Two Executive Orders, Two Very Different Days in Court: Birthright Citizenship and Mail Voting

The administration issued both orders. Both ran through federal agencies. Both used the word “shall.” One got sued the same week it came out and lost at the district court the same month. The other took five months before the Supreme Court said the challenge was premature and the district court had never had jurisdiction. The difference wasn’t the subject matter. It was a drafting choice. The birthright citizenship executive order told agencies what result to produce. The mail voting executive order told one agency to start a rulemaking proceeding whose result wasn’t predetermined. Nothing in the Constitution requires the one choice over the other. The choice decides which body of law tests the policy, when the clock starts, who gets through the door, and what a loss actually costs.

This is the executive-orders detour Gwen teased after the judicial review season opener. The series ahead returns to the threshold doctrines of judicial review — standing, finality, ripeness/exhaustion/mootness, scope limits. This episode is the detour because the president isn’t an agency, which means executive orders generally aren’t reviewable agency action under the APA. There’s no rule to set aside, no administrative record to attack. The implementing agencies produce the reviewable material — guidance, regulations, denials. Which means the structure of the order determines when the reviewable material arrives. A direct command on substance produces reviewable material immediately. A command to run a rulemaking might not produce it for months, and when it does, the rulemaking might have reached a different conclusion than the order projected.

The birthright order, signed January 20, 2025, used a verb pointed at a document. It said no department or agency “shall issue documents” recognizing citizenship or accept documents from other authorities purporting to do so for persons whose mother was unlawfully present and whose father wasn’t a citizen or lawful permanent resident, applicable to children born more than 30 days after the order issued. Four officials were named by title and told to conform their agencies’ regulations and practices. The order itself wasn’t adjudicating any one case — but it told every agency what result to reach when the application arrived. The plaintiffs filed three kinds of claims. First, administrative claims against the implementing officials — contrary to law, in excess of statutory authority, arbitrary and capricious, procedurally improper. Second, constitutional claims against the order itself under the Fourteenth Amendment. Third, a procedural theory: if an agency changes a binding legislative rule, it ordinarily has to use the procedures that made it. A plaintiff doesn’t have to pick. The complaint can plead all of them and let the court choose. Which choice the court makes decides whether the government can rebuild its position after losing.

The decision ladder matters. Losses on different rungs cost different amounts. A procedural APA loss means the agency can come back and do it correctly with better paperwork. A statutory holding closes one authority but leaves others open — Congress could pass a new statute. A constitutional holding takes the objective off the table entirely. The 2023 Biden student loan plan illustrates the middle rung. The announcement was presidential, but the legal act was the Secretary of Education’s determination under a 2003 statute, dated the same day, with an Office of Legal Counsel opinion dated the day before. Careful drafting, and because the statute specifically exempted the waivers from APA § 553, the obvious procedural attack was off the table before anyone filed. The plan still lost. In Biden v. Nebraska (2023), the Court held the Secretary lacked authority — but on major-questions grounds, meaning as a statutory holding, not constitutional. Within hours, the president announced the administration would start over under the Higher Education Act. The middle rung isn’t a door held open. It’s a door into a much longer hallway — the Higher Education Act requires negotiated rulemaking, which takes years. But it’s still a hallway. A constitutional loss has no hallway.

The Trump v. CASA arc tested where the birthright order landed on the ladder. On June 27, 2025, the Court restricted universal injunctions to the relief necessary to protect the particular plaintiffs. Litigation moved into class actions almost immediately. Within hours of CASA, Maryland plaintiffs filed an amended class complaint; a New Hampshire court provisionally certified a nationwide class covering essentially the same children. The injunction in the New Hampshire case became Trump v. Barbara, decided at the end of June 2026. The order was affirmed 6-3, but not 6-3 on the same ground. Five justices held the Fourteenth Amendment itself guaranteed citizenship to the children the order covered. Justice Kavanaugh was the sixth vote for the judgment, but on statutory grounds — the order violated the Immigration and Nationality Act’s codification of the Citizenship Clause, which in his view had closed the exceptions list in Wong Kim Ark (1898). On his reading, Congress could amend the statute to write exceptions in. If his reading had five votes, that road would be open right now. On the majority’s holding, it isn’t — the Fourteenth Amendment can’t be legislated around.

Barbara is also where constitutional avoidance didn’t operate. Avoidance comes in two varieties that get taught as one. The canon is a rule of statutory construction: if a statute can be read to avoid a constitutional problem, you read it that way. It has a floor — the alternative reading has to be genuinely available. The practice is broader: don’t answer a constitutional question you don’t have to answer. The practice has no floor. A court can decline to reach a question for any reason or none and never has to say which. The practice makes sense when the constitutional question is genuinely hard. It gets thin — or should get thin — when the question isn’t hard at all. In Barbara, the narrow statutory ground existed. One justice took it. Eight didn’t. Five of the majority decided the constitutional question. The three dissenters decided it the other way. Nine for nine on reaching it. The lesson for drafters: when the constitutional question is the strongest claim against your order, bare orders deliver it to the court quickly. If the court wants to answer it, nothing stops them.

The follow-up order confirms the dynamic. A second birthright order came out. The government made the same wait-for-the-paperwork argument it had made against the 2025 order — the order tells agency heads to enforce it through all appropriate measures, and those measures get spelled out in public guidance that hasn’t been issued yet. The district judge rejected it. Her reasoning tracked the order’s verbs: “No matter what the guidance says, the 2026 executive order commands agencies to deny citizenship documents to several broad categories of children.” She left the implementation requirements alone and enjoined the denials to anyone in the class. Twenty-seven days. The grounds were already gone — the majority’s constitutional holding in Barbara had taken them. On Kavanaugh’s reading, Congress could reach for an amendment. On the majority’s, nothing is left.

The mail voting order shows the opposite route. Executive Order 14399, dated March 31, 2026, is on citizenship verification in federal elections. Multiple sections. One directs the Secretary of Homeland Security to compile and send citizenship lists to states. One directs the Attorney General to prioritize prosecutions. One directs the Postmaster General to initiate a proposed rulemaking that would include a provision saying the Postal Service “shall not transmit mail-in or absentee ballots” from anyone not enrolled on a state list. Mandatory language throughout. Hard deadlines — 60 days for the rulemaking to start, 90 days for the DHS infrastructure, 120 days for the final rule. But look where the “shall nots” attach. They attach to a proposed provision — the thing the rulemaking will consider. Not to the final rule. The order commands a proceeding. The proceeding can produce a different result. The 120-day deadline is for “any final rule” — a deadline for something that might never exist. That isn’t a drafting trick; it’s what a rulemaking is. If an agency can’t change its mind after notice-and-comment, the exercise is a sham.

Twenty-three states and D.C. sued the mail voting order in April. By early July they had a district court judgment enjoining the directive sections. A separate case brought by voting organizations produced a nationwide injunction against part of the order in August. The government’s response was the same each time: you’re early, the rule doesn’t exist yet. On August 24, 2026, the Supreme Court agreed with the government per curiam, saying there was no jurisdiction because it was speculative whether the Postal Service would propose a rule, weigh comments, and then issue a final rule consistent with the order. “Federal courts review final rules, not proposed rules, and certainly not antecedent internal directives to propose a rule.” The government had also told the district court on August 7 that if a final rule affected an individual voter, “they certainly have standing to challenge that final rule” and “it would only take a very short amount of time” for emergency relief. Twice the government said: wait for the rule and you’ll have your day in court.

The Postal Service finished the rule three days before the Supreme Court order. Published in the Federal Register August 26. The plaintiffs did what they’d been told to do and filed on publication day. The government’s position at that point: too late, appellate review would come after the midterms. The district judge granted a 14-day TRO on August 27, pointing out that standing is measured as of filing and that the rulemaking the Court cited as speculative had completed. Her footnote noted that the government’s supplemental brief about the final rule was docketed hours before the Supreme Court’s order, with no indication the justices had seen it. The deeper problem is the one that resolved only after the rule issued: Congress has exempted the Postal Service from ordinary APA review. There’s no arbitrary-and-capricious count. There’s no § 705 stay. The government’s own position is that the plaintiffs are left with only constitutional claims and a narrow ultra vires theory the Court has recently compared to a Hail Mary. The argument that bought five months — that you just needed to wait for the final rule to have your APA case — turned out to be describing a case the plaintiffs can’t bring.

The two orders together make a complete picture of what drafting choices decide. The careful route isn’t automatically safer. It’s cheaper to lose on. The 2022 Biden loan plan ran through the Secretary under a 2003 statute, got a year of life, and lost on major-questions grounds — a middle-rung loss with the Higher Education Act hallway open. The 2025 birthright order ran straight through the president, got a direct constitutional loss, and the second attempt bought 27 days. The 2026 mail voting order ran through an agency whose rulemaking takes months, bought five months, and now sits in litigation about whether any review framework attaches at all because of the Postal Service’s APA carve-out. What each instrument bought was time and a different set of questions it could lose on. The real drafting question isn’t which instrument protects you. It’s which loss you can live with. A bare order with a mandatory verb attached to a document goes straight to the constitutional question. An order routed through an agency is more likely to lose on procedure or statute first — losses that leave the door open to try again. Both can produce the same outcome. The sequence they arrive in is most of what timing is.

Listen Now

What We Cover

Full Transcript

[Transcript]

Related Guides

Administrative Remedies is produced by Professor Gwendolyn Savitz and Professor Marc Roark at the University of Tulsa College of Law.