Brief
One cap table, several answers
Why sanctions ownership screening needs its own arithmetic
Bottom line
Legal-entity onboarding usually collects ownership once, for due diligence: the individuals who own 25 percent or more, plus one person who controls the company. Sanctions screening then checks those names against the lists and calls the result ownership screening. It is a name check on a partial record. The sanctions ownership test asks a different question, adds stakes with different arithmetic, and changes its answer on a different clock. An entity can be blocked by operation of law while appearing on no list, and a record built for due diligence is structurally unable to see it. Ownership screening needs its own data, its own computation, and its own trigger.
1. Two questions share one cap table
The due diligence question comes from FinCEN’s beneficial-ownership rule: who are the natural persons behind this customer? The rule reaches each individual who owns 25 percent or more of the equity interests, directly or indirectly, and one individual with significant responsibility to control, manage, or direct the company. Since February 2026, a FinCEN exceptive relief order has let institutions ask once, at the first account, and again only when facts call the answer into question or risk-based procedures call for it.
The sanctions question comes from the blocking programs. Under OFAC’s 50 percent rule, an entity owned 50 percent or more by blocked persons, directly or indirectly, individually or in the aggregate, is itself treated as blocked, listed or unlisted.
| Due diligence record | Sanctions ownership test (OFAC) | |
|---|---|---|
| Who counts | Natural persons at 25 percent or more, plus one control person | Every blocked holder, person or entity, at any size |
| How stakes add up | Multiplied down the chain | Through blocked entities at full value, through unblocked entities at zero |
| When the answer changes | At first account opening, then on doubt or under risk-based procedures | The day any holder anywhere in the chain is designated |
2. Who counts
Aggregation is what makes the 25 percent record insufficient: two blocked holders at 25 percent each block an entity under OFAC’s rule, and so do three at 17 percent. A record that stops at 25 percent has discarded exactly the stakes aggregation needs, and a record of natural persons has discarded the holding companies that indirect ownership runs through. The threshold answers its own question well. The error is borrowing it for another.
3. How the stakes add up
FinCEN’s 2018 FAQs illustrate indirect ownership by multiplying stakes down the chain. On that method, a person holding 60 percent of a company that owns 40 percent of the customer owns 24 percent of it.
OFAC counts differently. Indirect ownership, in its usage, runs only through entities that are themselves 50 percent or more owned by blocked persons, and such an entity’s stake counts in full. OFAC’s own example: a blocked person owns 50 percent of Entity A and 10 percent of Entity B, and Entity A owns 40 percent of Entity B. Entity B is blocked, at 50 percent. Multiplied down the chain, the same structure gives 30.
The divergence runs both ways. Through an unblocked intermediate OFAC counts nothing, where proportional arithmetic credits a share. Proportional look-through therefore understates blocked ownership through blocked intermediates and overstates it through unblocked ones, and a different threshold leaves both errors in place, because the error sits in the method. One person, several spellings describes the same trap in name matching.
4. Which regime is asking
A firm exposed to more than one program runs more than one test on the same chart.
| United States (OFAC) | European Union | United Kingdom | |
|---|---|---|---|
| Where the test lives | OFAC’s interpretation of the blocking prohibition, in guidance and in many program regulations | Council Best Practices, nonbinding, reading “owned or controlled” in the regulations | Legislation, such as regulation 7 and Schedule 1 of the Russia regulations |
| Ownership threshold | 50 percent or more | 50 percent or more, since July 2024 | More than 50 percent of shares or voting rights |
| Different designated persons’ stakes | Aggregated | Aggregated | Generally kept separate; combined where the evidence shows a joint arrangement or control |
| Control | Outside the rule, with OFAC advising caution | A separate test, with listed indicators | A separate test in the same legislation |
Read the table as layers, as From rule to workflow uses the word. The prohibitions are law in all three places. The ownership figure is legislation in the United Kingdom, agency interpretation in the United States, and nonbinding guidance in the European Union, which matters when a structure sits near a line and somebody asks how firm the line is. Which regimes a firm applies to which customers, and whether it reviews blocked stakes below the threshold, is policy. Screening the due diligence owners and calling the result ownership screening is convention, and it is the layer this piece argues against.
5. When the answer changes
Sanctions status can change while every field in the due diligence record stays the same, because designation happens to the holder. The day the company holding 30 percent of a customer is designated, the customer’s blocked ownership moves from zero to 30 percent, and the customer has done nothing at all. Recomputation therefore belongs to the list update, applied to every customer whose ownership graph contains the new name. Screening ownership at onboarding and at periodic review sets a refresh interval for a status that moves whenever a list moves.
The reverse movement calls for judgment. Once blocked ownership falls below 50 percent, OFAC’s rule stops treating the entity as automatically blocked, though property already blocked stays blocked unless OFAC authorizes otherwise. OFAC’s March 2026 sham-transaction advisory flags transfers to family members or close associates, and transfers made shortly before or after a designation; the EU Best Practices read share transfers close to a designation as a sign of retained control. Both call for a person rather than a formula.
6. Separate the arithmetic from the judgment
- Ownership graph: holders, stakes, sources, dates
- Arithmetic under each applicable regime
- Recomputation on every list update
- Judgment on whatever the arithmetic leaves open
Arithmetic is deterministic once the data exists. Each regime’s ownership result is a computation over a graph of holders and stakes, and it belongs in a system, rerun on every list update, with the regime and the list version stamped on the result.
Judgment covers what the arithmetic leaves open: control, joint arrangements, nominees, and transfers away from a blocked holder. Each is a decision with evidence, an owner, and a record, made by a reviewer equipped the way A human in the loop needs a job description specifies.
The data model follows from the split: ownership stored as edges, each carrying holder, held entity, percentage, source, and as-of date, with entity holders and small stakes kept, since aggregation needs both. How far below 25 percent to collect for sanctions purposes is a policy choice, and it deserves to be made as one.
A worked hypothetical
The following scenario is hypothetical and is included only to illustrate the framework. The companies and people below are invented.
A US business-banking platform opens an account for a components distributor incorporated abroad. The onboarding documents show its ownership:
- A holding company owns 40 percent.
- Holder A, an individual, owns 55 percent of the holding company.
- Holder B, an individual, owns 12 percent directly.
- Nine small holders own the remaining 48 percent.
At onboarding. Multiplied down the chain, Holder A’s stake is 22 percent and Holder B’s is 12. Nobody reaches 25 percent, so the due diligence record holds one name, the managing director, as control person. Screening clears the distributor and the director. The holding company sits in a PDF in the onboarding file.
Month seven. Holder A is designated. The holding company is now blocked, and its 40 percent counts in full: below the line, and squarely in the band OFAC’s guidance singles out for caution. No alert fires, because Holder A appears in no screened field.
Month nine. Holder B is designated. Blocked ownership reaches 52 percent, and the distributor is blocked while appearing on no list. Its payments keep flowing until a scheduled review eleven weeks later, when an analyst rereading the incorporation documents recognizes the holding company’s name. Every payment in between becomes a question for counsel.
Other readings. Proportional arithmetic gives 34 percent and clears the customer to the end. The UK test attributes the full 40 percent to Holder A through the majority-owned holding company and keeps Holder B’s 12 separate absent a joint arrangement, so the UK question turns on control.
What would have caught it. The holding company and both holders stored as edges with percentages, and OFAC arithmetic rerun on every list update. Holder A’s designation would have sent the customer to a named reviewer that day, under the firm’s policy for significant minority stakes; Holder B’s would have opened the blocking workflow. The firm held every fact it needed, in a form no system could compute on. As From recurring question to durable control argues, the fix is a structured field and a trigger.
7. Failure modes, and the measures that expose them
- The name check reported as an ownership check. Share of legal-entity customers whose stored ownership resolves, on every branch, to a natural person, a listed company, or a documented stopping point.
- The flattened graph. Share of ownership edges carrying a percentage, a source, and an as-of date.
- The onboarding clock. Time from a list update to a recomputed status for every affected customer.
- The unowned caution band. Customers with blocked ownership above zero and below the threshold, each with a named reviewer.
- One regime reported as clear. Results stored per regime, rather than as a single flag.
What this changes operationally
A sanctions owner states in the policy which ownership tests apply to which parts of the business, and the arithmetic of each, since the choice of algorithm is the control.
An onboarding or product lead separates the two questions on the form: the beneficial-ownership certification for due diligence, and a structured capture of who holds what, entity holders and small stakes included.
A data engineer stores ownership as a graph of dated, sourced edges and treats each list update as an event that recomputes every customer whose graph contains the new name.
An operations lead gives the caution band and divestment reviews a named queue and an owner, since those are the cases the arithmetic hands to judgment.
Limitations
This addresses the design of ownership screening. It offers no legal view on any structure, no thresholds, and no assessment of any data provider, and it describes no institution’s program. Each regime’s guidance carries detail a table compresses, including trusts and program-specific provisions, and the primary sources govern wherever this piece simplifies.
It also assumes the ownership data exists. Where registries are thin or slow, the graph is only as strong as its weakest edge, and the honest record says which edges rest on the customer’s own declaration.
Primary sources
- US Department of the Treasury, OFAC. Revised Guidance on Entities Owned by Persons Whose Property and Interests in Property Are Blocked (August 13, 2014).
- OFAC. Frequently Asked Questions 398 through 402, on control, aggregation, indirect ownership, and divestment under the 50 percent rule.
- OFAC. Guidance on Sham Transactions and Sanctions Evasion, Sanctions Advisory (March 31, 2026).
- OFAC and HM Treasury’s Office of Financial Sanctions Implementation. U.S. and UK Economic Sanctions Authorities: A Comparative Overview (June 23, 2026).
- FinCEN. Customer Due Diligence Requirements for Financial Institutions, 31 CFR 1010.230 (2016), and Frequently Asked Questions Regarding Customer Due Diligence Requirements for Financial Institutions, FIN-2018-G001 (April 3, 2018), Question 3.
- FinCEN. Exceptive Relief from Requirement to Identify and Verify Beneficial Owners at Each Account Opening, FIN-2026-R001 (February 13, 2026).
- Council of the European Union. Council Regulation (EU) No 269/2014, Article 2, and EU Best Practices for the effective implementation of restrictive measures, document 11623/24 (July 3, 2024), on ownership, aggregation, and control.
- HM Treasury, OFSI. UK financial sanctions general guidance, section 4, “Ownership and control.”
- United Kingdom. The Russia (Sanctions) (EU Exit) Regulations 2019, regulation 7 and Schedule 1.