VOL. II · NO. 04 · Essay № ii Registered No. 16191068 · Salford Sanctions · Q1 OFSI revisions
T
United TradeCrest Cross-border trade intelligence · est. MMXXV
EST. 17·I·MMXXV Salford · M6 5PQ

Sanctions screening in 2026: what changed after the Q1 OFSI revisions.

The UK OFSI Q1 update tightened beneficial-ownership thresholds and added 412 entities. The practical impact on commodity traders, shipping lines and trade-finance underwriters has been larger, and stranger, than the press coverage suggested.


§I

What actually changed.

The Q1 OFSI bulletin, published on 14 January 2026, did three things. It added 412 named entities to the consolidated list. It moved the beneficial-ownership threshold for the purpose of "owned or controlled" from fifty percent to twenty-five percent on a class of designated parent entities defined in Annex A. And it tightened the language on the "control" leg of "owned or controlled" — the leg that the press has tended to ignore — by adding two new tests that are functional rather than equity-based.

None of the three changes is, in isolation, unprecedented. UBO thresholds have moved before. The control-leg language is the more substantive change, because it elevates a test that, in screening practice as of 2024 and 2025, was almost always answered with a default no. The new tests put a real question to that default.

§II

Why screening accuracy collapsed for many firms.

For a firm whose screening platform is configured against fifty-percent UBO thresholds and a default-no control test — that is, for most firms — the Q1 bulletin produced an overnight increase in screening miss rate of between 6 and 14 percentage points, depending on counterparty composition. We say this not as a forecast but as the measured difference between the firm's existing screen and a Q1-aligned screen run against the same counterparty book on the same day. The difference was measured, for the firms that brought us their books, over the first three working weeks of February 2026.

The fixes available to most firms in that window were partial. A platform-vendor patch to the threshold was straightforward — change a number, redeploy. The control-leg tests required more: a structured ingestion of corporate-structure metadata that most platforms had not historically maintained against the precision the new tests demanded. The firms that ran their books against the threshold patch alone caught most but not all of the new hits. The control-leg additions were, in the third week of February, still mostly invisible to the unaided platform.

The Q1 changes produced an overnight increase in screening miss rate of between 6 and 14 percentage points, depending on counterparty composition. — Analyst desk · measured · February MMXXVI
§III

How CrestLens handles it.

CrestLens carries the sanctions methodology as a versioned object. The Q1 bulletin became methodology release v4.3, published in the Library on 16 January, two working days after the OFSI bulletin. The release moved the threshold, added the control-leg tests, and added the 412 entities to the canonical graph with the resolved bank-chain context the entities had at the moment of designation.

The "as-of" feature in the investigations workspace is doing more work since January than it has at any point in the life of the product. A compliance officer can now ask: what did my counterparty book look like, against the Q1-aligned methodology, on the day before the bulletin? on the day of the bulletin? on the day they signed the last deal with it? The graph answers each question against the methodology release that was current on the corresponding date, with the evidence and the contributors named.

Methodology timeline
v4.2
50 % UBO · default-no control
v4.3
25 % UBO · two new control tests
v4.3a
(in test) ultimate-UBO traversal
Library
All releases public

§IV

Practical checklist.

The list below is the one we have walked through with every commodity trading, shipping and trade-finance team that has brought their book onto the desk since February. The checklist is not novel; it is the discipline of running the existing screen with new defaults. Most of the work is in the second and third items.

CheckWhere the work is
1Threshold patched to 25 % on Annex-A parentsPlatform setting · one change
2Corporate-structure metadata fit for the control-leg testsData ingestion · ongoing
3Re-screen book against Q1-aligned methodology, with version receiptsOperational · 1 – 3 weeks
4Treat new hits as findings, not exceptions; trail each to evidenceAnalyst workflow
5Document a Q1-aligned screening policy in writing, datedCompliance · regulatory
6Brief the front office on what newly screens out and whyInternal · cross-function
§V

What's next.

OFSI has signalled — informally, in a Q1 working group — that the next bulletin (provisionally Q3 2026) will extend the ultimate-UBO traversal to the bank-chain leg, which is where, on the resolved graph, the more interesting designations have always lived. We have v4.3a in test against an extended traversal of the bank graph; we will publish it in the Library at least one full release cycle before any operational use, with the calibration window and the test set named.

For commodity traders, the Q3 change — if it lands as briefed — will move the centre of gravity of the screening problem from the counterparty itself to the bank chain behind it. The CrestLens workspace is already configured for the bank-chain pivot; the briefing room conversation we are now having with most clients is whether their internal policy is also configured for it.

§

Read further.

For the entity-resolution stack behind the screen, see The hidden economy of customs data. For the calibrated lane score that uses the same canonical graph, see Lane-risk forecasting explained. The Library carries the methodology releases, the calibration plots and the operational brief on the Q1 bulletin.

Customs data & AI → Lane risk → Request a briefing