixed point in the cycle, and any correction after that point travels through a documented supplement rather than informal adjustment. The practice matters in court as much as in the field. When a judge weighs the feasibility of a late change, the locked version is evidence of how far preparation has travelled and how many people would have to unlearn an instruction. An administrator who can produce the version history, the print dates, and the distribution records gives the court concrete material for the disruption inquiry instead of an assertion that disruption would occur.

Production deadlines are set backward from fixed points that litigation cannot move. Federal law requires ballots for military and overseas voters to be transmitted 45 days before a federal election, and that transmittal date anchors the entire production chain behind it: ballot styles must be finalised, proofing completed, and printing finished early enough for mailing to occur on schedule. The canvass, the formal process by which results are certified after voting, stands as the fixed end point at the other side of the cycle. Every internal deadline between transmittal and canvass inherits the rigidity of those anchors, which is why a court order arriving after ballot transmission confronts not a preference for stability but a sequence of physical acts already performed.

Litigation readiness completes the pattern. Offices prepare declaration templates, preserve records under litigation holds adopted before any dispute exists, and identify the officials who can describe printing, training, and mailing timelines under oath within days of a filing. That preparation serves the court’s inquiry directly, because feasibility under Purcell is an evidentiary question and the side with documents usually defines its contours. The deeper effect is cultural. Administrators plan as though any late dispute will be decided on the state of their preparation, so the doctrine shapes conduct in thousands of offices that never appear in a reported decision, and its influence on an election may be greatest in the disputes that preparation prevented from being filed at all.

What a Defined Purcell Doctrine Would Look Like

Richard L. Hasen proposed the leading reform in Reining in the Purcell Principle, published in the Florida State University Law Review in 2016. Hasen asked the Supreme Court to state conditions for applying the timing restraint rather than invoking Purcell as a conclusion. In substance, his proposal directed a court to weigh the strength of the claim on the merits, the disruption the requested change would actually cause, and which side created the emergency that forced the late posture. Late relief would remain available where a constitutional violation was clear. The proposal accepts the defence’s core insight, that late change carries real administrative and democratic costs, while denying that those costs excuse an unreasoned refusal. A court following the proposal would produce a record on which its choice could be examined, and the audit problem at the heart of the critique would shrink accordingly.

Every defined doctrine must then settle design choices that the case law leaves open. The first choice concerns the trigger. A fixed day window, counted backward from the start of voting, offers predictability: litigants, officials, and judges would share a boundary that no advocacy could move. Critics of a fixed window answer that disruption does not obey a calendar, since a change ordered 60 days before voting may be trivial in one state and impossible in another where ballots are already printed. A sliding disruption test answers that objection by measuring printing, mailing, and training as facts, at the price of reintroducing discretionary judgment. You can see the trade plainly once the choice is framed that way: certainty purchased with rigidity, or accuracy purchased with discretion that must itself be supervised.

The second choice concerns reasons. A written-reasons requirement for emergency orders would let litigants and the public test whether the stated conditions were applied evenly across disputes, and it would give lower courts instruction instead of inference. The opposing argument is institutional and practical: the emergency docket moves on deadlines measured in hours, full opinions take time the election does not allow, and demanding reasons may delay the very certainty the doctrine exists to supply. Intermediate designs answer with short statements of the controlling factors, issued with the order and supplemented later. The third choice concerns the anchor date for any presumption against late relief. A presumption keyed to the start of voting tracks voter confusion closely, while a presumption keyed to ballot printing tracks administrative feasibility, and the two dates can sit weeks apart inside one cycle.

The design debate stands in a particular posture. Justice Kavanaugh’s concurrence in Merrill v. Evans moved doctrine toward stated conditions by naming the elements a late movant must satisfy, and lower courts have worked with those elements since 2022. No fixed window has been adopted, no general reasons requirement governs emergency election orders, and the anchor question remains unresolved. That posture is itself instructive. The Court has accepted, in part, the premise that Purcell should operate through conditions a litigant can address, while reserving the discretion that a complete definition would surrender. A defined doctrine therefore exists in outline, assembled from a concurrence, a body of scholarly design work, and repeated dissents identifying the gaps, waiting on the institutional choice to complete it.

What the Principle Cannot Settle

A timing doctrine orders the sequence in which decisions get made. It does not supply the values that the sequence serves, and several of the hardest questions in election law sit precisely in the space Purcell cannot reach. The principle can say that a court should hesitate before changing a rule close to voting. It cannot say how much voter confusion justifies leaving how much disenfranchisement in place for one cycle, because that exchange rate is a judgment about the relative weight of participation and orderly administration, and no doctrine of timing contains a scale for weighing them. A court that invokes Purcell to end the analysis has answered the sequencing question while the value question continues underneath, usually decided by the same hesitation that the doctrine recommended.

A second set of questions concerns the legitimacy of the emergency process itself. Whether orders that govern an election should carry reasons, whether the votes of individual justices should be disclosed on such orders, and how much explanation a losing litigant and a watching public are owed are questions about institutional design rather than about timing. Purcell neither requires nor forbids reasoned elaboration; its silence on the subject is one source of the critique developed by Richard L. Hasen, yet supplying reasons is a reform of the Court’s emergency practice generally, not a corollary of election doctrine. The voters governed by a rule that governed one election and fell afterward raise a third question the principle cannot settle: what a system owes people whose ballots were cast, or turned away, under an arrangement later held unlawful. Finality doctrines answer that elections ordinarily stand, but the answer comes from the law of remedies and contests, not from Purcell.

Other doctrines carry the loads that Purcell sets down. The merits of a voting restriction are weighed under the Anderson-Burdick apparatus, from Anderson v. Celebrezze, 460 U.S. 780 (1983), and Burdick v. Takushi, 504 U.S. 428 (1992), which balance the burden on the right to vote against the state’s asserted justifications. Litigants and scholars have also advanced due process arguments contending that election administration can become so unfair in its departures from announced rules that it violates the federal Constitution; those arguments are presented here as positions their advocates hold, not as settled law, and courts have accepted them only in narrow circumstances if at all. State constitutional provisions supply a further layer, since the constitutions of several states protect voting in terms stronger than the federal floor and sustain challenges that federal doctrine would not. Recognising the boundary of Purcell is therefore a clarifying act. The principle decides when a federal court should act. The worth of the rule being defended, the fairness of the process that produced the governing order, and the remedies owed after the fact belong to other arguments, conducted under other names.

The Calendar Is a Party to Every Election Case

In federal election litigation, the calendar is not background. The calendar is a deciding force, present in every dispute whether or not any order mentions it, because a right vindicated after the votes are counted is a different right from one vindicated before them. A voter who casts a ballot under a lawful rule has exercised the franchise as the legal system promised it could be exercised. A voter who wins a judgment two cycles later has received a correction, valuable for the future and incapable of restoring the earlier act. Purcell v. Gonzalez, 549 U.S. 1 (2006), is the doctrine that prices that difference. It does not ask whether a rule is lawful. It asks whether a federal court may replace the rule in the time remaining, and it treats the answer as capable of deciding the election’s governing terms without any court reaching the underlying right.

The habit worth carrying away from the whole line of cases, from Purcell through Republican National Committee v. Democratic National Committee, 589 U.S. 423 (2020), and Merrill v. Evans in 2022 to Scanlan v. Coalition for Open Democracy in 2026, is a separation performed in two steps. The first question asks whether the rule is lawful, a merits inquiry governed by the right asserted and the justification offered. The second asks whether a court may replace the rule in time, a timing inquiry governed by closeness, confusion, feasibility, and the character of the order requested. Accounts that merge the two questions misdescribe nearly every late election ruling, because a stay that answers only the second question gets reported as a holding that answered the first. Justice Jackson’s dissent in Scanlan and Justice Kavanaugh’s concurrence in Merrill disagree about much, yet both insist, from opposite directions, that the character of the order and the posture of the case be identified before any conclusion is drawn.

The practical residue of that habit is record keeping, since a dispute separated into its two questions can be tracked, dated, and revisited as the litigation continues past election day. A researcher following a redistricting challenge across cycles, or an administrator maintaining the file on identification guidance through successive rounds of briefing, keeps the merits papers and the timing papers in separate order inside a litigation register so that a stay is never mistaken for a judgment when the file is opened months later. Orderliness of that kind is the reader’s version of the doctrine’s own discipline.

The structural condition underneath the doctrine guarantees its survival in some form. Elections run on fixed dates fixed by statute and constitution, and courts run on litigation time, which expands with briefing, records, and appeal. Any legal system holding both facts at once will generate a doctrine that mediates between them, deciding which late claims a court may grant and which must wait. Purcell is the name the American system gives that mediation. The live question, the one Richard L. Hasen posed in 2016 and the one the Kavanaugh conditions began to answer in 2022, is whether the mediation proceeds under stated conditions that litigants can test and observers can audit, or under a label that ends the conversation at the moment the conversation matters most.

Frequently Asked Questions

Q: What is the Purcell principle in simple terms?

In simple terms, the principle tells a federal judge to leave state election rules alone when an election is near, even when a lawsuit raises a serious claim against those rules. A late order can scramble instructions that voters have already learned, force officials to redo printing and training, and cast doubt over a result produced under shifting terms. Leaving the rule in place avoids that scramble, at the price of running one election under terms that a later trial might condemn. The idea therefore separates legality from timing. Judges reserve the legality question for a full hearing held away from election pressure, while the timing question decides which rules govern the contest already approaching. One election therefore runs under the challenged terms, while the argument over legality continues afterward for later cycles.

Q: Where does the name Purcell come from?

The name comes from Purcell v. Gonzalez, 549 U.S. 1 (2006), an Arizona dispute over Proposition 200, a 2004 measure that demanded proof of citizenship during registration and identification at polling places. A federal appeals court blocked parts of that regime in October 2006 without explaining the reasoning, and the Supreme Court vacated the block in a per curiam opinion issued for the Court as a whole. That short opinion warned that judicial orders entered near voting can confuse voters and weaken confidence in electoral processes. Later courts treated the warning as a general restraint on late intervention, and the case name attached to the restraint as shorthand for the whole line of reasoning that followed from the Arizona litigation. Emergency orders after 2006 kept returning to that warning in disputes over deadlines, maps, and identification rules.

Q: Did Purcell v. Gonzalez decide that voter identification laws are constitutional?

No holding in Purcell v. Gonzalez, 549 U.S. 1 (2006), pronounced Arizona identification requirements constitutional, and no part of the per curiam weighed the evidence that a merits case would have required. The justices vacated an unexplained Ninth Circuit injunction entered weeks before the 2006 midterm election, faulting the absence of stated reasoning at a moment when officials and voters needed clarity about governing rules. Questions about proof of citizenship at registration, identification at the polls, burdens on lawful voters, and the interests Arizona claimed for Proposition 200 remained open after the order issued. Litigation away from an election deadline, built on testimony and documents, remained the route for resolving those questions on a complete record. The order restored the Arizona regime for the 2006 election without settling the lawfulness of that regime in either direction.

Q: Why do courts worry that changing election rules late will confuse voters?

Courts worry because most voters learn an election rule once, from a clerk’s notice, a news report, a campaign message, or a neighbour, and then act on that learning without checking for revisions. A judicial change travels through each of those channels unevenly, so different voters end up holding different versions of the deadline, document, or district that applies to them. A voter acting on an outdated version may arrive without an accepted document, mail a ballot under a deadline that has moved, or appear at a precinct that cannot supply the correct ballot. Poll workers trained on an earlier procedure can deepen the mismatch by applying remembered instructions. Election officials cannot correct every mistaken belief in the final days, even with public notices and revised training. Confusion of that kind produces lost ballots and unequal treatment without any official intending either result.

Q: Is a Supreme Court stay under the Purcell principle the same as winning the case?

A stay and a merits victory decide different questions, and treating them as identical misreads both orders. A stay entered under the Purcell restraint preserves a state rule for the election immediately ahead, based on an emergency record about disruption, administrative burden, and the equities of intervention at a late stage. Claims in the case stay alive afterward, and a trial or appeal on a full record can still hold the preserved rule unlawful and bar its use in later cycles. Alabama supplied the clearest sequence, since the February 2022 stay in Merrill v. Evans kept a challenged congressional map in place for that cycle, while Merrill v. Milligan in June 2023 decided liability for the plaintiffs on the merits. Timing relief and final liability can therefore point in opposite directions within a single dispute without any inconsistency.

Q: Who named the Purcell principle, and in what publication?

Richard L. Hasen named the principle in Reining in the Purcell Principle, an article published in the Florida State University Law Review in 2016. Hasen gathered a set of short emergency orders, issued across the decade after Purcell v. Gonzalez, 549 U.S. 1 (2006), and gave the shared practice of refusing late changes a label that briefs and opinions could invoke directly. His argument accepted that late changes impose genuine costs, while urging that courts discipline the restraint with conditions instead of treating lateness alone as a complete answer to a claim. Justices later adopted the label in concurrences and dissents, which moved a scholar’s description of judicial behaviour into the vocabulary that courts use about their own decisions. The label also gave critics a target, since a named practice can be measured against stated conditions in a way an unnamed habit cannot.

Q: What must a voter or group show to persuade a court to change a rule close to an election?

Justice Brett Kavanaugh’s concurrence in Merrill v. Evans, the February 2022 Alabama map stay, described the showing that a late request for change demands. A claim must appear strong on the merits rather than merely arguable, because an emergency court will not disrupt preparation for a violation that a full record might not establish. Harm from leaving the rule untouched must appear severe, in the sense of ballots lost or rights denied during the election at hand rather than inconvenience that another procedure can cure. Implementation must also appear workable, supported by evidence from officials about printing, testing, training, and communication, showing that the change can occur without heavy cost, confusion, or trouble. Each element grows harder to establish as voting approaches. Declarations, testimony, and administrative records usually carry that proof, rather than argument alone from counsel in emergency papers.

Q: How did the Purcell principle affect the Wisconsin primary in 2020?

In the Wisconsin primary of April 2020, a federal district court extended the deadline for absentee ballots to arrive if the ballots carried a postmark by election day, responding to a surge in mail voting during a public health emergency and delays in ballot delivery. The Supreme Court stayed that extension five days before the primary in Republican National Committee v. Democratic National Committee, 589 U.S. 423 (2020), restoring the statutory rule that ballots arrive by election day. The majority treated a federal alteration of the receipt rule close to voting as a source of voter confusion and administrative disruption. Justice Ruth Bader Ginsburg dissented on the ground that voters who requested ballots in good faith should not lose votes because delivery systems failed under emergency conditions.

Q: Why are redistricting cases treated as more urgent earlier than other election cases?

A redistricting change reaches every administrative system built on district lines at once. Officials must reassign voters to precincts, rebuild ballot styles that in a large state can number in the hundreds or thousands, verify candidate residency and filing under the replacement lines, notify households of district or polling place changes, and reprogram and test tabulation equipment for each affected ballot style. Liability litigation over a map also leaves a separate remedial phase, in which a legislature receives a chance to redraw, a court may appoint a special master, and competing proposals require hearings and review. Because each step consumes weeks while filing and ballot transmission deadlines continue to approach, courts treat map relief as disruptive at distances from election day where a single deadline change might still be manageable for administrators.

Q: What happened in the Alabama redistricting dispute between 2022 and 2023?

A three judge district court found that Alabama congressional map enacted in 2021 likely violated Section 2 of the Voting Rights Act and ordered a second district where Black voters could elect a candidate of choice in time for the 2022 elections. The Supreme Court stayed that order in Merrill v. Evans in February 2022, with Justice Brett Kavanaugh writing on timing in concurrence and Justice Elena Kagan warning in dissent that the stay weakened Section 2 in practice. Alabama conducted the 2022 election under the 2021 map. After full argument, the Supreme Court affirmed the liability analysis in June 2023 in the dispute captioned Merrill v. Milligan, preserving the Gingles preconditions from Thornburg v. Gingles, 478 U.S. 30 (1986), and remedial proceedings for a compliant map followed for the next cycle.

Q: Does the Purcell principle stop a state from changing its own election rules late?

The Missouri order in People Not Politicians v. Onder on September 25, 2026, treated the restraint as a limit on federal courts rather than a freeze on state self government. In that sequence a state supreme court ruling on a referendum kept a legislatively drawn congressional map from taking effect after a primary had run and general election ballots had been prepared. The Supreme Court rejected an attempt to use Purcell against the state course, stating as reported that federal courts should not impose changes close to an election while states remain free to decide whether late changes serve state interests, subject to federal law. A state choice that violated the United States Constitution or a federal statute would still face federal review on the merits.

Q: What did the Supreme Court decide in the New Hampshire student identification case in 2026?

On October 22, 2026, in Scanlan v. Coalition for Open Democracy, the Supreme Court stayed lower court orders that had blocked guidance from Secretary of State David Scanlan concerning student identification cards. House Bill 323 had removed student cards from the documents expressly accepted for obtaining a ballot, and the guidance barred officials from treating a student card as evidence of identity alongside other proof under fallback procedures. District Judge Joseph Laplante left the statute standing and blocked only the guidance, and the First Circuit declined to pause that order. The Supreme Court stay, issued in an unsigned paragraph resting on the eve of election formulation, preserved the guidance for the election without deciding whether the guidance correctly interpreted state law. Justice Ketanji Brown Jackson dissented.

Q: What is the Elections Clause and how does it relate to the Purcell principle?

Article I, Section 4 of the United States Constitution provides that state legislatures prescribe the times, places, and manner of holding elections for Senators and Representatives, subject to alteration by Congress. The clause makes the state the primary rule maker and administrator for federal elections, even though Congress and federal courts enforce federal rights against state rules. Purcell operates downstream of that allocation. A federal court that displaces a state election rule close to voting overrides the constitutionally designated administrator after ballots, precincts, training, and voter instructions have been built around the state rule. Timing doctrine translates the standing allocation into a practical presumption against late federal substitution, while leaving federal enforcement intact for violations identified with time for orderly implementation.

Q: What is the difference between a court applying Purcell and a court upholding an election rule?

A court applying Purcell decides a timing question, and a court upholding an election rule decides a merits question. Timing relief, such as a stay of a lower court injunction close to an election, leaves the challenged rule in force for that election without holding that the rule satisfies the Constitution, the Voting Rights Act, or the Anderson-Burdick balancing approach from Anderson v. Celebrezze, 460 U.S. 780 (1983), and Burdick v. Takushi, 504 U.S. 428 (1992). A merits judgment after full briefing determines whether the burden the rule imposes is justified by state interests and whether the rule may govern future elections. Confusing the two outcomes overreads an emergency order. A rule preserved by a stay can later be invalidated prospectively, even though the election conducted under the rule stands.

Q: Why do critics say the Purcell principle can be applied selectively?

Critics begin with the absence of a stated test. Purcell v. Gonzalez, 549 U.S. 1 (2006), announced a caution against late federal court orders without fixing a number of days or a list of conditions, so two judges facing comparable disputes can reach opposite results without either one departing from an announced standard. Richard L. Hasen developed this objection in Reining in the Purcell Principle in 2016, arguing that an undefined restraint can be invoked against claims a court disfavours and set aside for claims a court welcomes. The emergency setting deepens the concern, because applications are decided on compressed briefing and orders are often brief, unsigned, or issued without disclosed votes. Under those conditions an uneven pattern, if one existed, would leave almost no evidence behind, and critics treat that unauditable quality as a defect in the doctrine itself rather than as an accusation against any individual judge.

Q: What did Justice Jackson argue in dissent in the New Hampshire case?

Justice Ketanji Brown Jackson dissented from the Supreme Court’s stay in Scanlan v. Coalition for Open Democracy in 2026. The dispute concerned guidance associated with Secretary of State David Scanlan describing New Hampshire’s student identification law, House Bill 323, and a district court order entered by Judge Joseph Laplante restraining that guidance. Justice Jackson argued that the district court’s order altered no state election rule, because the order left the enacted statute untouched and restrained only an official description that, in the challengers’ account, misstated what the statute allowed. On her reasoning, the order aimed to keep accurate information before voters rather than to change the rules for voting. The dissent therefore treated the stay as the disruptive act and the restrained guidance as the departure, reversing the roles that Purcell reasoning usually assigns to state action and court action.

Q: How does Justice Kavanaugh think courts should approach late election changes?

Justice Brett Kavanaugh set out his approach in a concurrence in Merrill v. Evans, the February 2022 stay involving an Alabama congressional map. His reasoning starts with institutional authority: state legislatures set election rules, and federal courts displace that authority when they intervene, so intervention close to voting demands a demanding showing. A plaintiff seeking late relief must demonstrate a strong likelihood of success on a plainly meritorious claim, and the requested remedy must be capable of implementation by election officials without serious disruption. Justice Kavanaugh also emphasised the asymmetry of error. A wrongful refusal leaves identified plaintiffs free to continue litigating after the election, while a wrongful late change can confuse voters who never sued and cannot be made whole afterward. His conditions present Purcell as a gateway with stated elements rather than as a general judicial reluctance to act near an election.

Q: Can a court ever change an election rule shortly before voting begins?

Yes, although the circumstances are narrow and each recognised route works by removing the harm that the timing restraint exists to prevent. A court may act where the claim is plainly meritorious under a clear constitutional command, a condition Justice Kavanaugh’s Merrill v. Evans concurrence built into the gateway for late relief. A court may also restrain action by officials that itself departs from the enacted rule, the logic of Justice Jackson’s Scanlan dissent, since restraining a departure preserves the legislature’s arrangement rather than replacing it. A third route covers orders restoring the arrangement voters already understood, because reliance interests attach to the familiar rule. Each route is contested at its edges, since parties dispute which arrangement is familiar and how clear a violation must be, but the existence of the routes shows that Purcell operates as a strong presumption rather than as an absolute prohibition.

Q: What reforms have scholars proposed for the Purcell principle?

The leading proposal comes from Richard L. Hasen in Reining in the Purcell Principle, published in the Florida State University Law Review in 2016. Hasen asked the Supreme Court to replace invocation of the principle as a conclusion with stated conditions for its application. Under his design, a court would weigh the strength of the merits claim, the disruption the requested change would actually cause, and which side created the emergency that produced the late posture. Relief close to voting would stay open in cases presenting a clear constitutional violation. Other design questions recur across the scholarly debate: whether a fixed day window would serve better than a sliding test based on administrative disruption, whether emergency election orders should carry written reasons despite the speed of that docket, and whether any presumption should be keyed to the start of voting or to the earlier milestone of ballot printing. None of these designs has been adopted in full.

Q: Why does an election conducted under a rule later held unlawful usually stand?

An election is a completed collective act rather than a transaction that can simply be unwound. Ballots were cast in reliance on announced rules, officials counted them under a single set of instructions, and candidates, voters, and governments ordered their conduct around the certified result. Courts therefore treat finality as the default and reserve invalidation for circumstances involving defects so extensive that the outcome itself cannot be trusted, a demanding standard administered through election contest procedures rather than through the original challenge. The ordinary consequence of a successful late challenge is prospective: the unlawful rule is removed for the next cycle, and the plaintiffs who established the violation govern future elections under the corrected arrangement. That division between a completed election and a corrected future is one reason timing doctrine matters so heavily, because relief deferred until after voting is real relief for the system and incomplete relief for the voters who participated under the defective rule.