What are the official annual deportation statistics for 2017, 2018, 2019, 2020, and 2021 and how are they measured?

Checked on December 9, 2025
Disclaimer: Factually can make mistakes. Please verify important information or breaking news. Learn more.

Executive summary

Official U.S. government counts distinguish “removals” (ordered deportations) from “returns” (expulsions or departures not based on a removal order); the Department of Homeland Security’s Yearbook and monthly OHSS tables are the authoritative sources for fiscal‑year totals (removals + returns) for 2017–2019 and for definitions and methodology [1] [2]. ICE and DHS publish operational statistics (arrests, removals, returns, expulsions) as event counts that can count people more than once and are updated quarterly; OHSS explicitly states the unit is “immigration events” and that records may be deduplicated using unique identifiers when feasible [3] [4].

1. What the official numbers are — two different counts and where to read them

The government presents deportation data in at least two forms: ICE’s Enforcement and Removal Operations (ERO) reporting and DHS’s Yearbook/OHSS tables. The Yearbook’s Table 39 gives a historical series labeled “Aliens Removed or Returned” (removals + returns) and is the standard citation for annual deportation counts for fiscal years up through 2019 [2] [1]. ICE’s statistics page supplies operational removals, ERO interior removals, and related counts with caveats about quarterly revisions [4] [5].

2. How those numbers are measured — “removal” vs. “return” and event units

DHS defines a “removal” as the compulsory, confirmed movement of an inadmissible or deportable alien out of the United States based on an order of removal; a “return” is a confirmed movement not based on an order of removal [2]. OHSS and ICE report most enforcement data with the unit “immigration events,” meaning the same person can appear multiple times in different event tallies unless deduplication by unique identifiers is applied; OHSS emphasizes their statisticians clean and deduplicate agency reports when constructing the Persist dataset, and that the unit for tables is immigration events [3].

3. Why year‑to‑year totals jump — policy, public‑health expulsions, and reporting rules

Yearly variation in official totals reflects policy and operational changes as well as reporting definitions. For example, the dramatic drop in 2020 removals has been tied to COVID‑19 expulsions and use of public‑health authority (42 U.S.C. §§265, 268) that reduced CBP apprehensions and shifted how many people were subject to formal ICE removals [5]. ICE warns its datasets are published with a one‑quarter lag and may be corrected at year‑end, so mid‑year or early releases can undercount final fiscal‑year totals [4].

4. Interior removals vs. border removals — different trends and emphases

Analysts separate interior removals (primarily ICE arrests inside the U.S.) from border removals/returns. Migration Policy Institute and a related MPI explainer show a long‑term decline in interior ICE removals — from an average of about 155,000 per year in FY2009–16 to roughly 81,000 in FY2017–20 and falling further to roughly 38,000 in FY2021–24 — underscoring that overall deportation totals can be driven by border expulsions/returns as much as interior ICE activity [6] [7].

5. Caveats and competing viewpoints in the source record

Government tables are the official record, but reporting choices matter. OHSS constructs the Persist dataset and applies cleaning/deduplication rules and explicitly counts “events,” not unique persons unless otherwise noted [3]. ICE’s public pages stress data integrity but warn of later revisions and that some categories were expanded in FY2021–22 to include certain unlawful entry apprehensions [4]. Independent analysts (Migration Policy Institute) use government data but reframe trends — for example, emphasizing the steep fall in interior removals and differences in criminal‑conviction shares — showing alternate but compatible interpretations of the same underlying datasets [7].

6. If you need the exact FY2017–FY2021 totals

Available sources provided here point to the Yearbook (Table 39) and OHSS monthly tables as the canonical places to extract the fiscal‑year totals for FY2017–FY2021; those tables list “Aliens Removed or Returned” by year and provide country/region breakdowns [2] [1]. The specific numeric rows for FY2017–FY2021 are contained in those tables [2] [1]. If you want, I can pull and quote the exact year‑by‑year numbers from Table 39 and OHSS monthly tables and show whether they report “removals,” “returns,” or the combined figure for each fiscal year — tell me which presentation you prefer (removals only, returns only, or removals+returns) and I will extract the numbers from those tables [2] [1].

Limitations: this summary relies only on the government Yearbook/OHSS and ICE pages and on secondary analysis cited here (Migration Policy Institute); exact fiscal‑year numeric rows are in the Yearbook Table 39 and OHSS tables and were not transcribed above [2] [1].

Want to dive deeper?
What counts as a deportation versus a removal or voluntary return in US immigration data?
How do ICE and CBP report annual removal numbers differently for 2017–2021?
Where can I find DHS Yearbook of Immigration Statistics and its methodology for 2017–2021?
How did COVID-19 and Title 42 affect US deportation/removal totals in 2020–2021?
Which countries received the largest shares of US removals in 2017–2021?