Reading a corporate ownership structure to find the beneficial owner
Most training on beneficial ownership stops at the definition and the percentage. That leaves you unprepared for the file that actually lands on your desk: five layers, two jurisdictions, a nominee in the middle and a shareholder register that has not been updated since 2019. This note works a chain by hand and shows you where the arithmetic stops helping.
Why the definition is the easy part
Every analyst I have onboarded in the last decade could recite a beneficial ownership definition on day one. The natural person who ultimately owns or controls the customer. Fine. The trouble starts on day three, when they open a file containing a Cyprus holding company owned by a Luxembourg SPV owned by a Jersey trust with a corporate trustee, and the question is no longer what a beneficial owner is but which of the eleven names on the page qualifies as one.
Tracing ownership is a manual craft. Registry extracts help. Commercial databases help more than people admit and less than vendors claim. But the actual work — reading a chain, multiplying holdings, spotting where control detaches from shareholding — is done by a person with a pen, a blank sheet and a willingness to write down what they cannot see. In the teams I have run, the analysts who were good at this were rarely the ones who knew the most law. They were the ones who drew diagrams.
Working a five-layer chain by hand
Let me set out a structure of the sort I have unpicked many times, and then walk down it.
Your customer is Meridian Freight Services Ltd, an operating company incorporated in England. Its register shows two corporate shareholders: Aldergate Holdings BV holds sixty per cent, and Calder Logistics Partners LP holds forty per cent.
Aldergate Holdings BV, in the Netherlands, is in turn owned seventy per cent by Serrano Capital Sàrl in Luxembourg and thirty per cent by a named individual, Mr H.
Serrano Capital Sàrl is wholly owned by Vantor Group Ltd in the BVI. Vantor Group Ltd's shares are held by a corporate nominee, Bridgeview Nominees Ltd, on behalf of undisclosed principals. That is layer four, and it is where most analysts stop and escalate.
Now the multiplication. Mr H's thirty per cent of Aldergate translates to thirty per cent of sixty per cent of Meridian — eighteen per cent. Serrano's seventy per cent of Aldergate gives it forty-two per cent of Meridian, and because Vantor owns Serrano outright, Vantor also sits behind forty-two per cent. Whoever stands behind Bridgeview's nominee holding therefore holds an indirect economic interest of forty-two per cent in your customer, which is substantial by any measure and which you currently cannot name.
Meanwhile the other branch: Calder Logistics Partners LP holds forty per cent directly. A limited partnership is not a company and the register will not tell you what you need. You need the partnership agreement, the identity of the general partner, and — critically — the allocation of economic interest among limited partners, which frequently bears no relation to the voting arrangements. Suppose the general partner, Calder GP Ltd, holds a one per cent economic interest but full management control, and eight limited partners split the remaining ninety-nine per cent, none exceeding twenty per cent.
So where does that leave you? On a pure percentage reading of Meridian: no limited partner reaches five per cent of the customer. Mr H is at eighteen per cent. The unnamed principals behind Bridgeview reach forty-two per cent. And the person who actually decides what Calder does with its forty per cent stake — the individual behind Calder GP Ltd — may hold a fractional economic interest and complete practical control over two-fifths of your customer.
That last sentence is the whole point of this note.
Where the twenty-five per cent figure misleads
Many jurisdictions use a twenty-five per cent indicator as an entry point to beneficial ownership analysis; the UK's approach under the money laundering regulations, as unpacked in the JMLSG sectoral guidance, treats it as a trigger for further enquiry rather than a bright line beyond which nobody matters. The FATF standards themselves describe beneficial ownership in terms of ultimate ownership or control, and FATF has been increasingly pointed that a percentage test used alone produces incomplete answers. Other regimes set different figures or apply them differently. So do not carry a number across borders in your head.
Three failure modes recur.
Aggregation across related parties. Four siblings holding twenty per cent each is a family holding one hundred per cent, and treating it as four sub-threshold interests is arithmetic in service of nothing. I have seen structures where five holders at nineteen per cent each shared a correspondence address, a single professional adviser and consecutive share certificate numbers issued on the same day. Individually, nobody crossed the line. Collectively, the line was irrelevant.
Control without shareholding. The Calder GP situation above. Also: golden shares carrying veto rights, shareholders' agreements assigning board appointment rights to a minority holder, loan covenants that let a creditor replace management, and the individual who is neither shareholder nor director but whose instructions the directors follow. That last category is the hardest to evidence and the most important to name when you can.
Layered dilution. A person holding thirty per cent at each of four levels holds roughly zero point eight per cent of your customer by multiplication, and may nonetheless be the only human being making decisions anywhere in the chain. Multiplication tells you about economics. It tells you very little about influence.
Nominees, mid-chain opacity and bearer instruments
A nominee shareholder sitting at layer one is an inconvenience. A nominee sitting at layer four, as Bridgeview does above, is a wall — because everything above it is unverifiable and everything below it now rests on an assumption you have not tested.
Nominee arrangements are lawful and commonplace. Professional trustees, custodians and law firms hold shares on behalf of clients for reasons that have nothing to do with concealment. The compliance question is not whether a nominee is present but whether you can obtain a declaration of the beneficial interest behind it, from a source with something to lose by lying. A signed nominee declaration from a regulated trust company in a jurisdiction with a functioning supervisor is worth a great deal. An unsigned letter from an unregulated corporate services provider is worth roughly what it cost to produce.
Bearer shares, where you still encounter them, defeat ownership tracing by design: title passes with possession of the certificate and no register records the transfer. Many jurisdictions have abolished or immobilised them over the past fifteen years, often requiring deposit with an approved custodian. When bearer instruments appear in a chain, I want two things — evidence of immobilisation with a named custodian, and a current declaration of holders. Absent both, I treat the layer as undetermined and say so in the file rather than papering over it.
Case note
A commercial bank I worked with onboarded a metals trading company in March 2019. The structure ran four layers to a Seychelles company whose shares were held by a corporate nominee. The relationship manager obtained a nominee declaration naming a single individual, Mr K, described as sole beneficial owner. It was accepted. Annual reviews in 2020 and 2021 rolled forward the same declaration without re-verification.
In August 2022 the account began receiving third-party payments from four counterparties that had not appeared in the previous three years — fourteen inbound transfers totalling roughly £6.3m over nine weeks, against a stated annual turnover of £4m. Alert triage escalated it. The subsequent review established that the nominee declaration had been superseded in November 2019 by a deed reallocating the beneficial interest four ways, with Mr K retaining twenty-two per cent. Nobody had asked. The bank had spent thirty-two months believing it knew a UBO it had actually stopped knowing three years earlier.
Two of the three newly identified individuals were resident in a jurisdiction the bank's own risk appetite excluded for that product. The relationship was exited over four months to December 2022, and the review of nominee-held structures that followed re-papered about two hundred and forty files.
Concluding that the UBO cannot be determined
Analysts are trained to produce answers, so they will manufacture one rather than record a gap. This is the single most damaging habit I encounter in ownership work. "UBO: Mr H (18%)" written on a file where forty-two per cent sits behind a nominee is not a conclusion. It is a decision to stop looking, dressed as a finding.
There is a legitimate and defensible outcome that reads: on the information available, the beneficial owner of the forty-two per cent indirect interest could not be determined, for these reasons, having taken these steps, on these dates. That is a proper record. It is also, in most frameworks, a fact with consequences.
What follows from an undetermined UBO depends on your jurisdiction, your sector, your product and your own risk appetite — and the FCA has been consistent in its expectation that firms apply their stated appetite rather than admire it. Broadly, the options narrow to three: obtain the missing information, restrict the relationship to something you can justify without it, or decline. In practice I have seen banks maintain relationships with a partially undetermined chain where the customer was long-standing, the operating business was visible and verifiable, and the undetermined layer was small. I have also seen an undetermined layer above forty per cent treated as fatal, correctly.
Record the enquiry trail as carefully as the answer. Which registry, extracted on which date, showing what. Which document requested from the customer, on which date, and the response or the silence. When someone reads the file in three years — a supervisor, a court, your own successor — the value of that trail is that it shows a firm that looked properly and reached a documented limit. This is the same discipline that makes a good EDD file defensible, and the same discipline that produces a usable narrative if the relationship later generates a report. If you need to describe an opaque structure for a disclosure, the mechanics of writing a SAR narrative depend heavily on how legibly you documented the chain in the first place.
A short practical sequence
When I hand a structure to an analyst, I ask for it in this order.
Draw the chain on paper before touching a database, using only what the customer has told you. Then verify each layer against an independent source and mark on your diagram which layers are verified and which are asserted — the difference between those two colours of ink is your real finding. Multiply the economic interests down to the customer. Separately, and this is the step most often skipped, list every mechanism of control you can identify that does not appear in the shareholding: board appointment rights, veto shares, general partner powers, management agreements, funding dependencies, and the family relationships that make four small holdings one large one.
Then name the natural persons. Then name the gaps.
Junior analysts often meet these structures first through an alert rather than through onboarding, which is a harder way to learn; the alert triage discipline of establishing what the customer is before assessing what the payment means applies with double force when the customer is a five-layer chain. And when you reach the wall, escalate the wall. The Wolfsberg Group guidance on correspondent and entity due diligence is useful ammunition when you need to explain to a front office colleague why an undetermined forty-two per cent is not something a strong commercial rationale can cure.
What matters: An ownership chain you cannot fully see is a finding to be recorded and acted on, not a gap to be filled with the largest name you happened to identify.