No Result on Record: The Inspection Blanks in ICE’s Own File

The rating column in ICE's own detention file has a blank problem. Of 203 facilities, 61 carry no inspection result on record — and those facilities held 9,186 people on an average day this fiscal year. This article is about what the blanks are, where they cluster, and what they do and do not prove.

First, the discipline. A blank rating in the file means exactly one thing: no inspection result appears on record in ICE's file of this date (2026-04-09). It does not mean a facility was never inspected — the file records results, not inspection history, and the Detention Ledger refuses to claim more than the column shows. We planned to title this piece “Inspected by No One” and did not, for precisely that reason. The blank is the fact. The blank is enough.

Where the blanks cluster

The 61 facilities without a result are not a random scatter. The largest block — 38 of them — are US Marshals Service intergovernmental agreement facilities: beds ICE uses under another agency's contracting umbrella, where ICE's inspection regime evidently does not attach a result to ICE's own row. The remainder spread across Bureau of Prisons facilities, state and local agreement facilities, staging sites, and one Department of Defense site. The pattern is structural: where the bed belongs to someone else's paperwork, the rating column goes quiet.

A second, smaller blank compounds the first: 21 facilities carry no inspection date at all. And a third gap is wider than either — for 73.1% of the average daily population, the file records no threat level, meaning the majority of people in the system are held without a classification the file discloses.

The inverse problem: inspected, failed, still holding

The blanks have an inverse. 5 facilities in the file carry a Fail rating — an inspection result on record, and an adverse one — while still holding 438 people on an average day. The file does not say why a Fail rating coexists with continued custody; it only establishes, in ICE's own numbers, that it does.

What the column would need to say

Read together, the two findings frame the same question from opposite sides: 61 facilities where the record is silent, and 5 where it speaks against the facility. A file that attached an inspection result to every row — including the USMS rows — and stated the consequence of a Fail would answer both. Until it does, the honest reading is the one the ledger publishes: results where they exist, blanks stated as blanks, and nothing inferred in either direction.

Every figure is computed from the Detention Ledger (FY26 year-to-date per ICE file dated 2026-04-09); keyless JSON at /detention/index.json, CC0, zero personal data. The phrasing rules — including why “no result on record” is never written as “never inspected” — are published in our data standards.