How do DHS and ICE define and report 'removals' versus 'returns' and 'criminal convictions' in deportation data?
Executive summary
DHS and ICE distinguish "removals" as compelled departures under immigration law from "returns" or "voluntary returns," which are administrative or non‑judicial dispositions that do not carry the same formal order consequences, and they flag a "convicted criminal" variable in their datasets that counts prior criminal convictions (with specific exclusions) rather than clustering crimes by severity in a single, consistent metric [1][2][3].
1. Definitions that matter: what DHS/ICE call a "removal" versus a "return"
A "removal" in DHS reporting is the compulsory, confirmed movement of an inadmissible or removable noncitizen out of the United States based on a final order of removal or equivalent administrative disposition; such removals often carry reentry consequences under the Immigration and Nationality Act (INA) and are recorded as removals or repatriations in OHSS and ICE data [1]. By contrast, USBP and CBP "returns" or "voluntary returns" are dispositions recorded when an individual with certain citizenships is allowed to go back without an ICE book‑in or without a formal final removal order — a classification the agency uses in monthly enforcement tables and repatriation guidance and which has included special rules for Mexican and Canadian citizens historically [1][4].
2. How the border and interior pipelines change the labels and the numbers
Border encounters are frequently processed through CBP expedited procedures and may be recorded as expedited removals, administrative deportations, reinstatements, or voluntary returns, with DHS noting that border removals are faster and simpler because they don’t require locating a person in the interior or running full adjudications [1][5]. Interior ICE ERO removals, by contrast, are more likely to follow INA §240 proceedings or interior enforcement and therefore show higher shares of cases catalogued as "removals" tied to final orders; ICE’s FY reporting and ERO statistics repeatedly emphasize that most interior removals are of people who passed through criminal‑justice contact or formal removal processes [3][6].
3. The "convicted criminal" variable: what it is—and what it leaves out
ICE and DHS data include a Yes/No variable indicating whether an individual had been convicted of a criminal offense in the United States; ICE defines "convicted criminal" as someone convicted in the U.S. for one or more criminal offenses and explicitly excludes civil traffic offenses from that flag [3][2]. Agency summaries and OIG reporting cite large numeric counts of removals or returns that “had criminal convictions,” but the datasets do not uniformly translate conviction counts into a single public‑safety severity scale: minor misdemeanors (including certain DUIs or non‑civil traffic offenses) and felonies are both counted within the convicted‑criminal bucket unless the reporting and narrative separate them [7][3][8].
4. What the headline percentages mean — and why they can be misleading
Different DHS and independent counts produce different headline shares: ICE has reported that a high share of its interior removals involve convicted criminals (ICE reported 92 percent for interior removals in certain FYs) while Migration Policy Institute analysis found 79 percent of ICE interior deportations in FY 2021–24 involved someone with a criminal conviction [6][5]. These percentages reflect differing denominators (interior vs. total, removals vs. returns vs. expedited removals), different fiscal years, and differing inclusion rules for expedited or reinstated removals — all reasons the same dataset can be read two ways [5][6].
5. Reporting limits, data projects, and the politics of framing
Public DHS/ICE data are parsed into monthly tables and downloadable files, and third‑party projects like the Deportation Data Project repackage and annotate agency releases to make caveats visible, but journalists and officials have also used the same numbers to promote opposing narratives: agency press releases touting “criminal removals” frame the data to emphasize public‑safety impacts, while watchdogs and researchers highlight the large share of non‑criminal immigration‑only removals and the role of expedited and voluntary returns to argue for restraint — both interpretations can be supported by agency data but depend on choices about which dispositions to count as "deportations" [4][9][10].