What share of removals under Obama, Trump, and Biden were classified as 'criminal' versus 'noncriminal' and how reliable are those classifications?
Executive summary
Across the three presidencies, removals labeled “criminal” versus “noncriminal” shifted with policy and data practices: the Obama-era enforcement emphasized recent border crossers and prioritized criminal convictions but still recorded large shares of noncriminal “returns” (85% of removals/returns in FY2016 were recent crossers) [1]; under Trump enforcement rhetoric and later practice emphasized criminality, with some reporting finding roughly two‑thirds of ICE arrestees/removals had criminal histories in the Trump era [2], though national reporting remained opaque [3]; the Biden years saw very large volumes of returns and a decline in the share of removals that were criminals by some accounts even as removals rose after Title 42 ended [4] [5]. All three snapshots are constrained by shifting definitions (removals vs returns, Title 42 expulsions) and incomplete public data, so any numeric “share” must be treated as an estimate rather than a settled fact [6] [7].
1. What the official numbers actually show: Obama’s mix skewed toward recent crossers, not just convicted criminals
The Migration Policy analysis of Obama-era enforcement documents that the administration deliberately refocused enforcement to “noncitizens with criminal records” and recent border crossers, but contemporaneous DHS figures show that in FY2016 fully 85 percent of removals and returns were of people who had recently crossed the border — a category that includes many non‑convicted migrants processed at the border rather than interior prosecutions [1]. That means headlines calling Obama the “deporter‑in‑chief” obscured an operational reality: a large share of formal removals and returns were recent entrants rather than long‑settled migrants with felony records [1].
2. Trump’s emphasis on “criminal” removals — policy broad, data patchy
Trump-era policy texts and memos signaled a broad prioritization of criminality as a removal trigger, but also expanded the list of removable categories and left key terms undefined, allowing immigration or minor offenses to be counted as “criminal” [8]. Independent trackers and advocates documented high removal totals in 2017–2020 and later, and some pro‑enforcement analyses argue that roughly two‑thirds of ICE arrestees and many removals under Trump had criminal histories [2]. Yet reporting also notes that the Trump administrations often concealed granular breakdowns of who was arrested and removed, limiting independent verification [3].
3. Biden: big returns, fewer removals classified as “criminal” in recent years, but statistical traps abound
Under Biden the post‑Title 42 surge produced large numbers of returns and removals — Migration Policy reports 775,000 removed or returned in the 12 months after Title 42 ended, and that in FY2023 more migrants were returned across the border than removed from the interior [4]. Multiple outlets observe that the share of removals involving people with criminal convictions appears to have declined in recent years even as total numbers rose, but those trends are sensitive to whether CBP “returns” and Title 42 expulsions are counted [5] [4]. Critics and analysts caution that inflating “deportation” totals by combining returns with formal removals can mislead [6].
4. How “criminal” is being defined — the core reliability problem
The reliability of the “criminal” label is undermined by inconsistent definitions across memos and agencies: “criminal” may mean an aggravated felony, any conviction, a misdemeanor, or even immigration-specific charges like illegal entry/reentry; Trump memos often left categories broad and undefined [8]. Additionally, returns and Title 42 expulsions are operationally different from formal removals with judicial orders, and are sometimes omitted from or conflated in public tallies — a key reason why independent fact‑checking warns against simple comparisons across administrations [6] [9].
5. Data gaps, political framing, and what can (and cannot) be concluded
Public reporting and advocates on all sides point to incentives to frame enforcement statistics favourably — administrations may publicize selective figures while withholding granular interior vs border breakdowns, criminal conviction types, or whether an arrest preceded removal [3] [7]. Reliable comparative statements therefore require standardized metrics (e.g., share of formal interior removals that followed felony convictions versus non‑criminal returns), which are not consistently published across these presidencies; available sources show clear directional differences but cannot produce a single, precise cross‑administration percentage without additional DHS‑provided microdata [6] [4].
6. Bottom line for readers and researchers
The best-supported conclusions are: Obama emphasized prioritized enforcement but many FY2016 removals/returns were recent border crossers (85%) rather than long‑term convicted criminals [1]; Trump’s enforcement stressed criminality and several analyses find a high share of arrestees/removals with criminal histories, but policy vagueness and data concealment complicate exact shares [2] [3] [8]; Biden presided over large numbers of returns and a declining share of removals that appear to be for criminal convictions, but that trend is susceptible to definitional and reporting artifacts especially around Title 42 and returns [4] [5] [6]. Any precise “share” comparison is therefore provisional and must be caveated by how “criminal” and “removal” are defined in the source dataset.