The Warrant Standard

The problem

Two different things happen under the name "immigration enforcement," and the country argues about them as if they were one. The first is a stop: an officer decides to approach someone and ask about status. The second is an arrest that crosses a threshold, a home's front door, a business's back room, a courthouse hallway, where the person did not agree to be found. Neither currently requires what every other American arrest requires. A stop needs no documented, individualized reason at all. An arrest, including one that enters a home, has been authorized on a Form I-200 or I-205, a warrant an immigration officer fills out and signs for another immigration officer, never seen by a judge.

The scale of the first problem grew fast. ICE made an estimated 481,800 administrative arrests in fiscal 2025, up 51 percent from fiscal 2024. Arrests of people with no criminal record rose more than 2,000 percent over the same stretch, evidence that the agency shifted from named targets to street-level stops. By December 2025 and January 2026 the daily rate hit 1,264 arrests a day on average, roughly 300 percent above the year before. A policy that scales that fast on discretion alone, with no record of why any individual stop happened, will produce errors at scale, and it has: applying a Northwestern University estimate that about 1 percent of people in immigration detention are U.S. citizens to the 65,135 people ICE held as of mid-November 2025 puts upward of 2,000 citizens in custody at any given time. In February 2025 a Milwaukee toddler, his mother and his grandmother, all citizens, were detained in one operation. In March 2026 the Justice Department admitted in court that the legal memo it had cited to defend arrests made inside immigration courthouses did not actually apply there.

The second problem came to a head in May 2025, when an internal ICE memo authorized officers to enter a home on an administrative warrant alone if the occupant had a final removal order. On January 11, 2026, agents forced their way into a Minneapolis home under that policy to arrest a Liberian national who was, in fact, complying with a supervised check-in schedule. U.S. District Judge Jeffrey Bryan ruled the entry unconstitutional: an administrative warrant, signed by the same agency doing the arresting, does not satisfy the Fourth Amendment's requirement of a neutral magistrate. DHS publicly disputed the ruling three weeks later. Nothing above the district level has settled it, so the rule now depends on which federal courthouse a case happens to land in.

The proposal

A stop requires a specific, contemporaneous reason on the record, never a person's appearance, language or accent standing alone. An arrest that is not made in public with the person's consent, one that enters a home, a business's non-public area, or a courthouse, requires a warrant a judge signed on a showing of probable cause, available within the hour by phone or secure video the same way federal criminal warrants already are. Administrative warrants remain valid for everything else: street arrests, port-of-entry processing, and any entry the occupant actually consents to.

How it would work

1. The stop standard. Before initiating any immigration stop, an officer must have and log a specific, individualized reason, tied to conduct or documentation, not to how someone looks or sounds. The log is made at the time, not reconstructed afterward, the same discipline already required of a Terry stop in ordinary policing.

1. The warrant standard. An arrest that crosses a threshold without consent, a home, the back of a shop, a courthouse corridor, requires a warrant signed by a federal judge or magistrate on probable cause. Administrative warrants (Form I-200, I-205) keep full force for arrests made in public, at ports of entry, or with the occupant's actual consent to enter. Nothing here touches removability. It changes who may authorize the door being opened, not who may eventually be removed.

1. The expedited channel. Judicial sign-off happens by telephone or secure video, modeled on Federal Rule of Criminal Procedure 4.1, which has let criminal investigators get a magistrate on the line for search and arrest warrants for decades. A duty-magistrate rotation, the same model federal courts already run for weekend criminal arrests, keeps the channel open around the clock. This is the direct answer to "agents can't wait for paperwork": the paperwork already exists, and takes minutes, for every other kind of American arrest.

1. The public count. Each quarter, ICE and CBP publish, by field office: stops made and the reason logged for each category, warrant applications sought, granted, and denied. The federal judiciary already produces exactly this kind of report for wiretaps, a warrant for something far more covert than an arrest, every year without incident. An immigration warrant standard gets no less daylight than a surveillance one.

1. What stays untouched. Criminal prosecutions for illegal reentry already go before an Article III judge; this changes nothing there. A person with a final order who is encountered in public, or who opens the door and lets an officer in, is arrested exactly as today. The rule closes one gap: entry without consent, authorized by no one but the agency doing the entering.

The numbers

The volume a judge would actually see. Of the 481,800 fiscal 2025 arrests, the great majority happen in public or at ports of entry and are untouched by this rule. The relevant slice is non-consensual entries, home and business arrests, which ICE does not break out publicly, itself an argument for item 4. For scale, federal and state judges authorized 2,297 wiretaps nationwide in 2024, a warrant that requires proving a case for continuous covert surveillance, not one probable-cause showing for one arrest, and the courts absorbed it inside the ordinary docket. 563 full-time federal magistrate judges sit across the country; a national duty rotation spreads the load thin.

What the current process has already cost. One home entry under the May 2025 policy produced a federal ruling against the government within nine months. One legal memo, misapplied to courthouse arrests, produced a Justice Department concession in open court. Neither of those errors would have needed a judge to prevent; a judge would have had to sign off before either happened.

What zero would look like. № 051 makes the same bet for citizenship: a published annual count of citizens wrongly held, with a target of zero. This proposal is the policing half of the same bet, a published quarterly count of stops, reasons, and warrants, with the same target.

The honest objections

"Agents in the field can't wait on a judge, and the people they're chasing will be gone by the time the phone call ends." The Rule 4.1 channel this proposal borrows was built for exactly this problem in ordinary criminal work, and it resolves in minutes, not days. It also only applies to the minority of encounters, non-consensual entries, that need it; a street arrest on an administrative warrant is untouched and stays as fast as it is today.

"Judges approve almost everything anyway, so this is paperwork theater." Likely true, and worth conceding rather than dodging: the wiretap comparison suggests approval rates will be high. The point of the warrant isn't a high denial rate. It's that a named judge looked at the facts and signed something, which is the entire difference between an administrative warrant and a judicial one, and it is also what creates the record in item 4 that does not exist today.

"Requiring officers to write down a reason just produces pretextual reasons, not honest ones." Also true in ordinary policing, and the answer there is the same one proposed here: a contemporaneous log plus published pattern data. Nobody claims logging catches every bad-faith stop in real time. It catches the pattern, over a quarter, in a public report, the way disparate stop rates get caught in ordinary police accountability work today.

"This is amnesty by process; it just makes deportation harder." No removable person becomes less removable. A public arrest on an administrative warrant, or a consented entry, proceeds exactly as before. What ends is the specific combination of no consent, no judge, and a door opened anyway.

"DHS says the current system already works and this fixes a problem that isn't real." DHS made exactly that argument in February 2026, defending administrative warrants publicly. Set against it: a federal district court found a home entry under that system unconstitutional the month before, and the Justice Department itself conceded an erroneous legal basis for a category of arrests the month after. A system that produces two admitted failures within ninety days, neither one of them caught by anyone inside the agency, is not a system a judge's signature would hurt.

Sources

• U.S. Immigration and Customs Enforcement, Enforcement and Removal Operations statistics, fiscal 2025 administrative arrests, ice.gov
• American Immigration Council, analysis of ICE arrest data by criminal history, 2025, americanimmigrationcouncil.org
• TRAC Immigration, Syracuse University, daily arrest rates December 2025–January 2026 and detention population as of November 16, 2025, tracreports.org
• Jacqueline Stevens, Deportation Research Clinic, Northwestern University, estimate of U.S. citizens in immigration detention, cited via TRAC and press reporting, 2026
• Office of Rep. Jamie Raskin and Rep. Pramila Jayapal, press release on wrongful detention of U.S. citizens, February 2025, democrats-judiciary.house.gov
• NPR, "DOJ admits ICE courthouse arrests relied on erroneous information," March 26, 2026, npr.org
• U.S. District Court, District of Minnesota, ruling of Judge Jeffrey Bryan on warrantless home entry, January 2026, reported via Just Security and Davis Vanguard
• U.S. Department of Homeland Security, "DHS Sets the Record Straight on Administrative Warrants," February 4, 2026, dhs.gov
• Federal Rule of Criminal Procedure 4.1, Complaint, Warrant, or Summons by Telephone or Other Reliable Electronic Means, law.cornell.edu
• Administrative Office of the U.S. Courts, Wiretap Report 2024, wiretaps authorized by federal and state judges, uscourts.gov
• Administrative Office of the U.S. Courts, Judicial Business 2025, status of magistrate judge positions, uscourts.gov
• Proposals № 001 (The American Dividend); № 007 (The Humane Perimeter); № 051 (The Citizen Card)