The Fine Print of ICE Detention: What the Data Reports Never Say Out Loud
ICE publishes reams of statistics on migrants in custody — but key metrics blur, mislead, or obscure the truth. We decoded the system behind the spreadsheet.
If you’ve ever tried to understand an ICE “detention statistics” report, chances are you gave up somewhere between “ICLOS” and “ATD.”
To the average reader, these documents look more like coded spreadsheets than transparency tools — dense with acronyms, footnotes, and circular definitions.
But beneath the jargon lies a story of how U.S. Immigration and Customs Enforcement (ICE) controls, measures, and justifies the detention of migrants.
Let’s break it down.
ICE’s Favorite Metrics: Population, Length, and Flow
The agency’s most prized data point is ADP — Average Daily Population. That’s the number of people held in ICE custody on any given day, calculated by counting each detainee at midnight (called “mandays”) and dividing by the number of days in a given period.
Closely tied to ADP is ALOS — Average Length of Stay. This tells you how long, on average, someone stays in detention before being released, deported, or transferred. For ICE, ALOS isn’t one-size-fits-all: different types of facilities use different math. Adult detention centers might use the “cohort method,” while Family Residential Centers apply a “snapshot method.”
Then there’s ICLOS — In Custody Length of Stay. This isn’t an average but a snapshot: how many days an individual has currently spent in detention, from book-in date to today.
These numbers may seem technical, but they guide ICE’s funding requests, justify private contracts, and inform public messaging about “capacity” and “national security.”
Welcome to the System: Book-Ins, Book-Outs, and Releases
Detainees enter the system through book-ins, and they leave through book-outs — the bureaucratic terms for intake and exit. But how someone exits ICE custody matters a great deal.
A book-out might mean someone:
Paid bond (either to a field office or via immigration court),
Was released under an Order of Supervision (must report to ICE),
Was paroled (humanitarian or public interest release),
Or was removed — deported, or rapidly expelled via “expedited removal.”
Starting in 2024, ICE added a new layer to its reporting: Final Book-Outs. Unlike “final releases,” this category includes additional statuses like transfer to U.S. Marshals or other non-ICE authorities — expanding the tally without necessarily reflecting fewer detentions.
Who Is in Custody?
ICE doesn’t just detain “criminal aliens,” despite political rhetoric. In fact, detainees are labeled by two parallel classification systems:
Threat Level (1–3): A federal metric based on criminal history. Level 1 is the most serious. Some detainees have no criminal record at all.
Classification Level (A–D): Assigned by the facility based on behavior, vulnerability, and security risk. An A might be a first-time migrant; a D might be someone ICE deems high-risk, not necessarily based on crime.
ICE also tracks specific categories:
U.S. military veterans
Parents of U.S. citizen children
U.S. citizens mistakenly detained
Nationals of TPS-designated countries (e.g., Haiti, Venezuela)
It’s a way of anticipating public criticism — and shaping the optics of who ICE is holding.
Facilities: A Patchwork of Prisons
Not all ICE facilities are alike. Some are federally owned (SPCs), while others are operated by state or private contractors (IGSAs, DIGSAs). Others are simply rented slots in BOP (federal prison) or USMS (Marshals Service) facilities.
Inspections exist, but the bar is low: facilities marked “meets standards” can still have serious deficiencies. Still, ICE uses these labels to signal compliance — not necessarily safety or dignity.
Fear Claims and Asylum: The Real Bottleneck
When a migrant expresses fear of returning home, they may undergo a Credible Fear Interview. This process, conducted by asylum officers at USCIS, determines if there’s a “significant possibility” they qualify for protection.
A positive fear finding doesn’t guarantee release. ICE retains custody unless it chooses to parole the individual or unless a judge grants bond. In 2025, over 443,000 fear interviews were completed — with roughly half found credible. But ICE only tracks people with overlapping detention periods, making the data incomplete.
Even more opaque is the Average Release Time — the number of days between a positive fear finding and actual release. Due to overlapping data silos, that number is unreliable and inconsistently reported.
Alternatives to Detention (ATD)
Not every person in removal proceedings ends up behind bars. ICE has quietly expanded its Alternatives to Detention (ATD) programs. These include:
GPS Ankle Monitors
SmartLINK: An app that uses facial recognition check-ins
Telephonic Reporting (TR) systems
ATD is used disproportionately for families and non-criminal migrants. It’s cheaper than detention, but civil liberties advocates have raised concerns about surveillance creep and unreliable technology.
The Gloss: Why All This Language Matters
ICE’s reports — while appearing transparent — function more as a filter than a window. Technical distinctions like “Final Book-Out” or “ICLOS” are less about precision and more about control: of information, of narrative, of policy direction.
By labeling someone “pending criminal charges” or “classified D,” ICE shapes not only internal protocols, but public judgment. A person’s detention length, security status, or even release pathway is more than a statistic — it’s a reflection of how the U.S. immigration system sees them: as a risk, a number, or an exception.
Understanding this language isn’t just a policy exercise. It’s the first step to holding an opaque system accountable.
If you’d like to see an actual statistics report here, read the FY2025 one here.
We need an ICE TV SHOW!