Congress caps permanent immigration by category. Employment-based green cards are limited to roughly 140,000 a year including family members, and family-sponsored preference categories have their own limits. Immediate relatives of US citizens are uncapped.
These figures come from the Immigration Act of 1990 and have not been revised since. The American population has grown by about a third in that period and the economy considerably more, but the annual allocation has not moved. A cap that stays fixed while demand grows produces a queue automatically, without anyone deciding that it should.
The categories matter as much as the totals. Employment-based visas are split across five preference tiers, from extraordinary ability down to unskilled labour, each with its own share. Family categories are similarly divided by relationship. Unused numbers in one category can sometimes roll into another, which is why the effective annual figure varies and why analysts argue about what the real cap is in any given year.
A second limit sits on top of the first. Under the per-country cap, no single country may receive more than seven percent of employment-based or family-sponsored preference visas in a year. The cap applies regardless of a country's population or how many of its nationals qualify.
It was introduced to prevent any one country dominating admissions. The effect is that an applicant born in India with identical qualifications to one born in Belgium waits many years longer, because India uses its entire allocation every year and Belgium does not come close.
For nationals of most countries the backlog is not an obstacle and waits are often under a year. For India, China, Mexico and the Philippines it is the defining feature of the system.
The employment-based backlog is projected to grow substantially through 2030 on current allocations. Beneficiaries entering the queue outnumber the green cards available by more than two to one.
The people in it are not undocumented. They have been assessed, sponsored by an employer and approved. Many hold temporary status that ties them to a specific job while they wait, which limits their ability to change employer, start a company or in some cases for a spouse to work.
Children in the queue can age out at 21 and lose their place entirely, which is a specific and well documented consequence that neither party defends when it is described plainly.
Average processing for an employer-sponsored green card runs to several years even where no country backlog applies, and the agency has carried a pending caseload in the millions across all application types. Supporters of reform argue that a system where approval and issuance are separated by a decade is not functioning as Congress intended, whatever anyone thinks about the right level of immigration.
The case against higher levels is not usually about totals in the abstract. It concerns where arrivals settle, because immigration concentrates in a relatively small number of metropolitan areas that absorb the housing and school demand.
There is also a labour market argument. An employer able to recruit internationally faces less pressure to raise wages or invest in training, and sponsorship ties a worker to a firm in a way that weakens bargaining power on both sides. Critics argue that economic analyses measuring national long run effects skip the local transition, and that the communities bearing it are rarely the ones producing the analyses.
Integration costs are the other half of the objection. English instruction, school capacity and credential recognition all require spending in the years before the fiscal contribution of a working immigrant shows up in the numbers. Critics argue that absorption capacity is a genuine constraint rather than a euphemism, and that acknowledging it openly would produce a better conversation than treating any limit as illegitimate.
Almost nobody defends the current arrangement as designed. About two thirds of permanent admissions are family based rather than employment based, and critics across the spectrum note that the balance was not chosen so much as inherited.
The per-country cap has attracted bipartisan reform bills for years. The Fairness for High Skilled Immigrants Act passed both chambers in the 116th Congress in different forms and died when they could not be reconciled. Similar bills have been reintroduced since without reaching a floor vote.
The deeper agreement is procedural. Whatever the right number is, the current one is not really a decision. It emerged from a 1990 statute, decades of unadjusted caps, administrative practice on unused visas and litigation. A figure this consequential arriving by drift rather than by vote is something both sides describe as a problem, even when they want it moved in opposite directions.
That procedural agreement rarely survives contact with a bill, because any legislation opening the numbers invites amendments on enforcement, and any bill on enforcement invites amendments on numbers. The result is that the per-country cap, which almost nobody defends on the merits, has outlasted several Congresses in which majorities said they wanted it changed.
Want the core arguments from both sides, side by side?
See both sides of the Legal Immigration Levels debate →