How many U.S. citizens have been wrongfully removed by ICE and CBP since 2000, according to GAO and TRAC?

1. What the GAO actually counted and what that number means

The GAO’s published work examined ICE and CBP administrative records and concluded that, based on the datasets it could analyze, ICE removed 70 potential U.S. citizens in the period it reviewed (2015 through the second quarter of 2020); the GAO emphasizes that this count reflects “available data” and that ICE does not systematically track citizenship investigations, so the 70 figure is a documented minimum rather than a comprehensive tally ^{[1] [3]}.

2. What TRAC measured and how it differs from the GAO metric

TRAC’s analysis — cited by other observers and advocacy groups — counted at least 2,840 cases between 2002 and 2017 in which ICE wrongly identified people as potentially removable U.S. citizens, a broader category that includes misidentifications and cases where individuals were entered into removal systems even if they were not ultimately deported; that count is therefore not strictly a tally of actual removals but of erroneous eligibility classifications within ICE datasets ^[2].

3. Why the two numbers are not directly comparable

The 70 removals from the GAO and the 2,840 TRAC identifications answer different empirical questions: GAO tracked documented removals of people who appeared to be U.S. citizens within a specific later time window and warned that poor recordkeeping likely understates the scope, while TRAC’s figure reflects a longer interval and counts erroneous “potentially eligible for removal” entries that may include arrests, detainers, or database misclassifications rather than completed deportations ^{[1] [2]}.

4. Limits in the government data and why the true total is uncertain

GAO’s central finding is that ICE and CBP lack consistent guidance and systematic tracking for citizenship investigations, which undermines the agencies’ ability to know “the extent to which its officers are taking enforcement actions against individuals who could be U.S. citizens”; GAO therefore frames its numeric findings as constrained by incomplete and inconsistent records, and explicitly warns that the documented counts may understate the true scale ^{[3] [1]}.

5. How advocates, reporters, and researchers have used these figures—and the implicit agendas to watch for

Advocacy groups and media have used the GAO’s 70‑deportation figure to highlight concrete instances of wrongful removal and the TRAC figure to argue the problem is much broader; each source serves different rhetorical purposes—GAO’s conservative, evidence‑based minimum versus TRAC’s wider net of misidentification—and those different framings are important when assessing policy claims or calls for reform ^{[1] [2]}. Observers should note that TRAC’s larger number is often invoked to underscore systemic risk, while GAO’s documented removals are used to show verified harms already realized; both perspectives are factually grounded in the available analyses but neither provides a definitive, single national total because of data gaps ^{[3] [2]}.

Bottom line

According to GAO’s review of available ICE records, at least 70 apparent U.S. citizens were removed by ICE between 2015 and mid‑2020 ^[1]; according to TRAC’s analysis reported in secondary coverage, at least 2,840 people were wrongly identified by ICE as potentially eligible for removal between 2002 and 2017 — a broader category that includes misidentifications short of confirmed deportation ^[2]. Both figures are useful but neither alone captures the full extent of wrongful enforcement because GAO found agency recordkeeping shortcomings that likely leave some incidents undocumented ^{[3] [1]}.