McKinsey, Puerto Rico and MIO: The Debt Island Case

A conflict review needs a map of decisions

The useful consulting question is not whether a large organisation can describe itself as independent. It is which people can see which information, which decisions they influence and which interests might benefit. A group can have separate legal entities and separate investment staff while still needing controls over shared oversight roles. Conversely, the existence of affiliated investments does not, by itself, establish that an engagement was improperly influenced.

Begin with a decision map. Identify the client’s decisions, the adviser’s recommendations, the investment office’s decisions and the committees that oversee each. Then mark who participates in more than one part of the map. This makes a review specific enough to test. It avoids both a blanket assumption that every affiliate creates a disqualifying conflict and an equally broad assurance that separate entity names resolve every concern.

The Rajat Gupta case concerns a different legal record, but illustrates why access to confidential information is valuable and why its use must be bounded. Do not import that case’s conviction into this one. Use the comparison to distinguish an information-access risk, an inadequate control and a proved misuse of information.

Look through the structure without presuming the answer

A conflict inventory should distinguish direct holdings, externally managed accounts, pooled investments and a person’s ability to oversee or influence allocations. Visibility and control are different. A person may know a holding without selecting it; an oversight committee may influence a policy without placing a trade. The response should be designed for the actual combination rather than for the most convenient organisational label.

For a hypothetical advisory engagement, prepare a register with columns for the interested party, nature of the connection, information available, decision authority and proposed response. Record uncertainty explicitly. A missing answer should trigger a follow-up owner, not be silently converted into an assurance that there is no connection. The register is an analytical aid; the governing legal assessment belongs with the appropriate qualified reviewers.

This is also a lesson about how a professional-services firm explains itself. The client needs enough information to evaluate independence without receiving irrelevant confidential investment details. Design that disclosure process deliberately. An uncontrolled dump of holdings is not necessarily a better review than a carefully scoped process with independent challenge and documented accountability.

Financial measures need their own labels

Debt face value, investment assets, advisory fees, creditor recoveries and regulatory penalties answer different questions. Each needs an entity, date, scope and measurement basis. A large fee can justify procurement scrutiny without proving biased advice. A small percentage of a portfolio can matter to a control assessment without establishing a material incentive for every employee. Neither conclusion can be reached from a dramatic side-by-side chart alone.

Consider a fictional slide with three boxes: an affiliate’s bond holding at face value, an adviser’s cumulative fees and a penalty imposed years later. An arrow connecting the boxes can make them look like one money flow. Before showing it, ask what each arrow claims. Does it represent an actual payment, ownership, an information link or merely a question? Use different notation and a clear legend for those relationships.

The France procurement case similarly requires aggregate spending, contract fees and investigative measures to stay separate. A useful presentation does not reduce complexity by removing the distinctions that determine what the evidence supports. It reduces complexity by making those distinctions legible.

A fixed fee needs a visible accountability mechanism

A fixed monthly fee is a commercial arrangement, not a verdict on value. It can help a client manage scope and budget, but it can also make effort and deliverables difficult to assess if the agreement does not specify what will be produced. The review should connect the price to outputs, decision support, capability transfer and the procedure for approving changes.

For a public or restructuring engagement, an accountability pack might include a workstream register, named deliverable owners, acceptance criteria, unresolved questions and the basis for any extension. These are proposed management practices, not claims about documents found in this historical engagement. They allow a buyer to ask whether work is useful without assuming that an hourly breakdown is the only possible measure of performance.

The Allstate claims case asks how a performance system can change the behaviour it rewards. Here the analogous question is what the fee and acceptance process makes visible. Define success in terms the client can inspect, including the quality of the decision record, rather than the volume of activity that an adviser reports.

An information barrier must cover exceptions and oversight

An information barrier is a working system of permissions, role restrictions, supervision and escalation. Its weakest point may be an exception made for a senior person or a committee role that does not resemble day-to-day investing. Testing therefore needs to cover both the ordinary workflow and the people who sit above it.

A review can ask how access is granted, how it is removed, how overlapping roles are identified and how exceptions are approved. It can examine whether the same person both benefits from an exception and decides whether that exception is acceptable. None of these questions presumes wrongdoing. They are ways to make a general assurance testable before a problem becomes a disputed historical account.

The Bain–Guinness case provides another setting in which an adviser’s organisational role deserves precision. A secondment, a board role and a consulting assignment can create different duties and reporting lines. The common lesson is to describe authority at the level where the real decision is made, while keeping the legal outcomes of the two cases separate.

Different reviews can answer different questions

A board-sponsored review, an investment-adviser proceeding and a court’s professional-compensation process need not have the same subject, period or test. A reader should be able to see what each examined and what it did not decide. One favourable conclusion does not automatically resolve every broader concern; one adverse finding does not retroactively establish every allegation in a different process.

Build an evidence table with one row per proceeding. State the body, respondent, question, period, outcome and limitation. This is especially helpful when several events share a brand name. It prevents a reader from treating the firm, an affiliate and an individual as interchangeable subjects. The table also makes uncertainty visible without burying it in a general disclaimer.

The Enron case shows how easily an admired management narrative can eclipse independent scrutiny. The solution is not to presume the opposite narrative. It is to make the evidence and the decision standard explicit enough that a reviewer can challenge both the reassuring story and the dramatic one.

The useful output is a decision the client can defend

For practitioners, the final product should identify the relevant connection, the risk it creates, the proposed response and the person accountable for maintaining that response. It should distinguish a disclosure issue from a control issue and distinguish both from an established misuse of information. Recommendations need this precision to be operational rather than merely cautionary.

The management consulting presentation guide offers a way to turn that work into an explicit recommendation. Use an action title that states what the client should do, an exhibit that supports the action and a note identifying what remains uncertain. The purpose of the Debt Island case is to improve that reasoning, not to make every organisational overlap look like proof of a corrupt result.