Can an 11,000-member House settle immigration?

This is a simulation. Claude played every part. The facts in section 1 are real. Everything else is made up: the members, their groups, the districts, the bills, the votes, the estimates, the attacks, the president's choices, and every voice. Made-up events are marked (simulated). This is not a prediction, a poll, or a recommendation, and it takes no side on immigration. No stakeholder speaks for a real organization, a whole region, or a whole faith.

Rerun on 29 October 2026, under the same two rules as the other reruns: stakeholders only watch (they do not vote, write bills, take sides, or act; they watch how the system works, take complaints, and report), and plain language. The bills and voting assumptions are unchanged, and so are the vote counts.

The question. Health care asked whether a large House could agree on what the problem is. Immigration asks something harder: whether members can agree on anything when the chamber is split not only by party, but between cities and farm towns, between regions that live with the border in very different ways, and among religious communities that weigh duties to strangers and to the law differently.

1. The real situation (as of 28 October 2026)

Enforcement

• Deportations. Immigration and Customs Enforcement (ICE) reported 356,389 deportations in the 2026 budget year through late July, on pace to pass the record of 438,421 set in 2013. June had the most arrests of any month: just over 43,000 by ICE and Customs and Border Protection together. These are ICE's own figures and were not independently checked. (ABC News)
• Detention. 65,765 people were held in immigration detention on 11 July; 70.6% of them had never been convicted of a crime. (TRAC, a data project at Syracuse University) ICE says 60% of the people it holds have a criminal record, a figure that counts pending charges as well as convictions. (ABC News)
• The border. Customs and Border Protection reported a 95% drop in arrests at the southwest border over the 2026 budget year, to 9,000–12,000 a month, and the lowest total border encounters nationwide in its history, 379,243. September's numbers were still preliminary. (Just the News / Center Square)
• Minnesota. During a large enforcement operation in Minneapolis and St. Paul, an ICE officer shot and killed Renee Good, a 37-year-old US citizen, on 7 January. A Border Patrol officer shot and killed Alex Pretti, a 37-year-old citizen and intensive-care nurse at a veterans' hospital, on 24 January. The operation ended on 12 February. (PBS NewsHour) It reached small towns too. In Worthington, Minnesota, where about one resident in three is an immigrant and the meatpacking plant is the largest employer, shops reported steep drops in customers. (KORN Radio)

Congress and money

• The shutdown. Democrats refused to fund parts of immigration enforcement without changes such as body cameras and limits on agents covering their faces. The resulting 76-day shutdown of the Department of Homeland Security, the longest shutdown of a single agency in US history, ended on 30 April without those changes. ICE and the Border Patrol were left out of the bill that ended it. (KPBS)
• The fast-track law. Republicans then funded ICE and Customs and Border Protection through a fast-track budget process that needs only a simple majority in the Senate. The Secure America Act passed the House 214 to 212 and the Senate 52 to 47, and was signed on 10 June. It provides $69.5 billion through September 2029: $38.5 billion for ICE, $26 billion for Customs and Border Protection, and $5 billion for the department generally, without the reporting rules that usually come with regular funding. (American Immigration Council)
• Legal status. The bipartisan Dignity Act would create a renewable seven-year legal status for people who arrived before 2021, with a $7,000 fine, back taxes, and a requirement to work or study, but with no general path to citizenship. It has a separate path for people brought to the US as children, and also requires employers to use E-Verify, the federal online system for checking whether a worker can legally work. It adds border measures and faster asylum decisions. It had 39 cosponsors, split almost evenly between the parties. (PolitiFact)

Courts

• Citizenship at birth. On 30 June, in Trump v. Barbara, the Supreme Court held that children born in the US to parents here illegally or temporarily are citizens at birth, and kept in place a court order blocking the president's executive order on the question. (Faegre Drinker)
• Houses of worship. A federal judge in Massachusetts barred arrests without a warrant at, and enforcement within 100 feet of, about 5,320 churches whose denominations sued after the government dropped its 2025 policy of mostly avoiding churches, schools, and hospitals. (Washington Times) On 20 October, a federal appeals court in Washington, DC, revived a separate lawsuit by Christian and Jewish groups over the same change, 2 to 1. (Bloomberg Law)

Population, refugees, and work

• Population. The number of immigrants living in the US fell from a record 53.3 million in January 2025 to 51.9 million in June 2025, the first drop since the 1960s. The number living here without legal permission was a record 14 million in 2023; about 6 million of them had some temporary protection from deportation. (Pew Research Center) The Congressional Budget Office lowered its forecast of immigration for 2025–2029, and projects that deaths will outnumber births from 2030. (Congressional Budget Office)
• Refugees. The limit on refugee admissions for the 2027 budget year is 17,500, with white South Africans of Dutch descent (Afrikaners) the main group considered. (Yen.com.gh)
• Farms. Workers born abroad make up about two-thirds of farm labor, and workers without legal status about 40% of hired crop workers. More than 300,000 seasonal farm worker visas were issued in 2024, and the program's wage rules are changing. (RFD-TV)

Voters

• In June, 73% of adults told Gallup that immigration is a good thing for the country, down from a record 79% in 2025. 31% wanted more immigration, 35% wanted the same level, and 29% wanted less. Among Republicans, the share calling immigration a good thing fell 14 points, to 50%. (FOX 9, on Gallup's poll) Midterm elections are on 3 November.

What a House can do. It can pass laws on enforcement, legal status, visas, refugees, asylum, and funding. It cannot change who is a citizen at birth, which the Constitution decides, and it cannot dictate how the president uses the leeway the law gives him. Every law needs the Senate, including 60 votes to end debate, and the president's signature or a two-thirds override. States and cities decide how far to help federal agents, so the same federal law lands very differently from place to place.

2. Who took part

The simulated House: the same 11,000 members, turnout model, and rules as Simulation 4, with two changes made after it: the runner-up rule (if the winner of a ranked vote falls short, it is dropped and the count runs again) and alternative versions (a bill can be marked as an alternative to another, and both are ranked together). The members fell into new groups, defined by views on immigration, not by party.

721 members miss each vote. The group sizes and views are assumptions, published in advance.

Districts. For the first time, each simulated district was given a region and a type of place, so results could be read by geography. These are assumptions, roughly in proportion to population, not facts about real districts.

A caveat. Real immigration votes follow party lines more closely than these groups do, and real districts are not this tidy. The labels let the run ask where support comes from; they do not say where it would come from in Congress.

Other simulated roles. These act in the simulation; none is a stakeholder:

• Local and state officials: city and county officials, state officials, farm and business employers, and tribal government officials.
• Congregations and their leaders of many faiths, and organizations that run shelters and food banks.
• People affected, posting as advisers with their backgrounds disclosed: a teacher brought to the US as a child, who has temporary protection from deportation; a dairy worker on a seasonal visa that does not fit year-round milking; a naturalized citizen whose husband is in detention; a rancher whose border land is crossed; and a refugee from Afghanistan resettled in a Midwest city.
• Also: the budget office, fact-checkers, video creators, translators, a legal reviewer, constituents, attackers, the Senate, the president, and a forecaster.

New stakeholders watching. Nineteen joined. All are made up. None votes, writes bills, or takes a side on any bill, and none matches a voting group. Each watched how the system worked for the people, places, or communities they know.

Cities, suburbs, and rural areas

Regions

Religious traditions, old and new

The council asked for traditions that arrived in, or began in, America at very different times: Indigenous traditions far older than the country; Catholic, Protestant, and Jewish communities present since colonial times; the Latter-day Saints and Pentecostalism, both born in the US; faiths brought by recent immigrants; and the growing number of people with no religion. Each speaks only for its own made-up organization, and only about how the system serves its community.

The people affected, and who could not join. The Ombud and Privy stakeholders warned that no one living in the US without legal status could safely join the panel, because a signed comment would be a permanent, public statement of their status. So the people most affected by the legal-status bills had no voice of their own.

3. Days 1–3: Seven bills

Seven bills were posted:

The budget office's estimates (simulated; the assumptions are the office's own):

• I1: tens of billions of dollars over ten years on top of the Secure America Act, mostly for detention. The office could not estimate the grant penalties without knowing which places would comply.
• I2: small direct costs. Its larger effects come through the workforce, which a federal cost estimate usually does not count.
• I3: brings in money overall, from fines, back taxes, and the extra taxes of people working legally, minus border and E-Verify costs.
• I4: about even; visa fees cover the cost of running it.
• I5 and I6: a few billion dollars each.
• I7: $8–12 billion over five years for paying back local costs, which then expires; if made permanent, $17–25 billion over ten years.

A second estimate. Members of the Lower numbers group published their own estimate of I3. It assumed most eligible people already pay payroll taxes, and that use of public benefits rises once the seven-year term ends, and put the overall effect near zero. Both estimates appeared side by side; neither is official.

Fact-checks by the simulated fact-checkers:

• A member said "70% of people in ICE detention have no criminal record." The 70.6% figure counts people never convicted; the Syracuse data separates convictions from pending charges, and ICE's own 60% figure counts both. Rated misleading as worded; the same claim with "conviction" would be accurate.
• A member called I3 "mass amnesty." I3 grants a renewable status, not a pardon, with fines and no general path to citizenship. Rated misleading. The fact-checkers cited PolitiFact's real rating of the same claim about the real Dignity Act: half true.
• A simulated video creator said border crossings "fell 95% because of the new ICE money." The 95% figure covers the whole budget year; the Secure America Act was signed in June. Rated unsupported.

What outside groups told members during the wait:

• City and county officials asked members to pay back local shelter and jail costs. Several counties that had cooperated with federal agents for years objected to I1's grant penalties, which could hit them over a single practice they disagreed with.
• Farm employers and rural county officials backed I4. One county official said half the children in his schools have a parent at the plant.
• Tribal government officials and border county officials objected to I3's money for border construction, which required no consultation with tribes or border towns.
• Congregations of many faiths asked members to limit enforcement at houses of worship. Secular shelters and food banks asked to be covered too.
• The majority party's leaders posted a party position backing I1. Members of the small party began meeting every group.
• Translators published plain-language versions and versions in Spanish, Chinese, Vietnamese, Tagalog, Arabic, Haitian Creole, and Punjabi, each tied to exact quotes of the bills, with a glossary for status, parole, detainer, and removal.

An attack (simulated). Attackers used 300 accounts, registered long enough ago to pass the waiting rule, to post comments backing I1. Each comment named a Catholic diocese or Jewish federation in the simulation as the poster's organization, without its permission. The comments disclosed everything required, falsely, and nothing in the system could tell. The dioceses and federations whose names were used posted denials. The attack was revealed after the vote.

4. Days 4–7: The ranked vote, a short voting period, and the runner-up rule

The seven bills went into one ranked vote. The voting period ran from Friday afternoon to Saturday night, Eastern time, so the result would be known before a committee meeting on Sunday (simulated).

I3 won the ranking. It beat every other bill head to head, most narrowly I6, 4,881 to 4,534.

It then needed 5,501 members to rank it above "no change," and fell short:

I3 fell 189 votes short: 57.1% of yes-and-no votes, but 48.3% of all members.

• Why: Enforcement first and Lower numbers members opposed any legal status. Many Humanitarian and Accountability members opposed a status with no general path to citizenship, and housing asylum seekers near the border.
• How solid: no test with one group 10% larger or smaller made I3 pass.

The runner-up rule ran for the first time. I3 was dropped and the count ran again. I6, the accountability bill, won every remaining head-to-head, most narrowly against I4, 4,903 to 4,611, and got enough members to rank it above "no change":

I6 passed by 65 votes: 59.5% of yes-and-no votes and 50.6% of all members.

• Close call: it would have failed with an Enforcement first group 10% larger.
• The runner-up rule chose the bill fewer members preferred. 4,881 members ranked I3 above I6, and 4,534 ranked I6 above I3. The rule passed I6 because more members ranked it above "no change." The bar stayed the same. But nothing on the results page said that the dropped bill had been preferred to the winner.

How each bill did against "no change" on its own:

Where I3's support came from. The district labels allowed the first reading of a vote by geography:

The gap between cities and rural areas, 30 points, was more than twice the gap between regions, 12 points. The suburbs, with nearly half the districts, split exactly in half.

A complaint (simulated). 268 members who keep a Saturday Sabbath could not vote at all during a period that ran from Friday afternoon to Saturday night. They complained to Ombud and Covenant. Members who worship on Sunday told Pulpit that a Saturday-to-Sunday period would have done the same to them.

• Had those 268 members voted like the rest of their groups, I3 would have gained about 153 yes votes, still 36 short, and I6 about 156. The outcome stands.
• Ombud's finding: the count was correct under the rules. But in a closer vote, the choice of voting period would have decided the result, and 268 members are recorded as absent from a vote they wanted to cast.

5. Days 8–14: A coalition bill, and an alternative version

Eighty-four members from every group except Lower numbers formed a coalition and started a shared draft. Their plan: start from I3, which beat every bill head to head; fix what cost it votes; and settle the citizenship question with an alternative version, instead of fighting over it inside one bill.

Bill 8a, Earned Status and Lawful Workforce Act (simulated, in plain words):

• Title I, Earned status. A renewable seven-year legal status for people living here continuously since 1 January 2021, with a $7,000 fine, back taxes, background checks, and a requirement to work or study. People brought here as children may apply for permanent residence after ten years.
• Title II, Workplaces. E-Verify phased in by size of employer: 500 or more workers from 1 July 2027, 50 or more from 1 July 2028, and all employers from 1 July 2029. Year-round farm visas, including for dairy and livestock, with wages tied to local rates.
• Title III, Border and asylum. More officers and inspection technology at official border crossings; 150 more immigration judges; asylum decisions within 180 days, with no requirement to house applicants near the border.
• Title IV, Local choice. No federal grant penalties for how much a state or city helps federal agents. The federal government pays back local shelter and jail costs, per person, for five years.
• Title V, Places of worship and service. Enforcement at houses of worship, schools, hospitals, shelters, and food banks needs a supervisor's approval, except in an emergency, whether the place is run by a religious or a secular group.
• Title VI, Tribal consultation. Tribal governments must be consulted before any construction on or near tribal land or sacred sites.
• Title VII, Paying for it. Fines collected under Title I are set aside to pay for Titles II to IV.

Bill 8b is marked as an alternative version of 8a. It changes one paragraph in Title I: anyone who holds the status for ten years may apply for permanent residence, and so, eventually, for citizenship.

What members changed, and why (all writing was done by members):

• Members from Southwest districts moved border money from barriers to official crossings, after border county officials argued that most trade, and most seized drugs, pass through them. They added Title VI after hearing from tribal government officials.
• Members from Southeast and rural districts phased in E-Verify by employer size and added year-round farm visas, after hearing from employers and farmers.
• Local choice members from states on both sides of the cooperation question wrote Title IV's ban on penalties for either choice. Members from cities added the paybacks.
• Humanitarian and Security and status members wrote Title V, after hearing from congregations, and covered secular shelters and food banks too.
• The simulated legal reviewer found that "emergency" in Title V was not defined. The members added a definition.

The budget office's estimate for 8a (simulated):

Who bears the costs the estimate leaves out? The estimate format had no place for local costs or effects on particular regions. The office's own rough figures put most of the paybacks in Northeast and Pacific cities and Southwest border counties, and most of E-Verify's effect on workers in the Southeast, the Pacific, and the rural Midwest. Those figures sat in the estimate's written notes, where no one could compare them.

6. Days 14–18: The original, the alternative, and the runner-up rule again

The alternative needed 24 hours' wait. After the complaint, the coalition set a 96-hour voting period, Wednesday to Sunday.

• Head to head, 8b beat 8a, 4,712 to 4,388. Most members at the enforcement end ranked neither, which counts as a tie.
• Against "no change," 8b fell short:

8b fell 82 votes short. It would have passed with an Enforcement first group 10% smaller.

• The runner-up rule dropped 8b, and 8a passed:

Bill 8a passed with 6,577 yes votes: 59.8% of all members, 68.2% of yes-and-no votes, and 1,265 more than I3. It passed in every test with one group 10% larger or smaller (6,488 to 6,666 yes votes).

• What moved votes: the border-crossing and tribal-consultation changes brought Southwest members; the phase-in brought Southeast members; Title V brought members from districts with large immigrant congregations; the paybacks brought city members; and 402 Enforcement first members, mostly rural, voted yes, up from 228 for I3, pointing to the workforce title.
• Twice in one session, the runner-up rule passed a bill that fewer members preferred to the one it dropped. Both times, the bill that passed had more support against "no change."

7. The Senate and the president (simulated)

• The Senate on I6: 51 to 49. Short of 60, so the bill died, as the real changes did when the shutdown ended.
• The Senate on 8a: 56 to 44. Short of 60.

Offers back and forth:

1. The Senate offers 8c: 8a with E-Verify for all employers from day one and without Title V. House: 4,871 yes. It fails.
2. The House offers 8a with more money for border crossings. Senate: 57 to 43.
3. Senators propose splitting the bill: Bill W, Workforce and Border Crossings, with only Titles II, III, and VI. No legal status, no paybacks, and no limits on enforcement at houses of worship.

• House: 7,435 yes, the most of any immigration vote in the run. Senate: 71 to 29.
• The geography flipped. Bill W won 77.8% of rural districts, 67.0% of suburban, and 58.8% of city districts, the reverse of I3. The legal-status bills were carried by cities; the workforce bill was carried by farm towns.
• The president: the simulated president signed, since border crossings, E-Verify, and seasonal visas fit the administration's direction. The least certain part was the year-round visas. Labeled as made up.
• Forecast: the simulated forecaster put the chance that any earned legal status becomes law in this Congress at 1 in 15.

8. Did the House settle immigration?

No. It settled the parts that do not divide cities from farm towns.

• No group's own bill had a majority. Enforcement, lower numbers, refuge, and local choice each fell short against "no change," by 880 to 3,200 votes. Two questions divided the chamber: how much enforcement, and whether long-time residents get legal status. Whether a district was a city or a farm town predicted members' answers to the second question far better than region did.
• Legal status can pass a large House, and stops at the Senate's 60-vote barrier. A status bill with a phased workplace title, border crossings instead of barriers, tribal consultation, and limits on enforcement at places of worship and service passed with 6,577 yes votes and died at 57 senators.
• Workforce and border crossings pass everywhere. The bill that became law (simulated) was the one rural members wanted most.
• Rules for how enforcement is done passed the House narrowly and died in the Senate, as the real changes did in the shutdown fight.
• The record made everything visible: the voting period that shut out 268 members, the version each side ranked, the runner-up rule that twice passed the bill fewer members preferred, every offer between the chambers, and the geographic flip between the status and workforce votes.

What members who voted no said:

• Enforcement first: legal status before proven enforcement repeats 1986, when legal status came and the promised enforcement did not.
• Lower numbers: no bill asked how many people the country should admit each year, which is the question that matters.
• Humanitarian: a status with no path to citizenship creates a permanent class of residents who cannot vote, and refugee admissions were not restored.
• Accountability: two citizens died in Minnesota, and the run ends with no new rule on how agents identify themselves.

What people affected said:

• The teacher brought here as a child: "My path to residence passed the House and died in the Senate. Again."
• The dairy worker: "Year-round visas help, if they arrive before my season ends."

9. What the stakeholders saw

What worked:

• County and Piedmont: E-Verify's three start dates, by employer size, could each be set on its own part of the bill, and the record showed each one.
• Sangat: a Punjabi version was published, tied word for word to the bill.
• Prairie: a member on a slow rural connection could download and check the whole record in time. The voting period, not the internet, was the limit.
• Acre: the plain-language versions worked on a phone, though start dates were buried deep in the detailed comparison of changes.

What failed:

• Covenant, Pulpit, and Mission: a short voting period over a day of rest shut out observant members. Any fix that lists holy days would force the project to decide whose holy days count.
• Metro and Harbor: the estimates had no place for local costs or effects on particular regions.
• Frontera and Pacific: I7 and Title IV work differently depending on each state's choice. The record could show only one national text, not the law in force in a given state.
• Ring: the city, suburb, and rural breakdown came from labels the simulation added. Nothing defined those labels or limited what labels a simulation could add.
• Parish: anyone could claim to speak for a church, and readers had no way to check.
• Agenda and Spatial: the runner-up rule twice passed a bill that fewer members preferred to the one it dropped, and nothing told members so.
• Cedar: a tribal government could only comment as an ordinary adviser, like anyone else.
• Iglesia, Privy, and Ombud: no one living here without legal status could safely take part.
• Range: there was no way to include the territories or the District of Columbia, whose residents live under federal immigration law with no voting member of the House.

Complaints the stakeholders received:

• To Ombud and Covenant, from 268 Sabbath-observing members: they could not vote at all during the first ranked vote.
• To Pulpit, from Sunday-worshipping members: a Saturday-to-Sunday period would have shut them out the same way.
• To Parish and Covenant, from the simulated dioceses and federations: attackers posted 300 comments in their name.
• To Metro, from simulated city officials: "The estimate says the bill saves money. It doesn't say the shelter costs land on us."
• To Frontera and Pacific, from simulated state officials: "We can't tell from the record what the law will be in our state."
• To Cedar, from simulated tribal government officials: "We are a government. The system treats us like a commenter."
• To Iglesia, from congregants without legal status: they wanted to speak, and could not do so safely.
• To Minbar and Humanist, from members: a request to add religion to the district labels for research. Every religious stakeholder opposed it.
• To Agenda, from Humanitarian members: "More of us preferred the citizenship version. Why did the other one win?"
• To Range, from simulated territory residents: "Federal immigration law applies to us too."

10. What changed in the system afterward

The council turned these findings into changes. The technical details are in the design log, Session 13.

• Geography only. A simulation may label each district with a region and a type of place, from lists it declares, and results can be broken down by either. No other label is allowed: not religion, ethnicity, birthplace, party, income, or anything else.
• Where a law applies. A part of a bill can be limited to certain states, or apply only where a state opts in. The record can show the law in force in any one state.
• Voting periods of at least 72 hours. Every voting period lasts at least 72 hours, so no single day of rest can shut anyone out. The system keeps no list of holy days and names no faith.
• Who pays, and where. Cost estimates can show effects by region and type of place, and costs that fall on states, counties, or cities.
• Checking who speaks for an organization. An organization can vouch for the accounts that speak in its name. Comments show whether a claimed organization has vouched for the account. Vouching is never required, and the system never decides who truly speaks for a group. The same rule applies to churches, unions, companies, and advocacy groups.
• Telling members when the runner-up rule changes the result. When the rule passes a bill other than the first winner, the results show how many members preferred each dropped bill to the one that passed.
• Still open: how people at risk can take part without being exposed; how tribal nations can take part as governments; how to represent the territories and the District of Columbia; and whether simulated residents of a district should include people who are not citizens.