The KFF/New York Times 2025 Survey of Immigrants (n = 1,805, conducted August to October 2025) found that 56% of likely undocumented immigrants in the US say they would still choose to come to America if they could go back in time, down from 72% in 2023. The remaining 44% is not a clean regret figure: the survey question offers three options (would still move, would not, not sure), and for the full sample the “would not” answer (9%) is dwarfed by “not sure” (21%) — but KFF does not publish that breakdown for the likely-undocumented subgroup specifically, so 44% can only be read as an upper bound on regret for this population, not a validated rate. On the inaction side, no survey asks people who actually left the US on visa expiration whether they regret it. The best available proxy uses the same survey’s directly reported finding for all immigrants: 70% say they would still choose to move to the US, 9% say they would not, and 21% are not sure — so 9%, not the 30-point “non-yes” share, is the defensible stand-in for people who watched others build lives, careers, and families in the US while they departed on schedule. Read this way, the two proxies may be further apart than a first glance suggests, though the action-side ceiling of 44% is soft enough that the true gap cannot be pinned down precisely: the underlying “would still move” figure for likely undocumented immigrants alone swung from 72% in 2023 to 56% in 2025 as enforcement intensified.
The mechanism on each side is structurally different. Overstayers who develop regret are generally responding to chronic enforcement anxiety: 75% of likely undocumented immigrants in the KFF survey report worrying about detention or deportation, 77% report negative health impacts from those worries, and 74% have avoided ordinary activities such as work or medical care. These costs are ongoing and compounding. The person who left on time bears a one-time opportunity cost, which may grow over time as they observe the lives built by those who stayed — a classic Gilovich inaction pattern — but does not experience the chronic psychological burden of enforcement fear. The tradeoff is between acute daily anxiety and long-term opportunity loss, which is a categorically different comparison than most decisions in this dataset.
Both figures carry significant uncertainty and are sensitive to enforcement context. The underlying “would still move” rate for likely undocumented immigrants was 72% in 2023 and fell to 56% in 2025 — a 16-point drop tied to a specific political and enforcement environment — but KFF has not published how much of that 16-point shift moved into “would not” versus “not sure” for this subgroup, so the corresponding regret ceilings (28% in 2023, 44% in 2025) should not be read as measured regret rates for either year. The inaction-side rate (9%) has a firmer anchor: it is the source’s own directly reported “would not” figure for all immigrants surveyed, not a derived complement, though it still answers a general immigration-satisfaction question rather than a departure-specific one. Nationality matters enormously: the relevant regret calculus for a Mexican national who overstays and builds a family near the border differs fundamentally from an Indian H-1B holder who ages out of status or a European tourist who simply delayed departure by a few weeks. This entry describes aggregate patterns and should not be read as evidence about any specific nationality, visa category, or enforcement policy.