Tarpon Springs is a Greek town because sponge divers came from the Dodecanese to work the Gulf a century ago and stayed. New Bedford holds Portuguese and Cape Verdean families because whaling ships took on crews in the Azores and off the African coast. Basques settled Boise and turned to herding sheep. Ford hired Yemeni men to build cars in Detroit. Vietnamese refugees took up shrimping off Louisiana and Texas. The country is built out of pockets like these, close to two hundred separate origins by the census count, blended across 340 million people into something no single word holds.
Zoom in on most American cities and the colors run together. One stretch southwest of Houston can be Nigerian, Jordanian, Mexican, Vietnamese, Salvadoran, and Irish on a single block. Chicago is not that city. Chicago reads in hard blocks. The New York Times mapped the country this summer and caught the difference at a glance. Los Angeles blends gold and green, Minneapolis runs purple, and Chicago comes up in blunt reds and blues and yellows, each color holding its edge. That sharpness is the story.
Chicago in blocks
Where each group concentrates, lake to the east. A schematic, not a census map.
▲ N
Polish
Italian
German
English
Puerto Rican
The Loop (mixed)
African American (West Side)
Chinese
Mexican
Irish
African American (South Side)
Lake Michigan
An original schematic in the spirit of the census ancestry map, not a reproduction of it. National flags mark each group; African American carries the Pan-African flag, since there is no single country of origin. Positions are approximate, meant to show the pattern rather than plot a tract.
The Times read its map as a record of two centuries of push and pull, and of policy that opened the door to some people and barred it to others. That is exactly right, and it is the whole of the work I do. People arrived in Chicago, wave after wave. Law placed them, and sometimes law removed them. Neither force explains a single block. Together they explain the city. Let me walk the map the way a Chicago immigration lawyer reads it.
~15%
U.S. foreign-born share in 2024 — the highest ever recorded
~22%
Chicago residents born abroad (≈597,400), highest since 2006
227,700
Mexican-born in Chicago, 2024 — up ~9% as the national count fell
Start with the country, because Chicago sits inside it. From 1820 to 1860 the door stood mostly open, and more than five million people came. The steamship dropped the price of passage, and the lines sold cheap immigrant fares bundled with rail tickets to the interior. Once a community took root somewhere, it pulled the next arrivals toward it, the way it always has. Germans left political collapse at home, the Irish left the famine, and after the 1848 gold strike in California the first Asian laborers were recruited onto the farms and the railroads.
Then the flood. From 1880 to 1920, twenty-four million immigrants landed, pushed out by pogroms across Eastern Europe and by the upheaval of a newly unified Italy. They went nearly everywhere except the South, where the landed elite already had cheap labor from the formerly enslaved and from poor tenant farmers. Congress met the wave the way it usually does, by moving to stop the next one. It had already barred almost all Chinese immigration in 1882, the first federal law to slam a door on a whole people. In the 1920s it built a national-origins quota system that rationed entry by birthplace, generous to Western Europe, stingy toward the South and East, set near zero for most of Asia. Ships raced through the night to reach Ellis Island before the year’s quota filled. The door stayed mostly shut for forty years, and the foreign-born share of the country fell to the lowest level in its modern history.
Then, in the 1960s, lawmakers who saw the quota system for the racial sorting it was replaced it with one built on family ties and skills. The next wave came from new places, and the makeup flipped. A century ago the immigrant was almost always European. Of the roughly fifty million immigrants counted in 2024, most come from the Americas and from Asia. Mexico still leads, then India, China, the Philippines. Europe is a thin slice now. Refugees came too, Southeast Asians displaced by the Cold War, then Cubans and Haitians and Venezuelans and Afghans and Ukrainians fleeing collapse and war, most of them steering toward the cities where their countrymen already lived. The foreign-born share is back near fifteen percent, and Washington is again writing restriction into law, in bans and quotas that echo the 1920s closely enough to notice.
The map you see now is the last century’s policy, printed in where people’s grandchildren live. Chicago recorded every wave of it.
The city, wave by wave — and where each one landed
The first great arrivals came to dig. Irish laborers cut the Illinois and Michigan Canal through the 1840s and stayed to run the wards, the parishes, and in time the machine that governed Chicago deep into the next century. Germans arrived in the same decades and took the North Side, brewing and printing and filling the skilled trades. Swedes settled Andersonville. Greeks took their own near-West-Side blocks. By the Civil War the city was foreign-born in fact, if not yet in its own imagination.
1880–1920
The flood
Then came the many. Poles filled the Milwaukee Avenue corridor until Chicago held one of the largest Polish populations of any city on earth. Ukrainians built the near West Side blocks still called Ukrainian Village, onion domes and all. Lithuanians arrived in such numbers that Chicago became, by most accounts, the largest Lithuanian city outside Lithuania, packing into Bridgeport and Back of the Yards to work the stockyards; Upton Sinclair set The Jungle among them. Italians settled Taylor Street. Bohemians built a Southwest Side neighborhood and named it Pilsen, after the city they had left. Eastern European Jews crowded onto Maxwell Street, then moved west into Lawndale.
Most of those enclaves have since thinned below the map’s threshold. The Lithuanian, Ukrainian, and Bohemian quarters did not vanish so much as disperse, their grandchildren scattering into the suburbs and the general mix, which is why a map drawn to today’s concentrations barely shows them. That is its own kind of erasure, quieter than the census religion ban, and just as real.
By 1910, three of every four Chicagoans were immigrants or the children of immigrants. The city was not a melting pot. It was a room full of separate pots, each on its own fire.
The map even offers “American” as an origin, chosen mostly by the descendants of the oldest arrivals, the English and Scotch-Irish who came first and stopped naming a homeland. It is a claim, not a birthright, and it makes the Polish block or the Mexican block next door no less American than the one that took the plain word for itself.
What the map can’t see The census does not ask about religion, by law, and that silence erases people. The Jews who fled the pogroms of the Russian and Austro-Hungarian empires land on the map as Russian, or Polish, or as nothing at all, swept into an “other” no color captures. A migration that built the West Side, then West Rogers Park, then suburban Skokie, half-disappears in the data. The Assyrians who came later meet the same fate. Read the map as a sketch, not a ledger.
Diverse before it was a city
Chicago was mixed before it was Chicago. The land belonged to the Potawatomi, and to the wider Council of Three Fires before them, until the treaties and forced removals of the 1830s pushed them west at gunpoint. The first permanent settler who was not Native was Jean Baptiste Point du Sable, a Black man of Haitian and French descent who built a trading house at the mouth of the river in the 1780s. A Native place, opened to the wider world by a Black Haitian. The city has never once been the white and Christian monolith some people picture when they imagine its founding. It began as a meeting point, and it stayed one.
1916–1970
The Great Migration
The next great wave were not immigrants at all. They crossed no ocean, filled out no visa, and descended from people who were brought here in chains rather than choosing to come. Black families left the South by train and bus, many of them reading the Chicago Defender before they came, and settled the South Side stretch that became Bronzeville before spreading west. They arrived into a city that had already decided where they would be allowed to live.
This is where arrival and law pull apart most cleanly, and where law wins. White Chicago wrote racial covenants into its deeds, private contracts that promised never to sell to a Black buyer. Banks drew red lines on maps and starved the enclosed blocks of mortgage money. When Carl Hansberry bought a house on the wrong side of a covenant line, the fight climbed to the Supreme Court, and his daughter Lorraine later turned it into A Raisin in the Sun. Hansberry won in 1940 on narrow procedural grounds, and the covenants held on across the city. Only in 1948, in Shelley v. Kraemer, did the Court finally forbid courts from enforcing them. By then the lines were already cut into the map. You can still see them in the hard color blocks the Times charted this year. Segregation this sharp is not an accident of heritage. Somebody drew it, on purpose, in contract and mortgage.
1910s–present
The Latino city
Mexicans came to Chicago to work the steel mills and the rail yards, and the city pulled them in when it wanted hands and threw them out when it did not. Washington ran mass deportation drives in the 1930s and again in the 1950s, even as the Bracero program bused Mexican laborers north to fill the very same jobs. The families who stayed built Pilsen, Little Village, and Back of the Yards into the Mexican heart of the Midwest. Puerto Ricans, citizens the day they were born, made Humboldt Park their own and hung the flag over Division Street.
The Mexican story is where the present bites hardest, and I laid it out at length earlier this month. The short version runs against the national trend. Across the country, the Mexican-born population has fallen from its 2007 peak. In Chicago it went the other way. The city’s Mexican-born count rose about 9 percent in 2024, to roughly 227,700, reversing a full decade of decline. That was the year before the raids.
After 1965
The wider world
The 1965 law reopened the door, and Chicago’s map gained colors. Chinatown, hemmed in for decades, grew and pushed settlement toward Bridgeport and the suburbs. Devon Avenue turned into one of the great South Asian commercial strips in the country, Indian sari shops beside Pakistani groceries beside Bangladeshi kitchens. Assyrians, Ukrainians, West Africans, and a Chinese population reaching well past the old Chinatown all took their place on the grid. The newer immigrant does not always land in the historic enclave. More and more, he lands in the suburbs and skips the enclave altogether.
The city now
Buses in, then the raids
Chicago in 2024 held about 597,400 immigrants, close to 22 percent of the city, the highest share since 2006. The newest layer arrived by bus. Starting in 2022, border-state governors sent tens of thousands of asylum seekers north, many of them Venezuelan, and Chicago scrambled to shelter them through two hard winters. It was the same magnet that has always worked here, people steering toward a city where others from home already lived, only this time the trip was engineered as a political message.
Then the wave turned inward. When border crossings collapsed, the government pointed its enforcement at the interior, at the cities, and Operation Midway Blitz sent federal agents through Pilsen and Little Village, the Mexican core. Over time the tactics grew quieter, moving from street sweeps to targeted arrests at courthouses and check-ins, where the government already knows where you will be. The reach now runs through data most people never think about. A boarding pass has become a kind of search, and I wrote recently about the TSA–ICE agreement that turns a domestic flight into an arrest. Even the ordinary paperwork of belonging carries a trap. For a noncitizen, registering to vote can wreck a green-card case, sometimes over a mistake the state itself made in adding the name. The immigrant in these neighborhoods today crosses a legal minefield the 1910 arrival never faced.
It is an old rhythm all the same. Fear empties the streets. Then, across years, the streets fill again, the newcomers turn into neighbors, and the city absorbs them the way it absorbed every wave before. The person an agent pulls off a job this month is cut from the same cloth as the steerage passengers we now print on postcards.
The counter-melody
The other Chicago tradition
Law does a second thing here, and it runs the other way. Chicago does not only place people and shut them out. At intervals it refuses to do the federal government’s removal work.
That refusal is older than the word “sanctuary.” Before the Civil War, Northern states passed personal liberty laws that gutted the federal Fugitive Slave Acts from below, guaranteeing jury trials, forbidding local officials from slave-catching, denying the use of county jails. The federal statute stayed on the books and went hollow. Across roughly eighty years of that resistance, the whole apparatus returned only a few hundred people. Emerson called the Act a dead letter in the states that fought it. I traced that lineage in a longer piece on how states quietly resist federal overreach, from the personal liberty laws straight through to the modern sanctuary ordinance.
Chicago inherited the reflex. Harold Washington signed an executive order in 1985 barring city workers from questioning residents about immigration status. The city wrote it into law as the Welcoming City ordinance in 2006. Illinois scaled it statewide with the TRUST Act. The fight is live again right now, as federal power presses against state autonomy in the courtrooms and on the streets. The language shifts across two centuries. The move does not. A federal government orders removal. A city declines to supply the hands.
Who comes next
Demographers now think the country could reach net-zero immigration, arrivals matched by departures, sooner than almost anyone expected a few years ago, as enforcement accelerates and legal pathways narrow. If that holds, the map will change again, the way the 1920s quotas changed it, printing themselves into where people live two and three generations on.
But the pull has not weakened. The country is aging, its workforce thinning, its need for workers running straight into a politics that wants fewer of them. A nation can answer that need by building a longer table or by building a higher wall, and it has tried both.
Chicago has run this collision before. It is the city that placed people by law, and the city that refused to hand them over, and both traditions are alive right now, on the same blocks, at the same hour. Watch the map. It is still being drawn.
The source This piece builds on “How a Nation of Immigrants Traces Its Roots” by Albert Sun, Jeff Adelson, and Larry Buchanan (The New York Times, July 1, 2026), and the interactive ancestry map that ran with it, drawn from U.S. Census Bureau American Community Survey data, 2019–2024. The maps and figures here are my own; the Times’ graphics are not reproduced.
This post is general information, not legal advice, and does not create an attorney-client relationship. If you or a family member face immigration enforcement, consult a licensed immigration attorney about your specific situation.
Every institution in American life is built to make voting easier. One institution is built to make a single wrong vote catastrophic. The same person stands in front of both.
By Michael D. Baker
Ask Claude how to register to vote and a banner appears. It points you to TurboVote, a nonpartisan service run by the nonprofit Democracy Works. TurboVote will help you register, remind you when an election is coming, and walk you through voting by mail. The banner exists because Anthropic decided that if people are going to ask an AI about elections, the responsible move is to hand them off to something accurate and real. That is a good decision. I want to be clear about that up front, because the rest of this is going to sound like criticism and it is not aimed at them.
It is aimed at a contradiction that runs straight through American democracy, and TurboVote just happens to be the cleanest example of one side of it.
Here is the contradiction. We have spent fifty years lowering the barrier to voter registration. Motor-voter laws register you at the DMV. Online portals register you in ninety seconds. AI assistants now surface a registration link the moment you ask. Every piece of that machinery is designed to reduce friction — to make the act of registering as close to automatic as a free country can manage.
And then, for one specific person, we treat the frictionless act as one of the gravest things he could possibly have done.
🚪 The Person the Machine Was Not Built For
A lawful permanent resident stands at the same DMV counter as everyone else. A clerk asks if he wants to register to vote. The whole system is engineered to make him say yes — that is the entire point of motor-voter. Nobody at the counter is trained to stop him. The form does not check his status. The machinery does exactly what it was designed to do: it registers a person with almost no friction.
Then that person votes once, believing the government would not have handed him the card if he were not allowed to use it. Years later he discovers that the single most accessible civic act in America is, for him, a federal offense — and grounds to deport him.
That is not a hypothetical. It is the fact pattern of real, published federal cases, and I have written about the newest one at length. What I want to sit with here is the design problem underneath all of them, because the design problem is the thing nobody in the shouting match will name.
🔁 The Paradox in One Breath
We built the registration system to be effortless, on purpose, for good reasons. Effortless systems do not screen. A system that does not screen will sweep in the exact people for whom the act is forbidden. Then we punish those people as though the effortlessness had been their idea.
⚖️ Why the Law Offers So Little Cover
You would think honest confusion would be a defense. Mostly, it is not.
Federal law treats voting by a noncitizen as a general-intent offense. The government does not have to prove you knew that only citizens may vote. It only has to prove you meant to do the thing you did — fill out the form, cast the ballot. Your sincere belief that you were allowed is, for purposes of liability, beside the point.
The courts have left one narrow escape hatch, and its shape tells you everything about how tight the space is. If a noncitizen makes complete and accurate disclosure of his status to an official who actually has authority over voter eligibility, and that official affirmatively registers him anyway, the vote may not violate the statute. That is the doctrine of official authorization, recognized in Keathley v. Holder, 696 F.3d 644 (7th Cir. 2012).
But the same court drew the boundary the same year. In Kimani v. Holder, 695 F.3d 666 (7th Cir. 2012), Judge Easterbrook explained that an official can only authorize what his office actually empowers him to authorize — his blunt illustration was that “the janitor of a public school can’t authorize anyone to sell cocaine.” A DMV clerk sliding a form across the counter is not conferring the franchise. And in Fitzpatrick v. Sessions, 847 F.3d 913 (7th Cir. 2017), the hatch slammed shut on a woman who had done nearly everything right, because on the same form she had checked a box claiming to be a citizen. One misrepresentation she made herself, and the defense was gone.
Read those three together and the civic-design problem becomes a legal one. The system invites registration with no friction. The law then demands a level of documented, deliberate, perfectly-disclosed care that the frictionless system never asked anyone to exercise. The gap between what the counter requires and what the courtroom requires is where real people fall.
📊 The Size of the Problem We Are Actually Solving
Now set the human design problem next to the thing it is supposedly protecting against, because the mismatch is staggering.
Every serious audit of noncitizen voting lands near zero. Michigan checked all 5.7 million ballots cast in 2024 against state driving records and found about fifteen that appeared to come from noncitizens. Georgia reviewed 8.2 million registrations and found nine people who had ever cast a ballot. The Heritage Foundation — an organization that built its database specifically to document election fraud — has logged 77 instances of noncitizen voting nationwide across twenty-four years.
And the accusations run overwhelmingly false. In Public Interest Legal Foundation v. North Carolina State Board of Elections, 996 F.3d 257 (4th Cir. 2021), the record showed that 97.6 percent of the people a motor vehicle agency flagged as noncitizens were in fact United States citizens. That is a finding in a published federal appellate opinion, not a talking point.
~15
Apparent noncitizen ballots in Michigan, out of 5.7 million cast in 2024.
77
Instances nationwide in Heritage’s own database — across twenty-four years.
97.6%
Of people flagged as noncitizens who turned out to be citizens, in the one court-tested screen.
So the frictionless system is not producing a wave of unlawful votes. It is producing a trickle so small that the country’s most motivated investigators need decades to reach three digits. What it does produce, reliably, is the occasional confused lawful permanent resident who trusted the counter — and a screening apparatus that mislabels citizens forty-one times for every real case it finds.
🗂️ The Real Repair Nobody Is Selling
If you actually wanted to improve the accuracy of the rolls, you would not be hunting noncitizens, because there is almost nothing there to hunt. You would be fixing residency.
Citizenship is objective and verifiable — there is a certificate, a file number, a date. That is why the audits keep returning near zero. Residency is the hard one. Americans move constantly, and their registrations do not move with them. Stale records, cross-state duplicates, addresses where nobody has lived in years: that is where the genuine mess sits, and being registered in two places is not the same as voting in two places. It is a records problem, not a fraud problem, right up until someone casts a ballot where they no longer live.
There was a tool built for exactly this. The Electronic Registration Information Center, founded in 2012 by the Pew Charitable Trusts and seven states, pooled voter and motor vehicle data across member states to flag movers, duplicates, and the deceased. It was bipartisan, and it worked. Then, beginning in 2022, nine Republican-led states walked out amid the fraud panic — and because the system only catches a cross-state move when both states belong, every departure degraded the data for everyone who stayed. The Heritage Foundation now publishes work asking for it to be rehabilitated, because the states that left have no comparable replacement.
Sit with that sequence. The one competent, data-driven instrument aimed at the real defect in American voter rolls was dismantled by the same movement now demanding emergency powers to fix American voter rolls.
🗳️ So What Should an Eligible Voter Do?
Use the tools. For the overwhelming majority of readers, who are citizens, that is the whole of it.
If you are eligible to vote, the frictionless machine is a gift. Register. Check your registration if you have moved. TurboVote and services like it exist to make sure a clerical gap or an old address does not cost you your vote, and they are worth using precisely because the residency problem above is real and mostly invisible to the individual voter. The friction we removed was friction that used to keep eligible people out. That was worth removing.
✅ How to Check Your Own Registration — Free, in Five Minutes
If you are a U.S. citizen, do not assume your registration is current, especially if you have moved, changed your name, or skipped a couple of elections. States quietly remove voters who look inactive, and the residency mismatches described above are exactly the kind of thing that knocks a real voter off the rolls. Checking is free and takes a few minutes.
Look yourself up. Go to TurboVote or your state’s official election website and confirm you are registered at your current address. If your address is old, update it now — not in October.
Set a reminder. TurboVote will text or email you before each election, with your polling place and dates. This is the single easiest way to never be caught by a deadline.
Re-check after any move. A registration does not follow you across town, across a county line, or across a state. Every move is a reason to look again.
Fix problems early. If you find you have been dropped or flagged, most states let you re-register or correct the record well before Election Day. The mistakes are routine and fixable when you catch them in time.
🛑 One Hard Line, If You Are Not a Citizen
If you are not a United States citizen, do not register and do not vote — no matter who hands you the form, no matter what a clerk at a counter tells you, no matter that the website let you through. A registration card is not permission. A clerk is not an authority on your eligibility. The frictionless system will not stop you, and the law will not forgive you afterward. If anyone offers to register you, the answer is no. And if you have already registered or voted, do not file anything and do not talk to any agency before you speak with an immigration lawyer.
🎯 The Point
We made voting easy on purpose, and we were right to. We then built a body of law that treats the easiest civic act in America as a trap for one specific person, and we pretend the trap is his fault.
The registration is frictionless by design. The punishment is severe by design. Only the confusion in between belongs to the immigrant — and we hand it to him.
The banner that sends you to TurboVote is a small, decent thing. It reflects a judgment that people deserve accurate help when they show up asking how to participate. I agree with that judgment completely. I only want the same care extended to the person standing one step to the side of it — the one the frictionless machine was never built to notice, and the one our law is least willing to forgive.
Disclaimer. This post is commentary and general legal analysis, not legal advice, and reading it creates no attorney-client relationship. Immigration consequences of voting are severe and fact-specific. If you are not a U.S. citizen and you registered to vote or cast a ballot, consult a licensed immigration attorney before filing anything or speaking with any government agency. TurboVote is a nonpartisan service of Democracy Works; the reference here is descriptive and does not imply any endorsement in either direction.
Law Offices of Michael D. Baker · 111 W. Jackson Blvd., Suite 1700 · Chicago, IL 60604
For years the rule at the airport was simple. ICE took the people who had already lost — final orders of removal, names already in a system, no appeal left to file. You could tell a client who was at risk. Usually it was not the woman with a pending case and a work permit in her bag.
That rule is gone.
On July 30 the Washington Post reported that ICE has started pulling a different traveler off domestic flights: people who came in legally and overstayed. Asylum seekers whose cases are still open. People holding valid employment authorization. No criminal charge of any kind. One man, identified only as Paul, landed at Dulles after a family trip to Hawaii and was met at the front of the plane by two plainclothes officers. During processing his chest tightened and he collapsed from a heart attack. He holds a withholding order protecting him from return to India under the Convention Against Torture. His lawyer called the airport spike an entirely new level of aggressiveness. In a separate case, agents arrested a Colombian woman with a pending asylum claim at Baltimore-Washington International as she waited to board a flight to Florida with her U.S.-citizen fiancé.
Here is how they find him.
The machine under the arrest
When you book a domestic flight, TSA collects your name, date of birth, and gender. That collection is not new, and it is not sinister on its own. It runs on the Secure Flight Final Rule, 73 Fed. Reg. 64018 (Oct. 28, 2008), codified at 49 C.F.R. Part 1560. Congress ordered it after 2001 for one reason: to check passengers against terrorist watch lists. Counterterrorism. That is the purpose the law authorizes.
In May 2025, TSA and ICE signed a Memorandum of Agreement that routes that same passenger data toward immigration enforcement. The public never saw it. American Oversight had to sue under the Freedom of Information Act to get it, and released it on July 28. The agreement cites Secure Flight as its legal footing, says TSA will hand ICE specified data elements, and says ICE will keep what it receives in its own repository. The appendices that would tell you exactly which data — those are redacted almost to the last line.
📄 Where this comes from The agreement was produced only after American Oversight sued TSA and ICE under FOIA. The watchdog reports that TSA shared information on more than 31,000 travelers, leading to more than 800 immigration-related arrests, and that the data-sharing was first revealed in December 2025 before the formal agreement surfaced this summer. The redacted appendices mean the full list of shared data elements is still not public.
There is a second problem, and it is about candor. In January, before the House Homeland Security Committee, the acting TSA Administrator was asked directly whether TSA had an agreement to share passenger information with ICE to check travelers for deportation orders. She said that was not what was happening — that TSA does not send the information to ICE, but only helps ICE check against TSA’s information. The agreement American Oversight pried loose describes a formal data-sharing arrangement and lays out how ICE retains what TSA gives it. Read the testimony against the document and they do not sit easily together. That is my read, not a ruling. But it is the kind of gap oversight committees exist to close, and so far the administration has declined to close it: at his confirmation hearing this month, the nominee to run TSA would not tell the Senate what TSA shares.
The government’s answer is short. DHS says information sharing between ICE and TSA helps identify who is trying to travel in the country, including violent criminals, and keeps the skies safe. That is the defense. It does not explain why the target has moved from people with removal orders to people with pending cases and work permits.
What this actually does
A domestic boarding pass now runs your identity against ICE’s target lists, and TSA’s passenger data feeds those lists. For a noncitizen without secure status, flying inside the United States means volunteering your name, your itinerary, and your exact location to the agency looking for you.
The shift is easiest to see side by side. What changed is not the raw authority to arrest at an airport. What changed is who gets caught, and how the government locates them.
The old pattern
What lawyers are seeing now
Who is arrested
People with final orders of removal or prior deportation orders
Also people who entered legally and overstayed a visa
Their status
Out of options, already ordered removed
Pending asylum, valid work permits, no criminal charge
How they are found
Known targets ICE was already tracking
Passenger data matched through the TSA feed
Data source
ICE’s own enforcement records
Secure Flight passenger data, collected for watch-list matching
Paper trail
Ordinary enforcement
A May 2025 agreement hidden until a FOIA lawsuit forced its release
The newest chokepoint
This is not an isolated program. It is the newest chokepoint in a pattern this blog has tracked all year. The licensing system reached Dreamers through their commercial driver’s licenses. USCIS reaches them through renewal queues that let a work permit lapse in the gap — the deferred-action squeeze I wrote about when three rulings in thirty days dropped the floor out for SIJ and DACA recipients. The mega-master calendar reaches them through the hearing itself. Each one finds people not by investigating them but by sitting on infrastructure they cannot avoid. The airport is the same move. The government already knows you have to fly, so it waits where the data already is.
And notice who it catches. Not fugitives. A man with Convention Against Torture protection. A woman with a pending asylum claim and an American fiancé. These are the same people my DACA and deferred-action posts keep returning to, and the same population behind Chicago’s collision between a growing immigrant city and the hardest enforcement in a generation — lawfully present or lawfully mid-process, doing what the system asks, and finding that the compliance is what makes them easy to locate.
The real danger is what follows the arrest
The arrest itself is not the sharpest risk. What can follow it is.
⚠️ The expedited-removal trap The administration expanded expedited removal nationwide in January 2025 under 8 U.S.C. § 1225(b)(1). A person who cannot prove two years of continuous physical presence on the spot can be removed without ever seeing an immigration judge. An airport is the worst possible place to be asked for that proof, with no file, no lawyer, and no time.
If detention follows instead of summary removal, the fight moves to bond — the same mandatory-detention battle I covered in Cunha v. Freden. The Colombian woman arrested at BWI won her release on bond this week. The door is not welded shut. But it is a fight, and it starts from inside a detention center far from home.
What to do now
If you do not hold a green card or citizenship, treat a domestic flight as a decision, not a routine. Weigh it before you book.
Carry proof of more than two years of continuous physical presence whenever you travel — leases, pay records, medical and school records. This is the single most important defense against expedited removal.
Know the airport is interior enforcement. You keep the right to remain silent and the right to refuse consent to a search. Say plainly that you wish to remain silent and to speak to a lawyer.
Sign nothing. A signature can waive a hearing you did not know you had.
Give family a plan before you leave: who to call, which attorney, where the documents are kept.
Counsel: ask new clients about recent and planned air travel as a matter of routine intake. It is now a risk factor.
Where this goes next
The FOIA litigation is not over. American Oversight says it will keep pushing for how the data moves, what safeguards exist, and how many travelers are swept in. Congress can demand the same, and a nominee who will not answer a senator’s direct question is a fight waiting to happen. Whether repurposing counterterrorism data for immigration targeting satisfies the Privacy Act is a real question, and no court has answered it yet.
The government has quietly made the airline manifest an enforcement tool. It did it with data you were required to hand over to fly, through an agreement it kept out of public view until a lawsuit dragged it into the light.
Until the courts or Congress force more into the open, the honest advice is the plain kind. If your status is not secure, the airport is no longer neutral ground. Fly carefully, carry your proof, and know your rights before an officer is standing over your seat.
This post is for informational purposes only and is not legal advice. Reading it does not create an attorney-client relationship. Consult a licensed immigration attorney about your specific situation.
🔄 Updates
August 6, 2026 — Initial publication. Monitoring the American Oversight FOIA litigation for the unredacted data-element appendices, the Senate’s follow-up to the TSA nominee’s confirmation testimony, and any Privacy Act challenge to the TSA–ICE data sharing.
The Migration Policy Institute says the Mexican-born population of the United States is getting smaller. Nationally, that holds. The count peaked above 11.7 million in 2007 and had slipped to roughly 11.1 million by 2024, even as the total immigrant population kept climbing. Chicago went the other way.
United States
11.7M → 11.1M
Mexican-born population, 2007 peak to 2024. Down.
City of Chicago
207.8K → 227.7K
Mexican-born residents, 2023 to 2024 — a 9% jump, the largest of any group here. Up.
Two numbers, opposite directions. The gap is the story. And it matters because the population that grew in Chicago in 2024 became, in the months that followed, the primary target of the largest federal immigration operation the city has seen in a generation.
What the national data show
MPI’s Spotlight, by Andres Ayala and Jeanne Batalova, describes a population that is large and long-settled. Mexicans remain the biggest immigrant group in the country, about 22 percent of 50.2 million immigrants. Most arrived before 2010. When Mexicans obtain green cards, they do it overwhelmingly through family — roughly 90 percent of Mexican green-card recipients in FY 2024, against 63 percent for all new permanent residents. Labor-force participation runs high. Educational attainment and income run below the national immigrant average. About a third are naturalized citizens.
Put plainly: a settled, working, family-based population, most of it here more than fifteen years, most of it still without citizenship.
Border numbers have collapsed, and two of them are worth keeping straight. Border Patrol apprehensions between the official ports of entry — people caught crossing illegally, away from the checkpoints — fell to about 237,500 on the Southwest border in FY 2025, the lowest since 1970; Mexicans were 47 percent of them. Total Southwest border encounters, which add everyone processed at the ports of entry themselves — those found inadmissible or asking for protection — came to about 444,000, down from 2.1 million a year earlier. The first number is a slice of the second. Both fell by roughly 80 percent. Whatever is shrinking the national count, it is not new arrivals from Mexico.
The 2.5 million problem
So what is shrinking it? The government has an answer. In December 2025, DHS announced that more than 2.5 million people had “left” since January. Look closer.
What DHS counts
Figure
How it is known
Deportations (formal removals)
605,000
Counted
“Voluntary self-deportations”
1.9 million
Estimated from population surveys; its own CBP Home app logged about 13,000 by mid-2025
Headline total
2.5 million
Press release
Independent removal estimates (Brookings, MPI)
310,000–340,000
Analysts counting actual removals
Treat the 2.5 million as what it is: a press release, not a count. Fear moves people, and the enforcement is real. A verified exodus of two and a half million is not.
Chicago against the trend
Illinois holds about 6 percent of the country’s Mexican immigrants, third behind California and Texas. Cook County is one of five counties — with Los Angeles, Harris, Maricopa, and Riverside — that together house roughly a quarter of every Mexican immigrant in the United States.
The 2024 city increase is real, but it needs context. Over the 2010s the metro area’s Mexican-born population fell, Cook County alone shedding tens of thousands. The 2024 figure reverses that slide. It comes from a single year of American Community Survey estimates, which swing more than multiyear data, so read it as a genuine uptick and not a boom. The longer truth is a population that thinned for a decade, then ticked back up in the year before the raids.
Where Mexican immigrants concentrate (MPI, 2020–24)
Places
Top 5 metro areas (≈35% of all U.S. Mexican immigrants combined)
Los Angeles, Chicago, Houston, Dallas, Riverside
Top 5 counties (≈24% combined)
Los Angeles (CA), Harris (TX), Cook (IL), Maricopa (AZ), Riverside (CA)
≈597,400 — about 22% of the city; highest since 2006
Cook County population
≈5.2 million, second-largest U.S. county
Illinois active DACA recipients (Sept. 2025)
≈26,850
Illinois share of U.S. Mexican immigrants
≈6% (third nationally)
The profile is the caseload
For a lawyer, the demographic profile is not abstract. It predicts who walks in the door.
A population that is largely noncitizen, family-based, employed, and long-term resident is exactly the population most exposed when enforcement turns inward. These are not recent crossers with thin equities. They are people with a decade or more in the country, U.S.-citizen children, mortgages, and jobs — and, in many cases, no citizenship and no clean path to it. About a third of the national Mexican-born population lacks health insurance, the highest rate of any group MPI measured, which tells you something about how many work off the formal grid.
DACA, concentrated
Mexicans are about 81 percent of the roughly 495,300 active DACA recipients nationwide as of December 2025. Illinois had about 26,850 as of September 2025. Every one of them is a person the government has on record as removable, protected only by a policy that has spent years in court. A shrinking national headcount does not shrink that exposure. It concentrates it in places like Chicago, where the population stayed.
The enforcement overlay
Operation Midway Blitz
Began September 8, 2025, with more than 200 federal agents, some pulled from Los Angeles. The early phase had a face — Border Patrol commander-at-large Gregory Bovino — and a look: riot gear, pepper balls, running clashes through Pilsen and Little Village, the city’s Mexican core. After two U.S. citizens were shot and killed during his later surge in Minneapolis, Bovino was pulled from the role in January 2026 and replaced by border czar Tom Homan, who promised “targeted” operations instead. The spectacle gave way to something quieter and harder to see — arrests at courthouses and at immigration check-ins, where the government already knows exactly where you will be and when.
The guardrail: Castañon Nava v. DHS
A 2022 settlement restricts warrantless immigration arrests in Illinois and five other states and requires probable cause plus a flight-risk showing before an arrest. Reported violations of that order have already produced court-ordered releases. Illinois lawmakers, meanwhile, stood up an accountability commission to document the operation, and its work carried into 2026.
None of this is peculiar to Chicago. The Migration Policy Institute, reading the same FY 2025 federal data, describes the through-line: once border crossings collapsed, the administration redirected its enforcement inward — toward the interior, toward cities. Chicago is where that turn pointed. Which raises a question the law has an answer to: where does “the border” end and “the interior” begin?
What “the border” means in law
The border is not a line; it is a zone. Immigration officers may run warrantless searches for noncitizens “within a reasonable distance from any external boundary” (INA § 287(a)(3), 8 U.S.C. § 1357(a)(3)), and a regulation fixes that distance at 100 air miles (8 C.F.R. § 287.1(a)(2)). By the ACLU’s count, that band holds about two-thirds of the U.S. population and nine of the ten largest cities.
Expedited removal — deportation with no hearing, under 8 U.S.C. § 1225(b)(1) — was long a border tool. The 2004 rule reached only people caught within 100 miles of the border and present under 14 days. On January 21, 2025, DHS stretched it to anyone, anywhere, who cannot prove two years’ continuous presence. A district court paused the expansion in August 2025; in 2026 the D.C. Circuit let it take effect while the case goes on.
Chicago sits more than 100 air miles from the nearest international land border.
Here is the rub. The administration is treating the whole country as a border. It has taken tools built for the physical line — warrantless stops inside the hundred-mile zone, summary removal without a hearing — and pushed them inland, to Chicago and cities like it. The rationale amounts to this: the statute allows it, so they are doing it. That is ambitious, and it is a mess. It runs into the Fourth Amendment’s limits on suspicionless stops and the Fifth Amendment’s promise of a hearing, and it turns a long-settled resident’s ability to document two years into the only thing between them and a one-day deportation. A border is supposed to be a place. The government is trying to make it a condition that follows you.
Watch how it works now. The New York Times reported that TSA — the agency that screens you for a domestic flight — has been feeding passenger data to ICE, which mines it to find travelers to arrest. Agents have pulled people aside at check-in counters, gates, and jet bridges at more than a dozen airports, including travelers with pending applications and no criminal charge of any kind. A domestic boarding pass is not a border crossing. The government is treating it like one.
The national “decline” narrative can lull people into thinking the pressure is off. In Chicago it is the opposite. The population here is present, rooted, and growing, which is precisely why enforcement in this city falls on settled families rather than on anyone who just crossed a border. The people most likely to be taken are the people most likely to have been here longest.
The other big cities
Chicago is not alone, and it is not the largest concentration. Los Angeles County remains the single biggest home for Mexican immigrants in the country. Greater Los Angeles, Houston, Dallas, and Riverside round out the top five metros with Chicago. Where MPI reports the share of a metro’s residents born in Mexico, the border regions lead — roughly a quarter of the McAllen–Edinburg–Mission area — followed by Riverside and greater Los Angeles near 11 to 12 percent. One population, spread thin across a few states, now worked city by city.
Where it goes next — watch three things
The Castañon Nava enforcement litigation — the closest thing to a brake on warrantless arrests in Illinois.
The accountability commission’s findings — which may feed further suits.
DACA — still unresolved, still the only thing standing between tens of thousands of Illinois residents and the removability they acknowledged when they applied.
Two numbers, opposite directions. The count went up. So did the raids. Both are true, and Chicago is where they meet.
This post is general information, not legal advice, and does not create an attorney-client relationship. If you or a family member face immigration enforcement, consult a licensed immigration attorney about your specific situation. — Law Offices of Michael D. Baker
Much of the data here — the demographic profile, the metro and county distribution, the unauthorized and DACA estimates — comes from the Migration Policy Institute, a nonpartisan research institution founded in 2001. MPI builds its numbers from Census and federal data, and it shows its work. That matters. When the government’s own headline is a press release — 2.5 million “left” — the count you can actually check comes from independent institutions like MPI. Explore their research at migrationpolicy.org.
Sources
National profile and Illinois share: Ayala & Batalova, Mexican Immigrants in the United States (MPI, 2025). Demographic profile: U.S. Census Bureau, 2024 ACS S0201; Income, Poverty, and Health Insurance: 2024. Green cards / family-based: DHS OHSS, 2024 Yearbook, Table 3. Border encounters: CBP, Nationwide Encounters; Pew (Gramlich, 2026). Unauthorized population: Gelatt, Ruiz Soto & Bachmeier (MPI, 2025). DACA (national, Dec. 31, 2025): USCIS, Active DACA Recipients, FY2026 Q1; MPI DACA data tools. Illinois figure (Sept. 30, 2025): USCIS, Active DACA Recipients by State. Metro and county distribution: MPI Migration Data Hub. Chicago ACS figures: U.S. Census Bureau, 2024 ACS, as reported by WBEZ (Sept. 12, 2025). Cook County population: U.S. Census Bureau, 2024 estimates. Enforcement: contemporaneous reporting (Chicago Sun-Times, USA Today, Block Club Chicago, WBEZ); Castañon Nava v. DHS settlement (2022); Ruiz Soto, A New Era of Immigration Enforcement (MPI, Oct. 2025). National “departures” claim and its limits: DHS, More than 2.5 Million Illegal Aliens Left the U.S. (Dec. 10, 2025); independent removal estimates near 310,000–340,000 from MPI and Brookings. Interior enforcement and the 100-mile zone: 8 U.S.C. § 1357(a)(3); 8 C.F.R. § 287.1(a)(2); expedited removal, 8 U.S.C. § 1225(b)(1); Designating Aliens for Expedited Removal, 69 Fed. Reg. 48877 (Aug. 11, 2004) and the DHS expansion notice (Jan. 21, 2025); ACLU, The Constitution in the 100-Mile Border Zone. Chicago operation and airport enforcement: reporting on Operation Midway Blitz, Gregory Bovino, and Tom Homan (Chicago Sun-Times, NBC Chicago, ABC7, AP); The New York Times on TSA–ICE passenger-data sharing and arrests of domestic travelers at U.S. airports.
The Board made unlawful voting a significant adverse factor and held that taking the Fifth can be used against you. Fourteen years between the ballot and the order.
Matter of Lopez Rico, 29 I&N Dec. 780 (BIA 2026), Interim Decision #4223, decided July 23, 2026. Read the decision (PDF).
By Michael D. Baker
A man voted once, in 2012, believing he was allowed to. Fourteen years later the Board of Immigration Appeals ordered him removed and told the judge below he should have punished that vote harder.
Along the way the man did something every American thinks he has a right to do. A judge asked him about his drunk driving. He took the Fifth. The Board just held that his silence can be used against him — and that holding reaches every immigrant who ever walks into an immigration courtroom with an open criminal matter.
Two rules came out of Matter of Lopez Rico on July 23, 2026. Both are bigger than the case.
1
Silence Costs You
Invoke the Fifth Amendment on questions that matter to whether you deserve mercy, and the judge may hold your silence against you.
2
The Vote Counts Heavily
Unlawful voting is now a significant adverse factor in every discretionary decision — even where the judge finds you thought you were eligible.
📅 What Actually Happened
Omar Eliodoro Lopez Rico came to the United States as a lawful permanent resident in 2001. He voted once, in California, in the 2012 presidential election. Then he reported it himself — on his naturalization application and again in a sworn statement. That disclosure is what triggered everything that followed.
Two drunk driving matters sat in the file — a 2016 conviction and a 2024 arrest still pending. The Immigration Judge asked him to explain both. He refused to answer, and he had a reason: anything he said could be used in the open criminal case.
So he took the Fifth. That decision cost him the country.
🤐 Issue One: Taking the Fifth in a Civil Courtroom
Here is the thing most people do not know, and it is the hinge of this entire case.
🔑 The Point Everyone Misses
Immigration court is not criminal court. It is a civil proceeding. In a criminal trial the government must prove you guilty, and your silence is free — a prosecutor cannot even comment on it. In immigration court, when you ask for relief, you carry the burden. You are the one who has to prove something.
So when you stay silent, you are not making the government’s case harder. You are leaving a hole in your own.
The Board was blunt about it. The Immigration Judge gave Lopez Rico repeated chances to explain the DUI conviction and the pending arrest. He invoked the privilege and declined. The Board held the Judge could draw reasonable adverse inferences from that refusal — that the silence itself left an evidentiary gap, and the gap ran against the man who created it.
The Board did not invent this. Matter of Marques, 16 I&N Dec. 314 (BIA 1977), said it forty-nine years ago: a respondent may assert the privilege, but in doing so he runs the risk of failing to carry his burden of persuasion. Matter of Li, 15 I&N Dec. 514 (BIA 1975), and Matter of Mariani, 11 I&N Dec. 210 (BIA 1965), say the same. The Supreme Court blessed adverse inferences from silence in immigration proceedings over a century ago in United States ex rel. Bilokumsky v. Tod, 263 U.S. 149 (1923), and again for civil cases generally in Baxter v. Palmigiano, 425 U.S. 308 (1976). Pereida v. Wilkinson, 592 U.S. 224 (2021), added the modern gloss: evidentiary gaps work against the person seeking relief.
What Lopez Rico does is make it precedent, in plain terms, in 2026 — at a moment when a great many immigrants have open criminal matters and every incentive to say nothing.
The Board did draw one line worth holding onto. Adverse inferences cannot carry the government’s burden of proving you removable in the first place. Matter of Guevara, 20 I&N Dec. 238 (BIA 1991). Silence cannot deport you. It can only sink your request for mercy.
⚠️ The Trap
An immigrant with a pending criminal charge now faces a choice with no good exit. Testify, and the transcript lands on a prosecutor’s desk. Stay silent, and the immigration judge holds the silence against you and denies the relief that would let you stay.
That is not a hard case. That is a rigged one.
🗳️ Issue Two: One Vote, Weighed Heavily
The second holding is where the Board made genuinely new law.
Lopez Rico applied for cancellation of removal — the form of relief that lets a long-term green card holder stay despite being deportable. Qualifying is not enough. A judge weighs your good against your bad and decides whether you deserve it. That weighing is called discretion, and it is where most of these cases are actually won and lost.
The Board ran the standard balance from Matter of C-V-T-, 22 I&N Dec. 7 (BIA 1998), and Matter of Marin, 16 I&N Dec. 581 (BIA 1978).
HIS EQUITIES
25 years as a resident
U.S. citizen wife and children
Steady employment history
Sole provider for the household
THE ADVERSE SIDE
2016 DUI conviction
2024 DUI arrest, still pending
Silence on both
One vote, in 2012
The Board called his equities substantial. It affirmed the denial anyway.
Then the Board reached past this record. It grounded the weight of unlawful voting in Congress’s twin decisions to criminalize the conduct at 18 U.S.C. § 611(a) and to make it a ground of removal at INA § 237(a)(6), 8 U.S.C. § 1227(a)(6). It invoked the governmental interest in public confidence in elections, citing Crawford v. Marion County Election Board, 553 U.S. 181 (2008), Brnovich v. Democratic National Committee, 594 U.S. 647 (2021), Doe v. Reed, 561 U.S. 186 (2010), and Purcell v. Gonzalez, 549 U.S. 1 (2006).
And then it wrote a sentence that will outlive this case. The Immigration Judge, the Board said, could have given the unlawful voting violation greater adverse weight than he did.
Read that again. The judge already denied relief. He already ordered the man deported. The Board still said he went too easy.
⚖️ Ignorance Is Not a Defense
The Immigration Judge made three findings that ought to have mattered. Lopez Rico mistakenly believed he was eligible to vote. He did not know it was unlawful. Somebody encouraged him to register.
None of it saved him, and the federal courts explain why. Voting unlawfully as a noncitizen is a general intent offense. The government does not have to prove you knew you were breaking the law. It only has to prove you meant to do the thing you did — walk in, take a ballot, fill it out.
Three federal appellate decisions built that architecture, and they still control.
📚 The Three Cases
Keathley v. Holder, 696 F.3d 644 (7th Cir. 2012) — the one that worked. A woman disclosed her alienage completely and accurately, and an official registered her anyway. The court sent the case back. Judge Easterbrook later named the doctrine: official authorization.
Kimani v. Holder, 695 F.3d 666 (7th Cir. 2012) — released the same day, and it drew the outer limit. Two rulings. First, § 611 is general intent: the government need not prove the voter knew the rule. Second, the escape hatch only opens if the official actually held the power to grant what he appeared to grant. Judge Easterbrook illustrated the boundary of official authority bluntly — “The janitor of a public school can’t authorize anyone to sell cocaine.” Handing someone a registration card is not conferring the franchise. Kimani lost on a second ground as well: he had represented himself as a citizen, and under Bayo v. Napolitano, 593 F.3d 495 (7th Cir. 2010) (en banc), you are bound by what you sign whether or not you read it.
Fitzpatrick v. Sessions, 847 F.3d 913 (7th Cir. 2017) — the door closes. A clerk asked if she wanted to register. She asked whether she was supposed to. He said it was up to her. She lost, because on the same form she had checked a box claiming to be a citizen. I wrote about it in 2017.
The Board also relied on Matter of Fitzpatrick, 26 I&N Dec. 559 (BIA 2015), and distinguished McDonald v. Gonzales, 400 F.3d 684 (9th Cir. 2005), where a state statute did require proof the voter knew he was ineligible. That distinction is the whole ballgame: the federal statute has no such requirement.
So the architecture was already settled. What Lopez Rico adds is a second bite. Even after the vote makes you deportable, the same vote comes back — now as a heavy weight on the scale when you ask to stay.
👤 Now Strip Away the Baggage
Lopez Rico was an easy case to affirm. Two drunk driving matters. A refusal to explain either one. Put that record in front of any judge and the outcome looks narrow — a bad-facts case, decided on its bad facts.
It is not narrow. To see why, take everything ugly out of the file and run the same rules again.
Call her Rosa. She is a composite, assembled from real matters and rearranged so that nobody is identifiable.
Rosa got her green card in 2004. In 2006 she stood at a government counter renewing her driver’s license. The clerk slid a form across and asked whether she wanted to register to vote. Rosa asked whether she was supposed to. He shrugged. He said it was up to her.
She checked the box. A registration card came in the mail with her name printed on it and a state seal on the front. She voted twice. She told her sister she felt like she belonged.
Nine years later Rosa filed for citizenship. The form asks whether you have ever voted in a federal election. She answered yes, because she had no idea she had done anything wrong.
That honest answer is what put her in removal proceedings. Now watch the three rules land on a woman with a clean record.
⚖️ The Rules, Applied to Rosa
General intent gets her. She did not know noncitizens could not vote. Under Kimani, that does not matter. She meant to fill out the ballot, and the statute asks nothing further. She is removable.
Official authorization does not save her. A shrug is not permission, and a clerk is not an election authority. Fitzpatrick forecloses the argument on facts nearly identical to hers.
Lopez Rico weighs the vote against her a second time. The same ballot that made her deportable now returns as a significant adverse factor when she asks a judge for permission to stay.
Rosa has no DUI. No pending charge. No reason to invoke the Fifth Amendment. Nothing in her file resembles the record that made Lopez Rico an easy affirmance.
She still loses ground on all three, and she loses it for one confused afternoon at a government counter twenty years ago.
That is the reach of this decision. Strip away the drunk driving and the silence, and the vote alone still sits on the scale — heavily, and by instruction from the Board.
No prosecutor ever wanted Rosa’s case. No U.S. Attorney has ever charged her with anything. The government does not think she is a criminal. The government thinks she is deportable, which is an entirely different thing, and the difference is the country she raised her children in.
🔍 Now the Part Nobody Says Out Loud
The Board justified all of this with the governmental interest in election integrity. So let us look at the size of the threat.
Every serious audit conducted in this country lands in the same place.
0.00028%
Michigan, 2024. About fifteen apparent noncitizen ballots out of more than 5.7 million cast.
9
Georgia audited 8.2 million registered voters. Twenty noncitizens on the rolls. Nine had ever voted.
1
Ohio flagged 521 potential cases over four years. Exactly one produced a charge.
77
Total instances nationwide, 1999–2023, in the Heritage Foundation’s own fraud database. Twenty-four years.
And the accusations themselves are mostly false. In Public Interest Legal Foundation v. North Carolina State Board of Elections, 996 F.3d 257 (4th Cir. 2021), the record showed that 97.6 percent of the people the motor vehicle department flagged as noncitizens turned out to be United States citizens. That is not an advocacy statistic. That is a finding in a published federal appellate opinion.
The 250,000
On July 17, 2026, Homeland Security Secretary Markwayne Mullin sent letters to four states claiming a quarter of a million noncitizens sat on their voter rolls. California, 190,832. New Jersey, 35,152. Nevada, 15,903. Pennsylvania, 14,576.
Look at that list again. Those are not the four states with the most immigrants. Those are the four states that refused to hand their voter files over to the federal government. The number does not track immigration. It tracks defiance.
Where did the figures come from? A White House official told reporters the estimate came from an analysis of commercial databases. Not government records. Not citizenship files. Marketing data — the kind that misspells your name and mails catalogs to people who moved out in 2011. David Becker of the Center for Election Innovation and Research put it plainly: the data almost certainly includes a great many people who are eligible voters, and states would likely break the law by removing them.
Then Nevada did the arithmetic in public, and the whole thing fell apart.
🧮 The Nevada Problem
Nevada has about 2.1 million active registered voters. Of those, 138 registered without providing a driver’s license or Social Security number — and even those 138 may have used another acceptable ID.
So the entire universe of Nevada registrants with any documentation gap at all is 138 people.
The federal claim was 15,903.
That is not an overcount. That is a number with no relationship to the underlying records. It exceeds the maximum arithmetically possible answer by a factor of 115.
The 250,000 figure also describes registration, not voting. Nobody claimed these people cast ballots. And even taken at face value, 250,000 is about one tenth of one percent of registered voters nationwide — roughly six tenths of one percent of the forty million registered across those four states.
The courts have been unimpressed throughout. A federal judge found the administration acted unlawfully in building its centralized citizenship database, and found that some states had used it to strip actual citizens off the rolls. Two executive orders on proof of citizenship have been blocked. The Justice Department sued thirty states and the District of Columbia for their voter files; judges have thrown out sixteen of those suits.
🧮 How an Actuary Would Look at This
Set the politics down for a minute and think like someone who prices risk for a living.
An actuary does not ask whether a thing is possible. Everything is possible. An actuary asks three questions: how often does it happen, how much damage does it do, and how good is your method of detecting it. Run those three questions here and the program collapses on every one.
1 in 380,000
Frequency. Michigan’s rate — roughly fifteen ballots out of 5.7 million.
0
Severity. Elections ever shown to have been decided by noncitizen votes.
2.4%
Precision. Share of flagged “noncitizens” who actually were noncitizens, in the one case a court examined.
An actuary multiplies frequency by severity to get expected loss. Multiply a rate of one in 380,000 by an impact of zero demonstrated changed outcomes and you get a number no insurer would bother to write down. In the trade it has a name: de minimis exposure. You note it in a footnote. You do not build a program around it.
But the third number is the one that ends the argument, and it is the number nobody in this fight wants to discuss.
⚡ Why Screening for Something This Rare Cannot Work
When you hunt for something that occurs roughly once in 400,000 cases, using a test that is even slightly imprecise, nearly everyone you catch is innocent. This is not opinion. It is arithmetic, and it is the same arithmetic that governs cancer screening, airport bomb detection, and fraud alerts on your credit card.
The Fourth Circuit record gives us the real-world precision of exactly this kind of screen: 97.6 percent of the people flagged were citizens. For every one correct hit, roughly forty-one Americans got accused of a federal crime by a database.
Now apply that ratio to the 250,000. You would expect on the order of 6,000 real registrations and 244,000 wrongly accused United States citizens.
That is not a detection system. That is a machine for stripping citizens off the voter rolls, and it is performing exactly as its mathematics require.
So here is where we are. The country has spent executive orders, thirty-one federal lawsuits, a rebuilt federal database later ruled unlawful, letters from a cabinet secretary, a prime-time presidential address, and an unmeasurable quantity of congressional floor time — hunting a thing that happens about once in every 380,000 ballots, with a tool that is wrong roughly forty-one times for every time it is right.
No underwriter alive would approve that. No business would run it. If a client brought me that risk model I would tell him he was not managing a hazard. He was building a pretext.
🗂️ The Real Problem, Which Nobody Is Working On
Now let me say the thing that gets lost when a debate turns into theater.
The voter rolls in this country do have problems. Serious ones. They are just not the problem anybody is shouting about.
Citizenship is the wrong place to look because citizenship is an objective fact with a documentary trail. You either naturalized or you did not. There is a certificate, a file number, a date. That is why every audit that goes looking for noncitizens finds twenty people out of eight million: the category is verifiable, and the answer keeps coming back near zero.
Residency is different, and residency is where the actual mess lives.
Why Residency Is Genuinely Hard
Americans move. They move constantly, and their voter registrations do not move with them. You register in one state, you take a job in another, and nobody tells the first state you left. You move across town, across a school district line, across a municipal boundary. The old registration just sits there.
Stack that up over a decade and you get rolls carrying people who live somewhere else, duplicate registrations across state lines, records at addresses where nobody has lived in years, and a category of registrant whose file predates modern identification requirements entirely. Utah found 5,007 of that last kind — people it simply could not confirm one way or the other, most of whom registered before driver’s license or Social Security numbers were required, and nearly all of whom are certainly citizens.
Now, I want to be as disciplined about this as I have been about the other side’s numbers. Being registered in two places is not voting in two places. A stale registration is a records problem, not a fraud problem — right up until somebody casts a ballot in a jurisdiction where they no longer live. That does happen. It probably happens more than anyone has measured, and it happens most in the races where residency is the entire eligibility question: school boards, municipal offices, special districts, local referenda.
Nobody is auditing for that. There is no cabinet secretary writing letters about it. There is no prime-time address.
The Problem Everyone Is Fighting About
The Problem Actually Sitting on the Rolls
Citizenship status. Objectively verifiable against federal records.
Residency. A legal judgment about domicile that changes constantly and gets recorded almost nowhere.
Every audit finds numbers near zero.
Every serious study finds millions of stale or misplaced records.
Screening tools are wrong roughly 97.6% of the time.
Screening tools exist, work reasonably well, and are being abandoned.
Consumes executive orders, lawsuits, and floor time.
Consumes almost no political attention whatsoever.
The Other Categories — and an Honest Admission
Residency is the big one, but it is not the only one. Rolls also carry the recently deceased, until vital records catch up. They carry in-state duplicates from people who registered twice under slightly different names. In states that still disenfranchise for felony convictions, they carry mismatches in both directions. They carry plain clerical error — transposed digits, misread handwriting, the wrong box checked by a clerk in a hurry.
And there are almost certainly categories I have not identified. Anyone who tells you he has mapped this problem completely is selling something. That admission is the difference between analysis and advocacy dressed up as analysis.
The Part That Should Make You Angry
Here is the turn.
There was a tool built for exactly this problem, and it worked.
🔧 What Happened to the Thing That Worked
The Electronic Registration Information Center launched in 2012 — founded by the Pew Charitable Trusts and seven states, bipartisan from the start. Member states pooled voter and motor vehicle data to generate exactly the reports the real problem requires: cross-state movers, in-state movers, in-state duplicates, and the deceased. At its peak, thirty-three states belonged.
Beginning in 2022, amid the fraud panic, nine Republican-led states walked out.
And here is the design feature that makes those departures so costly: the cross-state movers report only catches a move when both states are members. Every withdrawal degrades the data for everyone who stayed.
The Heritage Foundation — Heritage — now publishes work calling for ERIC to be rehabilitated or replaced, because the states that left have no comparable substitute. There is no substitute. The League of Women Voters studied Missouri’s exit and could not find one either.
Sit with the sequence. A functioning, bipartisan, data-driven instrument aimed squarely at the genuine defect in American voter rolls got dismantled by the same movement now demanding emergency powers to fix American voter rolls. The one competent effort became a casualty of the theater.
Meanwhile the federal government is running commercial marketing data against forty million registrations to hunt a category that keeps testing near zero — and a federal judge has already found that its centralized database caused states to strip actual citizens off the rolls.
What Competent Work Would Look Like
None of this is mysterious. The National Voter Registration Act already constrains removals for good reason: it limits cancellation to death, felony conviction where applicable, mental incapacity, a move outside the jurisdiction, or the voter’s own request, and it requires notice before anyone gets dropped for a suspected move. Those are citizen protections, and they should stay.
But protection against wrongful removal is not the same as accuracy. Serious reform points in one direction: build the data infrastructure that follows voters to their new addresses and updates the record, rather than treating every uncertainty as grounds for cancellation. Rebuild the interstate matching. Fund vital-records integration. Define a real evidentiary standard for what it takes to move somebody off a roll.
That is the whole reform. It is boring. It costs money. It generates no rallies. It would actually work.
We are not doing it. We are doing this instead — and a lawful permanent resident who cast one ballot in 2012 is paying for it with his family.
🧠 Why This Matters More Than the Numbers
Repeat a claim often enough and people stop asking for evidence. They start treating it as weather — just how things are. Nobody has to prove that noncitizens are voting in meaningful numbers, because everyone has already heard that they do.
Then the law catches up to the story. Agencies write rules for the emergency. Judges invoke the governmental interest in stopping it. And real people pay the freight for a crisis assembled out of repetition.
Naming that is not spin. It is the job.
🔨 Where the Board Went Wrong
It counted the same conduct twice. The vote made Lopez Rico removable under INA § 237(a)(6). The Board then counted that identical vote again as a significant adverse discretionary factor — and said it deserved even more weight. Every removable person is removable for something. If the ground of removability automatically becomes a heavy negative, the scale starts tilted in every case before anyone puts a thumb on it.
It severed weight from blame. The judge found this man honestly believed he could vote and that someone encouraged him to register. Those findings stand. The Board acknowledged them, called them mitigating, then declared unlawful voting significant as a category. A factor that weighs the same whether you schemed or were handed a form at a counter is not a discretionary factor. It is a penalty.
It confused two different rules. General intent is a rule about what the government must prove to establish a charge. The Board imported it into discretion, where the entire point is individualized judgment. A statute that dispenses with knowledge for liability tells you nothing about how much a particular person’s conduct should count against him when he asks for mercy.
The election cases do not support the move.Crawford, Brnovich, Purcell, and Doe v. Reed concern a state’s authority to regulate elections against constitutional challenge. None involves immigration relief. None weighs one person’s family against one ballot. The Board borrowed the language of election-law deference and spent it somewhere those cases never went.
Nobody charged him. The Board leaned on Congress having criminalized the conduct. Congress also gave prosecution to the Justice Department, which declined here — as it declines in nearly every one of these cases. The Board treated a theoretical criminal penalty as proof of seriousness while ignoring the executive branch’s own steady judgment that these cases are not worth charging.
The dicta is the whole point. The Board affirmed. It did not need to say the judge could have gone harder. That sentence does no work in Lopez Rico. It does enormous work in the next thousand cases, where the government will quote it in every brief and every judge will read it as an instruction.
✅ What This Means in Practice
Never register, never vote, until you are a citizen. A registration card is not permission. A clerk is not an authority. If anyone at a government counter offers to register a noncitizen, the answer is no — every time, no exceptions.
Screen the voting history before any application goes out. Nearly every one of these cases — Lopez Rico, Fitzpatrick, Keathley — began with an honest person answering a naturalization question honestly. Find the issue before the government does.
Resolve the criminal case first when you possibly can. After Lopez Rico, walking into immigration court with open criminal exposure means choosing between a transcript a prosecutor can read and a silence a judge will hold against you. Sequence matters more than it ever has.
Get the judge to make findings on what the person actually believed.Lopez Rico says significant. It does not say dispositive, and it does not license abandoning the individualized balance C-V-T- and Marin require.
🎯 The Bottom Line
A man cast one ballot fourteen years ago believing he was allowed to. He reported it himself. No prosecutor ever charged him. He exercised a constitutional right when a judge asked about his drinking, and that silence helped cost him his home.
Meanwhile the government announced 15,903 noncitizen registrations in a state where only 138 registrants have any documentation gap at all. It sourced a quarter-million accusations to commercial marketing data. And the one screen a federal court ever examined was wrong 97.6 percent of the time.
Run the numbers honestly and this is a once-in-380,000 event being hunted with a tool that accuses roughly forty-one citizens for every one it gets right. No actuary would insure it. No underwriter would price it. It is not a risk. It is a story.
And the rolls do have a real defect — residency, not citizenship. We had a working, bipartisan tool aimed at it. Nine states walked out of it during the panic, and nobody has built a replacement.
We are not fixing a crisis. We are building machinery for one that does not exist, while the actual problem sits untouched. The machinery works fine on people like Rosa.
Where does this go? The Board designated Lopez Rico as precedent, so it binds immigration judges nationwide starting now. No federal appellate court has yet reviewed a discretionary denial resting on this framing. Someone will bring that petition. When they do, Keathley and Kimani and Fitzpatrick will still be on the books — and the question will be whether a court that already understands how these cases actually happen will accept that one confused ballot is a significant strike against a human life.
I think that argument is winnable. I think somebody should make it.
July 23, 2026 —Matter of Lopez Rico, 29 I&N Dec. 780 (BIA 2026), Interim Decision #4223, designated as precedent. Watch for the first petition for review testing the significant-adverse-factor framing, and for government briefing citing the Board’s greater-weight language. Check back for updates.
Disclaimer. This post is legal analysis and commentary, not legal advice, and reading it creates no attorney-client relationship. The immigration consequences of voting are severe and depend entirely on the facts. If you are not a U.S. citizen and you registered to vote or cast a ballot, do not file anything and do not speak with any government agency before consulting a licensed immigration attorney. “Rosa” is a composite drawn from multiple matters and depicts no identifiable person. Case links to Descrybe use session-based URLs and may require independent verification by readers outside my account.
Law Offices of Michael D. Baker · 111 W. Jackson Blvd., Suite 1700 · Chicago, IL 60604
