Now add the emergency question. If an administration says foreign interference or voting-system vulnerabilities require federal action, government lawyers have to decide what statutes actually permit. Prosecutors have to decide whether allegations justify investigations. Cybersecurity officials have to decide whether the evidence supports the President’s claims. Somebody may have to say no.
Which brings us back to those 8,000 career employees.
There is a large difference between identifying powers that could become relevant in a disputed election and proving an administration intends to use them together. The evidence establishes the former, not the latter.
But the timing matters. Between November 3 and January 3, an administration that may just have lost Congress still controls the executive branch while the new Congress does not yet exist. Courts remain courts. States remain states. Congress retains its constitutional powers. There is no magical lame-duck dictatorship clause hiding in Article II.
Power does not have to be unlimited to matter. It merely has to be applied at the right pressure point in a sufficiently close election.
That is why the Schedule Policy/Career piece bothers me more than it did when I first read it. The question isn’t simply who tries to control a database or Postal Service regulation. It is whether the lawyer who says an order is unlawful, the prosecutor who sees no basis for an investigation or the election-security official who says the evidence doesn’t support the President’s claim can safely keep saying so.
Each piece has a respectable explanation. That is how these things usually look. Nobody announces at a press conference, “Good afternoon. The autocracy starts Thursday at nine.” Lawyers write memoranda. Agencies issue regulations. Somebody cites subsection 4(c). Everything remains reassuringly bureaucratic right up until the moment you notice where the decisions have accumulated.
The danger is not necessarily that Trump declares martial law on November 4 and Hegseth sends tanks toward Philadelphia. That would at least have the virtue of being obvious. The more interesting possibility is that a president doesn’t need martial law at all. He needs enough leverage over the institutions surrounding an election that the system starts bending before anybody agrees on what to call it.
The Supreme Court’s postal ruling is an important reason not to overstate that danger. The administration wanted a new federal role in getting ballots into the mail. States resisted, judges blocked it and the Supreme Court left the block in place. That is what constitutional friction is supposed to look like.
Maybe that is exactly what happens everywhere else. The courts do their jobs, the states run the election, somebody wins Congress and Trump complains about it with his customary restraint and dignity.
But a democracy ought to get nervous when an administration spends the months before an election testing new ways to involve itself in an election it is constitutionally not supposed to run—particularly when control of Congress may change hands. The reassuring lesson from the postal case is that the system can stop it.
The less reassuring question is how many times the system may be asked to.
Somewhere in Washington there is already a glass box marked EMERGENCY.
What worries me is that somebody has apparently been reading the instructions.
Footnotes
¹ USPS’s August Federal Ballot Mail rule imposed new ballot-envelope, barcode, voter-data and verification requirements and authorized USPS to reject noncompliant outbound ballot mailings. Judge Indira Talwani entered a nationwide preliminary injunction blocking key provisions for the 2026 election; the First Circuit denied the government’s request for a stay. Trump-appointed Judge Carl Nichols separately blocked implementation of the rule in Washington.