Sanctions ownership: why the fifty per cent rule catches people out
Most sanctions training explains the fifty per cent rule in a sentence and then moves on. The sentence is correct and almost useless, because the cases that cause trouble involve two designated shareholders who individually fall below the line, holdings buried two layers down, or control exercised with no shares at all. This note works a chain by hand and shows where the arithmetic bites.
The rule sounds simple, which is the problem
Ask a new analyst what the fifty per cent rule is and you will usually get something like: "if a designated person owns more than half of a company, the company is sanctioned too." That is close enough to be dangerous. It is the shape of the rule, not the rule, and the gap between the two is where I have seen firms process payments they should have frozen.
Three things get lost in the short version. The first is aggregation: the test is not about one owner but about the combined holdings of designated persons. The second is indirection: ownership counted through intermediate companies, not just the shares on the customer's own register. The third is control, which in some jurisdictions does not require a shareholding at all.
I want to be careful about the framing before going further. Working out whether an entity is caught by an asset freeze is a legal determination about status. It is not a view about anybody's conduct. The customer may have no idea that one of their minority shareholders was designated three weeks ago. The arithmetic does not care, and nor should your file notes.
Working a chain by hand
Here is a structure close to ones I have seen several times. I will use neutral labels and do the sums on the page, because the only way to get comfortable with this is to push the numbers around yourself rather than trust a screening tool's green tick.
Your customer is an operating company — call it Opco — incorporated in a jurisdiction with a public register. Opco's shares are held fifty-five per cent by an intermediate holding company, Midco, with the remaining forty-five per cent split between two investment funds. Midco's own register shows three blocks of shares: thirty per cent to a Mr A, thirty per cent to a Mr B, and forty per cent split across three individuals with no apparent connection to either.
Step one: aggregate at the layer where designations sit
Mr A is designated. Mr B is designated, under the same programme but on a different date. Taken separately, each holds thirty per cent of Midco, which is below any ownership threshold you will encounter. Taken together, they hold sixty per cent.
Under the approach OFAC has set out for US sanctions, the holdings of multiple blocked persons are added together, and an entity owned fifty per cent or more in the aggregate by blocked persons is itself blocked, whether or not it appears on any list — see OFAC for the published guidance. So Midco is a blocked person. Nobody designated it. Nobody published its name. It is blocked by operation of the arithmetic.
Step two: carry the status down, do not multiply the percentages
This is the step that catches people, and I have watched capable analysts get it wrong on a whiteboard. The instinct is to multiply: sixty per cent of Midco, Midco holds fifty-five per cent of Opco, so the designated interest in Opco is thirty-three per cent, which is under half, so Opco is fine.
That is not how the US test is applied. Once Midco is itself treated as a blocked person, Midco's own holdings are assessed afresh. Midco holds fifty-five per cent of Opco. Fifty-five is more than fifty. Opco is blocked. The chain propagates stepwise down the structure rather than diluting as it goes.
Run the same structure with Midco holding forty-nine per cent of Opco and the answer flips, assuming nothing else is going on. One percentage point of share capital, two layers up from your customer, decides whether you can move money today.
Step three: look for parallel paths
Structures rarely have one route from the designated person to the customer. If Mr A also holds forty per cent of one of the investment funds sitting on Opco's register, you now have a second line of influence to assess, and you may find that the funds and Midco are not as unconnected as the register suggests. Reading these chains properly is a skill in its own right, and I have written separately about tracing an ownership structure to the ultimate beneficial owner when the layers do not line up neatly.
Case note
A payment institution I worked with onboarded a commodities trading company in February 2022. The ownership chart in the file was dated to onboarding and showed a single holding company parent, which in turn showed four individual shareholders at twenty-six, twenty-six, twenty-four and twenty-four per cent. All four screened clear. Over the following fourteen months the customer sent about four hundred and ten payments through the firm, totalling roughly €18.4m, almost all to two counterparties in the same corridor.
In April 2023 two of the four shareholders were designated eleven days apart. Neither designation triggered an alert against the customer, because the customer's own name was not listed and the screening configuration matched counterparties and account parties rather than shareholders recorded in a PDF at onboarding. The combined holding was fifty-two per cent. The firm found it in June during an unrelated periodic review, by which point a further €1.37m had moved across nine payments. The remediation ran to about seven weeks and the hardest part was not the freeze — it was reconstructing, payment by payment, what had been actioned after the first designation date and what had cleared before it.
Control without a single share
Ownership is the easier limb, because it is countable. Control is where judgement comes in, and where I spend most of my supervision time with analysts.
In the United Kingdom, the asset-freeze regulations reach entities that are owned or controlled directly or indirectly by a designated person, and the control limb does not depend on holding shares. A designated person who can appoint or remove a majority of the board, or who can direct the entity's affairs in the way they wish, may bring the entity within scope on that basis alone. OFSI publishes guidance on how it reads those tests and it is worth reading in full rather than second-hand.
What does control look like in a file? Usually something mundane. A shareholders' agreement granting veto rights over budgets and senior appointments. A loan from a connected party with covenants that make the lender the real decision-maker. A sole signatory on the main operating account who holds no equity. A founder who stepped off the register in favour of a spouse and a long-standing associate, while the registered office, the auditor and the email domain all stayed the same.
Those signals belong in the enhanced due diligence file rather than in a screening tool, which is one reason I am insistent about what an EDD file should actually contain for complex corporate customers. A tool compares strings. It cannot read a shareholders' agreement.
The jurisdictions do not agree, and you need to know which you are in
I have sat in rooms where a group-wide policy said "fifty per cent" as though that settled the matter across every entity in the group. It does not.
The US aggregation approach is explicit and widely published. The UK ownership limb is framed around holding more than fifty per cent of shares or voting rights, with that separate and significant control limb alongside it — and "more than fifty" is not the same line as "fifty or more", which matters in an exactly-even split. The EU regime also carries a control test and the Commission has issued its own guidance on how the criteria interact. Beyond those three, treatment of aggregation and of indirect holdings varies, and some regimes say considerably less than practitioners would like.
The practical consequences are these. A single global rule set to the strictest reading is defensible and costs you some false positives. A rule set to the most permissive reading is not defensible anywhere. And a payment that is clean under one regime may be prohibited under another touching the same transaction through a correspondent, a clearing currency or a group entity — a point that overlaps heavily with the risks in nested correspondent relationships. The Wolfsberg Group materials on sanctions screening are a reasonable starting point for thinking about where in the payment chain a determination actually has to be made.
A structure chart is evidence of a date, not of a fact
This is the operational lesson from the case note and the one I repeat most often. The chart you obtained at onboarding was accurate on the day you obtained it. Shares move. Designations land. A chain that was fifty-two per cent clean in March can be fifty-two per cent caught in April without anybody at the customer doing anything at all.
So the question for your firm is not "do we understand the fifty per cent rule". It is: when a new designation is published, what in our estate re-runs the ownership arithmetic against the shareholders we recorded at onboarding, for which customers, and how quickly? If the honest answer is "the next periodic review", you have a gap measured in months. The FCA has been clear in its published findings that sanctions systems are judged on operating effectiveness rather than on the existence of a policy, and the same logic applies to any supervisor.
Some things I ask teams to put in place, in rough order of how much they repay the effort:
- Capture shareholders as structured data. Names, percentages, dates of birth where available, and the date the information was obtained — held as records that can be screened, not as an attachment nobody will open again.
- Screen the ownership layer, not only the customer and counterparty layer. This is a configuration question and frequently a licensing cost question, and it is the single change that would have caught the case above in days rather than weeks.
- Make the aggregation step explicit in the procedure. Analysts will not aggregate holdings across two designated persons unless somebody has told them to, with a worked example in front of them.
- Record your reasoning when you conclude an entity is not caught. The percentages you relied on, the source, the date, and why you concluded that no control limb applied. A negative determination with no reasoning behind it is indistinguishable from not having looked, and that distinction will matter in a review two years later.
- Treat a partial name match on a shareholder as a live question. The disposal logic is similar to any other match and I have set out how I approach clearing a partial sanctions name match elsewhere.
None of this is exotic. The fifty per cent rule catches people out not because it is conceptually hard but because the arithmetic has to be redone every time the world changes, and most firms do it once, at onboarding, on a document that was already out of date when it arrived.
What matters: The percentage you verified at onboarding is a fact about that day, and a designation published since then can change the answer without anybody at the customer lifting a finger.
