The Turnstile

The problem

A nation is a family with a front door. This catalog has argued (№ 003) that the door should be one of the easiest in the world to walk through, and (№ 007) that the perimeter should be fully seen, fast, firm, and humane. Both arguments depend on a fact that is not currently true.

America does not know who has left.

Arrivals are recorded meticulously — biometrics, interviews, databases. Departures are inferred, mostly from airline manifests, and at land borders they are substantially not recorded at all. The Department of Homeland Security must therefore estimate the number of people still in the country by subtracting a number it does not reliably have from a number it does.

Congress noticed this in 1996 and ordered an automated entry-exit system in that year's immigration act. It then re-ordered it, in one form or another, roughly nine more times across three decades, through both parties' majorities, usually without appropriating what the build would cost. Biometric exit is now substantially deployed at airports. At land crossings — the majority of American border traffic — it remains incomplete.

The consequences show up in the numbers. In fiscal year 2024, DHS recorded 538,548 overstay events against 46.6 million expected departures, a rate of 1.15 percent; after accounting for later departures and status changes, roughly 427,204 people were classed as suspected in-country overstays. The rate is not uniform: 0.43 percent for Visa Waiver Program countries, 2.22 percent for non-waiver countries excluding Canada and Mexico. And over the longer run, research by the Center for Migration Studies found that visa overstays exceeded unauthorized border crossings every year from 2007 through 2017 — with overstays accounting for about 62 percent of newly undocumented arrivals in the most recent year of that series.

So the American immigration argument has spent a generation and a great deal of money on the portion of the flow that climbs the fence, while the larger portion walks through the door, legally, and simply never registers an exit. Not because they evaded a system. Because the system has no turnstile on the way out.

The proposal

Finish biometric entry-exit on a funded deadline, publish the departure ledger as a public statistic, and replace blunt enforcement with a refundable departure bond.

This is the least ideological proposal in the catalog. It does not decide how many people should come, or from where, or on what terms. It builds the instrument that any of those decisions requires. You cannot run a generous immigration policy on unreliable data, because the political coalition for generosity collapses the moment nobody can answer "and then what happened?" An honest ledger is the precondition for an open door.

How it would work

1. Finish the build, with the money attached. A hard statutory deadline for biometric exit at land ports, with appropriations in the same bill. The lesson of thirty years is that a mandate without funding is a press release. Air and sea are largely solved; land is the expensive remainder and should be scoped honestly rather than promised cheaply.

1. Publish the ledger. Monthly overstay data by visa class and country of nationality, released on a fixed schedule with documented methodology — the way the Bureau of Labor Statistics releases unemployment. A statistic that arrives on a predictable date, prepared by career statisticians, is much harder to weaponize than one that surfaces in a press conference.

1. Consequences at the class level, never the person level. Where a visa class or country shows a persistently high overstay rate, the automatic response adjusts the terms — shorter initial validity, a bond requirement, more frequent renewal — rather than imposing bans on nationalities. The adjustment is formulaic, published, and reversible when the rate falls. A country that improves gets its terms back.

1. The refundable departure bond. For higher-risk classes, a bond posted at entry and returned in full, automatically, on recorded departure — payable by a sponsor, employer, or family member, on a sliding scale by means. This converts exit from an act with no incentive into an act with a cash reward, and it makes the traveler an active participant in producing the data. It also gives the government a humane instrument in place of detention.

1. A voluntary-departure window. Someone who has overstayed but departs within a defined window leaves with a clean recorded exit and no bar to future application. The goal of a turnstile is an accurate count and a voluntary exit — not a maximized removal statistic. A system that punishes self-reporting will not be self-reported to.

1. Statutory walls on the data. Biometric exit records are limited to non-citizens at ports of entry; U.S. citizen images are deleted on a short fixed clock; the data may not be repurposed for domestic law enforcement, commercial licensing, or any use not named in the authorizing statute. Written as a prohibition with a private right of action, not as an agency policy that a future administration can revise on a Friday.

The numbers

The scale of what is currently unmeasured is the argument. Roughly 427,000 suspected in-country overstays in a single year is comparable to the entire annual employment-based immigration allocation. Over a decade, the accumulated uncertainty is in the millions — and it is uncertainty, not a count, which is why estimates of the unauthorized population vary by margins wide enough to sustain two incompatible national narratives.

The overstay rate differential is the useful policy signal: 0.43 percent for waiver countries against 2.22 percent for non-waiver countries excluding Canada and Mexico. That five-fold gap is precisely the kind of finding that should drive graduated terms rather than categorical exclusion. Most of the world's travelers, from most countries, leave when they said they would. Policy that treats a 2.22 percent rate as if it were a 50 percent rate is not firmness; it is innumeracy, and it costs America enormous amounts of tourism, tuition, and goodwill.

On cost: land biometric exit has been scoped by CBP in the billions, and past estimates have proven optimistic. We are not going to invent a cheaper figure. It is a large, unglamorous infrastructure expenditure whose entire return is informational — and it is still the highest-leverage dollar in immigration enforcement, because every other policy in the field is currently being set without it.

The honest objections

"This is a surveillance system, and the safeguards will erode." The strongest objection, and history supports it — biometric programs built for one purpose have repeatedly been extended to others. We do not think the mitigations fully answer it. What we would say is that the alternative is not privacy; it is an enforcement apparatus that, lacking data, acts on suspicion, proxies, and nationality — which is worse for civil liberties and worse for the people it touches. The safeguards in item 6 are the price of our support, and if they are stripped in conference, the proposal should be opposed by its own authors.

"Thirty years and ten mandates suggests this is technically or politically impossible." It suggests it is expensive and that nobody has wanted to pay. Airports demonstrate that the core technology works at scale. Land crossings are genuinely harder — pedestrians, vehicles, enormous volumes, minimal dwell time — and any honest proposal budgets for a phased build with published milestones rather than a completion date chosen for a press cycle.

"Departure bonds ration travel by wealth." They do, partially, and that is a real cost we accept rather than deny. A bond that a wealthy traveler ignores is a barrier to a poor one. The sliding scale, sponsor payment, and full refundability limit it; they do not eliminate it. The comparison we would ask people to make is against the instruments a bond replaces — detention, categorical bans, and consular refusal, all of which ration by wealth too, and none of which give the money back.

"Better data will simply build a better deportation machine." It could. Data is dual-use and we will not pretend otherwise. This is why the design deliberately pairs the ledger with a voluntary-departure window and a bond that pays people to leave on time — the enforcement value we are seeking is at the door, prospectively, not in interior raids. An organization that argues for the open door of № 003 has an obligation to say where it would draw this line, and we draw it here: count everyone, publish everything, and let the consequences fall on the terms of future visas rather than on the lives of people already inside.

"Focusing on overstays lets border enforcement off the hook." It is not a substitute for anything in № 007. It is a correction of proportion. When roughly three in five newly undocumented people arrived on a valid visa, an enforcement debate conducted almost entirely about the other two is not a debate about the problem.

Sources

• U.S. Department of Homeland Security / CBP, Entry/Exit Overstay Report, Fiscal Year 2024 — 538,548 overstay events, 1.15% overstay rate, 427,204 suspected in-country overstays, VWP 0.43% vs. non-VWP 2.22% (dhs.gov)
• Illegal Immigration Reform and Immigrant Responsibility Act of 1996, § 110 (automated entry-exit mandate) and successor statutory mandates
• Center for Migration Studies of New York, visa overstays exceeding unauthorized border crossings 2007–2017; 62% of newly undocumented arrivals (cmsny.org)
• Government Accountability Office, reports on biometric exit implementation at air, sea, and land ports (gao.gov)
• National Foundation for American Policy, analysis of the DHS overstay reports (2025) (nfap.com)
• Proposals № 003 (The Open Door Compact), № 007 (The Humane Perimeter), № 012 (The Resident's Passport)