Worked structure eleven · Structural Second Opinion
A structure already drawn by somebody else.
The manager holds an architecture drawn by another adviser and wants to know whether it holds before committing resources to it. The drawing carries its entities, its jurisdictions, its instruments and its flows on its face. It carries nothing about the question it was built to answer.
A second opinion does not grade the adviser who drew the structure and does not redraw it by preference. It tests the structure against what binds it.
Worked structures · fifteen rooms
- 01An existing platform
- 02An ineligible asset
- 03An investor class
- 04A narrow exit
- 05A repeatable structure
- 06A co-investment
- 07A continuation
- 08Unavailable security
- 09The seed terms
- 10A change of control
- 11Drawn elsewhere
- 12An in-kind distribution
- 13A change of domicile
- 14An investor's perimeter
- 15A strategy, no vehicle
01 · The transaction as it arrives
The structure is finished and coherent. The question it answered is not written on it.
Four arrivals and one disguise. One is a decision not yet taken, one is a decision taken and not yet paid for, one is a document round that will not close, and the fourth is a structure that was right on the day it was drawn and is now asked to hold facts it never saw.
The manager holds a drawing, a memorandum and a proposal, and has committed nothing beyond reading them. Every route is still a route, including those the drawing does not contain. The cheapest of the four arrivals, and the rarest.
The route is chosen and counsel is instructed. What a review can still change is the instruction rather than the structure, which is a smaller thing to change and far cheaper than anything that follows.
A provision will not close. Three parties are drafting against each other and none can say whether the difficulty sits in the drafting or in the design. The structural question is underneath the drafting one, and it is the question nobody has been asked.
The structure was drawn against a set of facts and one has changed: a class of investor arrived, an asset was added, a facility was taken on, a place changed a rule underneath it. Nothing in the structure is defective. It answers a question the transaction has stopped asking.
Almost never "review this structure". It arrives as a view wanted before an investment committee, a diligence item an incoming institution raised, or one sentence in a memorandum nobody can source. The structural question sits behind whichever came through the door, and it is the same in all four.
The structure in this class is not defective work. It is competent work, drawn by an adviser asked a question and answering it. What makes it this case is that the question was put in one room, the structure is relied on in another, and nothing travels between the two except the drawing.
No manager, adviser, asset, counterparty, value or date appears in this room. What transfers from one transaction to the next is the shape of the problem, and the shape is what is set down here.
02 · The structural problem
Testing a structure drawn elsewhere begins by recovering the question it was asked.
A structure is a completed answer. Two things are needed to test one: the answer, and the question it was given. Only the first of them arrives.
Whether an architecture is coherent can be read off a drawing by anybody who knows how to read one. Every entity has a reason to exist or it does not. Every duty has an owner, and no two parts assume different facts. That is the easy half of the work, and the half that reliably gets done.
The other half decides more. A structure that is correct for the question it was asked can still be the wrong structure for the transaction, and nothing on the drawing distinguishes those two conditions. The drawing does not say what facts the adviser was given, which questions sat expressly outside the scope, what the manager instructed it to assume, or what has changed since the day it was handed over.
There is an obvious way to fill that gap and it is the wrong way. The reviewer compares the structure in front of it with the structure it would have drawn, and reports the differences. That produces a document which reads like a finding and is not one. A preference cannot be tested by the manager, answered by the adviser who drew the structure, or routed to anybody who could settle it. It changes nothing except the relationship between two advisers.
What the drawing carries, and what it does not
The drawing carries the entities, the place each sits, the instruments between them, the flows, the security and, where it is careful, the order of the steps. Every one is testable against a document somebody can produce, which is what makes a structure reviewable at all.
The drawing does not carry the instruction, the facts supplied, the scope boundary, the assumptions the adviser was permitted to make, or the date at which each was true. Those live in the engagement paper, the fact memoranda, correspondence, and the memory of the people who were in the room. All are recoverable in an afternoon. None is recoverable from the drawing.
The distinction decides what a finding means. A row the structure assumes and no document supports is one thing where the adviser held that question, and quite another where the manager kept it. They call for different repairs, and marking them the same way turns a review into an accusation. An accusation is not a structural output.
A structure cannot be tested against a preference. It can only be tested against what binds it, and what binds it is a set of documents somebody can produce.
03 · What binds
Five of the six are documents. The sixth is the instruction, and it decides what every finding means.
Six things bind the reading. Five come from a file. The sixth comes from what people remember being asked, and it is the one that separates a finding from a verdict.
The constitution, which settles what may be held, who may be admitted and what may be amended by whom. The instruments the structure runs on: the asset contracts and the facility, with the transfer, pre-emption and change-of-control provisions inside them. The perimeter, which decides what activity each entity performs and whether the party performing it is permitted to. What the investors were offered, which fixes the population the vehicle can now admit. What the operating parties can run, which decides whether the mechanic survives a register, a valuation cycle and a reporting date. And the scope the first adviser was given, the one item with no document of its own.
Each one, by source and by what can move it
Fixed at formation. Movable by consent, which means movable at a price, paid to the holders who are asked to give it.
Contracts signed by parties who are not in the review and have no reason to attend it. Each carries a consent, a right or a trigger the structure accounted for or did not.
The rules of the place each entity sits in, and the permissions the parties hold. Not movable by anybody in the transaction, and it settles which entity may perform which act.
The offer document, the subscriptions and the side letters. Fixed at the point each holder signed, read as at that point rather than as at today.
The administrator, the auditor, the depositary and the register. Movable by appointment and by system, on their timetable rather than the transaction's.
Nothing on the drawing. The engagement paper the first adviser worked to, the facts supplied, the scope given. Recoverable in an afternoon, lost entirely if nobody asks in the first week.
The first five are read as they stand rather than as the drawing assumed them. Where a document has been amended since the structure was drawn, the difference is real and the structure is not at fault for it, and a review that does not make that distinction produces findings nobody can use.
The sixth is asked for in the first conversation, before anything is read. It costs one question put twice: what was the adviser asked, and what was it given. Where nobody now knows, that is itself the first finding, and it belongs to the manager.
The structure is read against what binds it, never against what anybody would have preferred. The sixth item is what keeps a finding from becoming a verdict on a person.
04 · The architectures considered
Three ways to read a structure somebody else drew, and one of them produces a finding anybody can answer.
Three review architectures, differing in what they read and in whether the conclusion can be answered by the parties who have to live with it.
Reads the drawing alone: an entity with no purpose, a flow with no destination, a duty with no owner, two parts assuming different facts. It cannot reach anything the drawing does not say, and most of what decides a structure is not on the drawing.
Reads the drawing against the instruments that govern it, row by row, each cited to the provision that settles it. The only one of the three whose output can be handed to counsel, the administrator, the lender and the adviser who drew the structure, and answered by each on the same terms.
Reads the objective and starts again. Legitimate work, and not a review: it is transaction architecture run a second time, and it is honest to give it that name and its own paper. Delivered as a review it leaves the manager holding two drawings and a test of neither.
The binding read is what the rest of this case describes. Its output is a ledger rather than a report: one row per assumption, the provision that speaks to it, the mark, and the party who can close it.
Four rows of the kind the ledger carries. Three marks exist, and one of them is a problem inside the structure.
| The assumption · the reading | What the structure assumes | What the governing document says | The mark, and what turns on it |
|---|---|---|---|
| That the vehicle may hold the asset | That the investment powers reach this asset class, in the form the drawing holds it and at the tier it holds it. | The constitution states the powers, more often than not by reference to a schedule drafted for the portfolio the vehicle was raised against. | Tested. The provision is short, it sits in one place, and it either reaches the asset as held or it does not. |
| That the investors named may be admitted | That every investor in the plan may hold interests in a vehicle of this type, formed here, at this tier. | The constitution and the offer set the population the vehicle may admit. What each investor may hold is set by its own regime, and that document is not in the fund's file. | Untested. Nobody has read the investor-side instrument, because nobody inside the transaction holds it. |
| That title moves without a third party | That the asset, or the entity owning it, can be transferred at the step the drawing shows, between the parties it shows. | The asset contracts and the facility carry transfer, pre-emption and change-of-control provisions, each installing a party whose agreement the step requires. | Contradicted. A provision names a consent the drawing does not show. The step is not unavailable; it is conditional, and the condition sits outside the transaction. |
| That the mechanic can be operated | That valuation, register and reporting carry what the structure asks of them, on the timetable it sets. | No document answers this. The administrator, the auditor and the keeper of the register answer it, in writing, against the drawing. | Untested. The cheapest row to close, most often left open because the party who can close it is not yet appointed. |
05 · The critical dependencies
Three dependencies, and the first decides whether the other two mean anything.
Each is a condition on the review rather than on the structure, and each has an owner who is not the reviewer.
That the instruments produced are complete and current. Owned by the manager, satisfied by a list rather than an assurance: every document named on the drawing, in the version in force, dated. A review run against a superseded constitution produces findings that look exactly like findings and are not.
That the instruction given to the first adviser can be recovered. Owned by the manager: the engagement paper, the facts supplied, the scope, and anything the adviser was told to take as given. Without it every silence in the drawing reads as an omission, and a silence that was instructed is not one.
That every finding has a destination before it is written. Owned by the manager through its appointments: a contradicted row is a question for counsel, an operability row for the administrator, a facility row for the lender. A finding written without a destination arrives at the manager and stops there.
What each costs to hold, and what it costs to skip
The first costs one list and one afternoon. Skipped, it costs the whole review, because a ledger built against the wrong version of a document cannot be repaired row by row. It has to be built again, and by then the manager has read the first version.
The second costs one conversation, held while the people who were in the room are still reachable. Skipped, it costs the tone of the whole engagement. Every gap becomes an accusation, and the adviser who drew the structure is left defending its work against a scope nobody has stated.
The third costs nothing at the time and everything afterwards. A ledger with no owners is a document the manager reads once and files. The same rows with an owner and a date against each is a work list that closes itself over a quarter. Both are written in the same afternoon.
A finding with no document behind it and no owner in front of it is not a finding. It is an opinion about somebody else's work, and nobody can act on one of those.
06 · The architecture that survives
What survives is the structure as drawn, with its assumptions marked and its questions routed.
The constraints do not leave a better structure standing. They leave the same structure, tested, and a short list of who has to answer what. To a manager who will have to defend the architecture to somebody else that is worth more than a second drawing.
Three outcomes exist, and all three are proper results of the work rather than degrees of failure.
The first and commonest is that the structure holds. Every row is tested or closed, and the manager holds a document that says so, cited provision by provision, which it can put in front of an investment committee or a diligence team. It is the outcome managers least expect and the one most often produced, because the structure was drawn by a competent adviser answering a question properly put.
The second is that the structure holds for the question it was asked and the question has moved. The repair is to the instruction rather than to the drawing, made by going back to the adviser who drew it with the fact that changed. That is a short conversation, and it is available only to a manager who knows what the original instruction was.
The third is that a row nothing supports is carrying weight. It becomes a question, put to the party who can settle it, before it becomes a change to anything. A question routed to counsel comes back as a position taken in counsel's own name. A change made by a reviewer on its own authority comes back as nothing, because nobody who has to stand behind it has signed it.
The order the findings travel in
The ledger goes to the adviser who drew the structure before it goes anywhere else. That is not a courtesy. It is the fastest way to close rows, because a substantial share of the contradicted marks resolve into an instruction the reviewer did not hold, and they resolve inside a day. What is left after that pass is the real ledger, shorter than the first and harder in every row.
Each remaining row then travels to the party who can settle it, framed as a question with the provision attached, and returns as a position taken in that party's own name. The manager reads the closed ledger and decides: adopt as drawn, repair the rows that need repairing, or instruct a redraw as a new piece of work. All three belong to the manager and none to the reviewer.
The output of a second opinion is not a better structure. It is the same structure with its assumptions visible, and a list of who has to answer what.
07 · The implementation framework
Six items in order, and the fifth is where the tone of the whole engagement is set.
Written as dependency statements rather than as tasks. Two of the six are done before a line of the structure is read, and skipping either of them is what makes the other four unusable.
Fix the scope in writing. Assemble the instruments in the versions in force. Recover the instruction the structure was built to answer. Build the ledger against the documents. Route the findings, to the adviser who drew the structure first. And leave the decision with the manager, as three options rather than one.
The signatures then belong elsewhere, as they should. The manager's counsel opines in its own name on the rows that turn on what a document means. Its tax advisers take the formal position wherever the treatment of the chain is something the structure assumed. The administrator answers the operability rows. The lender answers what its facility requires. The adviser who drew the structure answers the rows that turn on the instruction it was given. Our work is the scope, the instruments, the recovery, the ledger and the routing.
The sequence, and what each step depends on
- 01 The scope, written before anything is read Depends on nothing but the manager stating what it wants tested and what it does not. It also states what a finding will look like, the sentence that stops a review drifting into a redraw.
- 02 The instruments, in the versions in force Depends on the manager producing a complete, dated set. What is missing is named rather than read around, because a gap here is a gap in every row that touches it.
- 03 The instruction, recovered Depends on the people who were in the room and on the engagement paper the first adviser worked to. Done in the opening week or not at all, because every later step reads differently once it exists.
- 04 The ledger, built row by row Depends on the two above. Each row carries the assumption, the provision, the mark and the party who can close it. Nothing enters that cannot be cited.
- 05 The findings, routed Depends on the ledger being complete rather than interesting. First to the adviser who drew the structure, with the instruction attached, so a row explained by scope is closed by scope. Then outward to the parties who can settle what remains.
- 06 The decision, left with the manager Depends on the five above sitting on one page. Three options, set out in what each costs the sequence rather than in what anybody prefers: adopt, repair, or draw again.
The order protects everybody standing in it. A finding produced before the instruction is recovered is aimed at a person rather than at a structure.
08 · What this case generalises to
Wherever a structure was built to answer a question, the question has to be legible before the structure can be tested.
Four families share the shape of this one. In each, a structure arrives complete, its instruction does not travel with it, and the party relying on it is not the party who set it.
A platform that came with an acquired manager, a vehicle taken over mid-life, an architecture that arrived attached to a seed investor. It works, nobody in the house built it, and the first person to ask why an entity is in the chain is usually somebody in a diligence room.
A class of investor admitted after formation, an asset added to a portfolio drawn for a different one, a facility taken on afterwards, a rule changed underneath. The structure did not fail. It answers a question the transaction stopped asking, and nothing in the ordinary running of a fund prompts anybody to notice.
An investment committee, an incoming institution's diligence, a lender's credit process, an auditor at a first year end. A defence is built of cited rows rather than of confidence, and a manager who holds the ledger answers in a week what otherwise takes a quarter.
Counsel answered the legal question it was asked, the tax advisers answered theirs, and nobody read the two answers against each other. Neither was wrong and the structure still assumes two different things. This is the ordinary origin of a contradicted row, and it is nobody's error.
The engagement that answers this class is a Structural Second Opinion: the scope, the instruments, the recovered instruction, the ledger and the routing. It tests coherence, then it tests the structure against what binds it. It does not grade the adviser who drew the structure and it does not redraw it by preference. It stops where the positions start.
The adviser who drew it answered the question it was asked, and answered it properly. Whether that is still the question is the manager's to know, and it is the whole of what a second opinion is for.
Written as a type · no party, no value and no date · stated as at August 2026
Write to us.
Complex transactions fail at the interfaces between otherwise workable components. We resolve the structural complexity between investment intent and transaction execution.
Write to us