Greensill Capital made supply-chain finance and the regulatory perimeter a supervision accountability test

Summary

• Greensill Capital entered administration in March 2021 after a funding model dependent on credit insurance and investment funds could no longer continue. The immediate trigger involved the loss of insurance support and suspension of funds that bought Greensill-originated assets, but the accountability problem was broader: asset verification, obligor concentration, insurance renewal, valuation, liquidity and regulatory ownership were distributed across firms and jurisdictions.
• Supply-chain finance itself is not synonymous with misconduct. Traditional structures can accelerate payment to a supplier against a buyer's confirmed obligation. Greensill also financed what were described as prospective receivables, where the expected buyer obligation had not necessarily arisen. That distinction changes the evidence needed to establish an asset, the credit risk being taken and the disclosures an investor requires.
• Parliamentary reports, regulator statements, public audit work, insolvency filings and enforcement records answer different questions. Committee conclusions are scrutiny findings, an administrator's report records an office-holder's work and estimates, an investigation records unresolved suspicion until disposition, and a regulatory censure or final notice establishes only the matters within its legal scope.
• Durable reform requires more than moving one activity inside a statutory perimeter. It requires verified receivables, visible concentrations, stress tests for insurance and fund withdrawal, principal-firm oversight of appointed representatives, public-sector due diligence that survives urgency and access, comparable supplier-finance disclosures, and recovery reporting that separates claims, cash collected, costs and distributions.

A collapse assembled across several markets

Greensill Capital presented itself as a provider of supply-chain finance: it paid suppliers before invoices fell due and expected payment later from the buyer. A plain version of that transaction can serve a legitimate working-capital purpose. The buyer confirms an invoice, a finance provider advances money to the supplier at a discount, and the buyer pays the finance provider on the agreed date. The transaction still has risks, but its basic evidence can be tested: a named supplier, a delivered good or service, an accepted invoice, a named obligor, an amount and a maturity.

The Greensill structure became harder to assess as origination, insurance, securitisation and investment distribution were connected. Financing assets could be sold or transferred into vehicles and acquired by funds. Credit insurance could make exposures appear more acceptable to investors or support risk limits. Cash collected from obligors had to reach the right vehicle. Investors depended on asset-level information, valuation and liquidity management by institutions beyond the originator. A failure at one node could therefore remove confidence and funding from the others even before the underlying receivables matured.

The Treasury Committee's report on Lessons from Greensill Capital provides the most useful UK parliamentary synthesis. It described traditional supply-chain finance and separated it from financing against prospective receivables. It recorded evidence about Greensill's concentration to the GFG Alliance, the role of investment funds and insurance, and the use of an appointed representative for regulated activity. It also concluded that the collapse did not itself demonstrate a need to regulate all supply-chain finance.

That is a committee conclusion, not a judicial finding about every transaction, and its 2021 view must be read alongside later regulatory outcomes.

The sequence exposed a structural assurance problem. Each entity could examine the portion assigned to it while missing how the whole model behaved under stress. An insurer could assess policies under its own terms. An asset manager could review eligible investments and liquidity. A bank could measure direct exposures. A regulator could act only inside its jurisdiction. A government body could assess a particular scheme.

Yet the central question was relational: if insurance did not renew, would funds stop buying; if funds stopped buying, could assets be held to maturity; if a concentrated obligor weakened, would valuations, insurance and liquidity move together; and who had the authority and data to act before that feedback loop closed?

Receivables must exist before they can be financed safely

The first accountability gate is asset existence. For an ordinary approved invoice, a control should match a purchase order, evidence of delivery, supplier identity, buyer acceptance, invoice amount, currency and due date. Confirmation should come through a channel controlled independently of the originator and borrower. Amendments, cancellations, credit notes, duplicate invoices and related parties require explicit handling. A finance platform that merely records what a customer supplies has digitised a claim; it has not independently verified the economic event beneath it.

Prospective-receivables financing requires a different description. If a specific sale has not happened and an invoice does not exist, there is no present trade receivable to confirm. The financing resembles an unsecured exposure based on expected future business, perhaps informed by historical trading relationships or forecasts. Calling it supply-chain finance does not create a supplier, delivery or accepted obligation. The control package must therefore identify the legal borrower, permissible use, forecast assumptions, repayment source, covenants, concentration and loss-given-default without implying that a buyer has confirmed an invoice.

This is more than terminology because different labels produce different diligence. A trade-receivables investor may expect short duration, self-liquidation and diversified buyer exposure. A prospective exposure may depend on the borrower's enterprise cash flow and future orders. If the latter is presented with the language of the former, an investor can misunderstand both asset existence and correlation. The appropriate remedy is not to ban forecasting. It is to prevent a forecast from inheriting the assurance attributes of a confirmed receivable without the missing evidence being conspicuous.

The BEIS Committee report on Liberty Steel and the future of the UK steel industry examined how Greensill financing related to GFG Alliance companies and the steel businesses. It recorded uncertainty about debt and the significance of the financing relationship, drawing on evidence submitted to that inquiry. Its findings illuminate concentration and industrial continuity; they do not establish the validity or invalidity of every financed invoice, the liability of every company in the group, or a final creditor recovery.

An auditable control would assign a stable identifier to each financed asset and preserve its evidence from origination through collection. It would show whether the asset was an approved invoice, an unapproved invoice, a prospective receivable or a corporate loan. It would link the obligor's confirmation, insurance coverage, purchaser, vehicle, valuation, cash receipt and exception history. Any change in classification would retain the earlier version and named approval. That data lineage would let an investor, insurer, auditor or supervisor distinguish a failure of a buyer to pay from an asset that was never independently established.

Concentration can hide behind transaction volume

A platform may process many invoices and still depend economically on a small number of obligors or corporate groups. Transaction count is not diversification. Thousands of assets can share the same ultimate repayment source, be insured by the same carrier, be purchased by a small family of funds, or rely on one originator's servicing. Concentration should therefore be measured through several lenses: named obligor, connected group, sector, country, insurer, investment vehicle, maturity period and data source.

The Treasury Committee heard public officials describe contacts, the pandemic facilities and the reasons proposals were considered. The oral evidence from HM Treasury officials is valuable for reconstructing who communicated what and when. It is witness evidence given to Parliament rather than an adopted court record. Questions, propositions put by members and answers from witnesses retain different evidential status. The record supports scrutiny of process and record-keeping; it should not be converted into proof that any person acted corruptly or that a particular financing asset was false.

Concentration control begins before a limit. Legal-entity names should be resolved to ultimate groups; guarantees and common cash sources should be mapped; entities created for individual projects should not automatically count as independent risk. Exposure should include funded principal, approved but undrawn commitments, past-due balances and prospective assets. It should also show how insurance changes loss allocation without pretending that an insured exposure disappears. A limit exception should identify who approved it, why it remained consistent with mandate, what exit path existed and what evidence would trigger a stop.

Stress testing must join credit and funding. A concentrated obligor downgrade can increase expected loss and challenge insurance terms. Investors may redeem from funds or refuse new notes. The originator can then lose capacity to roll maturing assets while borrowers still rely on continuity. A useful test asks what happens if the largest connected group stops paying, one insurer declines renewal, asset sales stop and funds face redemptions at the same time. Separate mild scenarios miss the feedback mechanism that makes a concentrated model brittle.

The PRA censure of Wyelands Bank concerned a regulated bank owned by Sanjeev Gupta and established significant failings within the dates and legal scope stated by the PRA, including large exposures, governance, controls and risk management. It is relevant to the surrounding GFG financing ecosystem and change-in-control debate. It is not an enforcement finding against Greensill Capital, not a judgment that all Greensill-related financing was improper, and not a substitute for asset-level evidence.

Insurance was a dependency, not a guarantee of permanence

Credit insurance can transfer defined losses subject to a policy's wording, limits, exclusions and compliance conditions. It does not verify every receivable for every other entity, eliminate concentration, guarantee renewal or provide instant cash in all disputes. If a financing model requires continuing cover to place assets with investors, renewal risk is a funding risk even when every current policy remains valid. The model needs a plan for assets whose maturity extends beyond policy certainty and for a carrier that changes appetite.

The German bank was a separate regulated institution within the broader group. BaFin's official Greensill Bank moratorium FAQ explained the 3 March 2021 ban on disposals and payments, closure for customer business and deposit-protection process. That protective administrative action speaks to the bank and German legal framework. It does not determine claims in the UK administration, establish wrongdoing by every group director or show what every investment fund would recover.

Insurance governance should make four facts visible. First, which assets and perils are actually covered, after deductibles, limits and exclusions. Second, when coverage expires relative to asset maturity. Third, how much exposure depends on each insurer and policy. Fourth, what happens operationally if renewal is refused or a claim is disputed. A board should see the uninsured and renewal-at-risk positions, not only the headline insured amount. Investment committees should understand that insurance credit quality and legal enforceability become part of the asset risk.

Renewal should have staged triggers. Months before expiry, management should demonstrate replacement capacity or an orderly reduction in originations. If adequate cover is not contractually secured by a defined date, new assets dependent on it should stop. Existing assets should be mapped to cash and liquidity resources under conservative claim timing. Exceptions should not be justified by an assumption that longstanding commercial relationships will continue. A relationship is not committed capital and does not override policy terms.

The distinction also disciplines post-collapse accounts. Loss estimates can change because obligors pay, insurers accept or dispute claims, courts interpret contracts, assets are sold, or administrators incur costs. A gross face amount is not the same as a filed claim, an admitted claim, an estimated realisation or a cash distribution. Reports should preserve those categories so that a revised estimate is understood as new evidence rather than proof that the earlier figure was deceptive.

Fund valuation and liquidity needed one joined control

Funds that buy short-dated finance assets can appear liquid because instruments mature frequently. That appearance depends on the assets being real, paying on schedule and remaining saleable or replaceable. If valuation relies on expected collection and insurance, a dispute about asset eligibility or cover can change both price and liquidity. If investors redeem while uncertainty prevents reliable valuation, a manager may suspend dealing to protect equal treatment. Suspension is a fund-control decision, not by itself an adjudication that every asset is worthless.

The accountability chain belongs to both originator and purchaser. The originator must provide complete asset data, related-party mapping, performance history and exceptions. The asset manager must test mandate eligibility, independently challenge concentration, understand valuation inputs and plan liquidity without assuming continued origination. A depositary, custodian or administrator has its own defined duties, which should not be overstated into a universal guarantee of asset quality. An auditor tests financial statements under an applicable standard; it does not continuously insure the business model.

The FCA's final action involving GAM International Management and a former investment director established conflict-management and due-care failings described in the notices, including transactions linked to Greensill. Those findings show why potential incentives and conflicts must reach the right decision-makers. They do not adjudicate the later collapse of Greensill, determine the value of Credit Suisse funds, or prove misconduct in every transaction involving the originator.

An effective purchaser performs recurring tests rather than a one-time onboarding review. It samples obligor confirmations through an independent channel, reconciles cash to individual assets, compares promised and actual maturity, examines extensions and substitutions, maps connected groups, and reviews insurance exceptions. It watches for performance that is implausibly smooth, rapid growth without comparable control capacity, assets repeatedly rolled rather than paid, and concentrations concealed by special-purpose entities.

When evidence is unavailable, valuation and eligibility should become more conservative rather than treating opacity as immaterial.

The International Accounting Standards Board later issued supplier-finance disclosure amendments to IAS 7 and IFRS 7. The requirements seek transparency about arrangements, liabilities, cash flows and liquidity risk for reporting entities. They are a broad accounting reform, not a case-specific verdict and not a substitute for an originator's asset verification. Their significance is that financing presented near trade payables can affect liquidity analysis and should be visible in comparable disclosures.

A perimeter problem is also an ownership problem

Greensill Capital's stated core supply-chain lending was not generally regulated by the FCA. Certain regulated activity was undertaken through Greensill Capital Securities Limited as an appointed representative of an authorised principal, while the UK company also had anti-money-laundering registration implications. This divided scope matters. Being visible to a regulator for one purpose does not mean every product, balance-sheet risk or business-model dependency is prudentially supervised.

The FCA later described its position in responses following its 2022 Annual Public Meeting: it said investigations concerned Greensill Capital UK, Greensill Capital Securities and the appointed-representative oversight, while explaining limits arising because supply-chain finance was outside its conduct perimeter. An investigation is not a finding of breach. It records that matters were being examined under specified powers and cannot be used to infer guilt or predict a final outcome.

That outcome boundary became concrete in 2025. The FCA's closure statement on Mirabella Advisors said its review did not identify Mirabella breaches requiring further action and that the investigation had closed, while reserving the ability to review if new information emerged. This disposition must replace any earlier implication that an open inquiry proved principal-firm failure. It resolves the FCA's Mirabella investigation as described; it does not resolve other regulators' work, the UK-company investigation, insolvency claims or civil disputes.

The appointed-representative model assigns the authorised principal responsibility for regulated activity carried on by its representative. The control weakness across the market was that a small principal could host businesses much larger or more complex than its own operations. Oversight must therefore be proportionate to the representative, not to the principal's historic size. Before appointment, the principal needs a credible map of activities, customers, revenue, permissions, overseas links and financial resources. During the relationship, it needs data, skilled staff, testing access and an exit plan.

The FCA's confirmed 2022 rules strengthening appointed-representative oversight require enhanced principal oversight, risk assessment, annual review and more information for the regulator. Those rules are durable evidence of a changed control framework, although they do not move all commercial lending into regulation or prove how every principal performs in practice. Effectiveness should be tested through timely notifications, observed challenge, remediation, termination where necessary and fewer harms, not rule publication alone.

Public access and emergency schemes required auditable distance

Greensill's relationship with government raised two related but distinct questions. One concerned the development and use of supply-chain-finance schemes in public services. The other concerned lobbying during the pandemic for access or changes to emergency facilities. Public officials had to hear proposals quickly during a genuine crisis, but urgency increased the need for consistent channels, declared interests, preserved records and documented decisions. Access is not proof of preferential treatment; undocumented or informal access can nevertheless weaken confidence that competing proposals received equal scrutiny.

The Public Administration and Constitutional Affairs Committee interim report examined Lex Greensill's relationship with government and described gaps in the evidence available to it. The committee made governance findings and expressly defined limits to that phase of its inquiry. Those limits matter: a finding about unusual autonomy or access is not a finding that a contract was unlawful, that an official was bribed, or that a financial asset was invalid.

The Cabinet Office published Nigel Boardman's review of the development and use of supply-chain finance in government, including a factual report and recommendations. It provides an executive-commissioned reconstruction with access to government people and records. It is not a court inquiry, and its recommendations do not prove criminal or civil liability. Its practical value is control design: conflicts should be identified, appointments and access should be governed, commercial propositions should have named owners, and records should permit later scrutiny.

A strong public-sector gate separates policy interest from counterparty approval. An innovative mechanism may deserve examination without its promoter being entitled to a contract, guarantee or scheme change. Officials should log contacts, route substantive proposals through an institutional channel, identify economic interests, compare alternatives and state the authority for the decision. Former office should neither block a person from being heard nor grant a private evidential shortcut. The same record standard should apply to messages, calls and meetings when public resources are at stake.

Emergency lending requires an additional separation. The policy team defines objectives and risk appetite. An accrediting body assesses whether a lender can deliver within scheme rules. The lender conducts borrower diligence and retains obligations under the guarantee. Monitoring checks compliance after approval. If information suggesting concentration or noncompliance arises, it must reach the body capable of suspending or investigating guarantees. Each handoff should preserve what was known, the level of confidence and the response deadline.

Accreditation and monitoring must work as one system

The National Audit Office's investigation into the British Business Bank's accreditation of Greensill found that the Bank used a streamlined version of its established process under pandemic time pressure. The NAO said a less streamlined and more sceptical process might have prompted further questions about default rates, concentration, products, business model and ethical standards. It also credited monitoring with identifying loans allegedly in breach of rules. These are public-audit findings; alleged noncompliance remained contested, and the report did not adjudicate guarantee liability.

The lesson is not that emergency accreditation must wait for perfect information. It is that reduced pre-approval checking creates an explicit assurance debt. The decision record should identify which claims were accepted without independent verification, why speed justified that choice, what limits compensate for it, and when post-approval tests will close the gaps. A lender approved quickly might receive a lower initial cap, borrower-group limits, more frequent reporting and early sample audits. The control should tighten if growth or concentration outruns verified information.

Monitoring data must be designed before money flows. Lender and borrower identifiers should resolve connected groups. Every guaranteed facility should show approval date, amount, purpose, drawdown, group exposure, arrears and exceptions. Automated flags can identify multiple borrowers sharing ownership, address, directors, repayment sources or operating dependencies, but investigators must review context. When a rule may have been breached, the system should preserve the allegation, lender response, evidence, interim protection and final decision separately.

Public-money exposure also needs careful language. A guarantee ceiling is not a realised taxpayer loss. A loan issued is not automatically a valid guaranteed claim. A guarantee suspended during investigation is not necessarily cancelled. Recovery from borrowers can change net exposure. Reports should state the measurement date and distinguish original principal, outstanding principal, guarantee percentage, accrued amounts, claims made, claims accepted, cash recovered and net cost. Without those categories, an early worst-case figure can persist long after evidence changes.

The Treasury Committee's broader conclusion that ministers and officials acted with integrity in handling the CCFF lobbying can coexist with recommendations for stronger procedure. Accountability is not limited to finding bad faith. A system can reach the right decision while exposing weaknesses in record retention, informal contact, conflict visibility or cross-body information sharing. Treating improvement as an admission of corruption discourages institutions from learning; treating an absence of corruption as proof that procedure was adequate produces the same result.

Public-service continuity was not the same as company rescue

Greensill-linked services touched the public sector through pharmacy early-payment arrangements and an earned-wage product offered to NHS organisations. When a private provider fails, continuity planning should identify the public outcome that must continue, not assume that preserving the provider is the only route. Pharmacies still need predictable reimbursement and employees still need lawful payroll. Government can transition, replace or end a mechanism while protecting those functions.

The National Audit Office's investigation into supply-chain finance in the NHS examined the Pharmacy Earlier Payment Scheme and the Earnd salary-advance service. It described procurement, participation, benefits, costs and what happened after Greensill's failure. Its scope was these NHS arrangements, not Greensill's entire business, all lobbying or creditor losses. It provides public-value evidence and continuity lessons rather than an insolvency valuation or liability finding.

The Public Accounts Committee report on the pharmacy and salary-advance schemes scrutinised conflicts, risk assessment and the government's understanding of benefits. Committee recommendations are parliamentary accountability findings, not judicial determinations. They are most useful as a specification for future controls: document why a privately financed scheme is preferable to direct prompt payment, test supplier resilience, identify subcontractors, quantify actual take-up and benefits, and plan exit before dependency grows.

A public-service contract should contain a current service map, data portability, notice triggers, transition assistance, substitute financing options and an owner for continuity exercises. Free provision is not risk free. A zero-price service can create switching costs, data dependency, employee expectations or reputational risk. Procurement teams should record the provider's revenue model and incentive to continue. If a service depends on a parent company's balance sheet, insurer or funding market, financial-resilience monitoring should extend beyond the contracting subsidiary.

The continuity test is observable. On provider failure, are payments made on time, are employees' accrued wages protected, can users understand the transition, and can government reconstruct decisions and data? That test avoids overstating disruption merely because a brand disappears. It also prevents a successful transition from erasing weaknesses in onboarding. Both facts may be true: an arrangement ended without catastrophic interruption, and earlier due diligence or conflicts management still required improvement.

Administration, investigations and recoveries are different tracks

Greensill Capital (UK) Limited entered administration on 8 March 2021. Administration created a statutory process led by office-holders who sought to preserve or realise assets, investigate affairs, agree claims where appropriate and report to creditors. Their estimates evolve as assets are collected, litigation develops and costs accrue. Administration is not a single finding that validates every creditor claim or attributes fault to a director.

The company's Companies House filing history provides the authoritative public index for administrator proposals, statements and progress reports. It is the right source for dated filed documents, not a guarantee that every figure in an office-holder report will be realised. Estimated outcomes, contingent claims and recoveries should always retain the reporting period and basis. Later reports may supersede earlier estimates without rewriting what was reasonably known at the time.

The Insolvency Service's June 2026 statement on Lex Greensill's disqualification undertaking says he agreed to a nine-year ban after signing a legally binding undertaking and did not dispute specified facts for purposes of those proceedings. The statement describes conduct involving Katerra-related notes, removal of protections and use of $440 million. Its findings and admissions have that defined disqualification context; they should not be expanded into a judgment on every Greensill asset, creditor claim, director or group transaction.

The Serious Fraud Office maintains an open GFG Alliance case page concerning suspected fraud, fraudulent trading and money laundering, including financing arrangements with Greensill. An open criminal investigation establishes neither charge nor guilt. It must be reported in the present tense and kept separate from parliamentary descriptions, PRA action against Wyelands, civil claims and insolvency recovery. The presumption of innocence and later procedural updates control any account of individual responsibility.

Recovery reporting should use a waterfall. Start with cash and assets controlled by the estate. Show realisations by asset family, disputed or contingent claims, secured claims and costs. Then show admitted unsecured claims and distributions actually paid. Avoid adding creditor claims, investor fund exposure, insured amounts, government guarantees and group-company debts into one supposed loss total; they can overlap, sit in different estates and change with collection. A transparent schedule explains movement from the previous period and identifies which figures remain estimates.

This separation also improves remedy. Investors may claim against a fund or manager under one legal relationship. Banks and insurers may have contractual disputes under another. Employees have statutory and contractual rights. Government bodies may assert scheme or guarantee rights. Creditors lodge claims in particular estates. A headline about “Greensill losses” cannot establish who has standing, priority or a recoverable amount. The responsible account follows the legal entity and claim path.

Governance must connect data to stop authority

A board cannot govern a complex finance model from aggregate originations and revenue alone. Its risk dashboard should show the composition of financed assets, including confirmed invoices and prospective exposures; the largest connected obligors; insurance expiry ladders; purchasers and funding channels; overdue and extended assets; disputed confirmations; and cash that has not reconciled on time. The data should be measured under documented definitions that remain stable from one meeting to the next. If a definition changes, the board should receive the old and new results with a reconciliation rather than a smoother historical series.

Board challenge also needs a route from exceptions to action. A rising number of unconfirmed assets should pause new financing for the affected originator or obligor. A connected-group limit breach should require approval outside the commercial team and a dated reduction plan. Insurance due to expire without committed replacement should reduce eligible tenors or stop dependent originations. A fund purchaser that changes eligibility or concentration requirements should trigger a portfolio review, not a search for a different purchaser with less scrutiny. Every trigger needs an owner, response deadline and evidence required to reopen activity.

Independence is practical rather than ceremonial. Risk staff need direct access to asset records, counterparties and the board committee; they must not depend on a sales summary to test a sales claim. Internal audit should be able to select samples from the complete population and obtain confirmation from external parties. Model validation should examine how missing or stale data are treated. Compliance should map each legal entity and activity to permissions instead of treating group association with an authorised firm as a general regulatory status.

Remuneration should not reward volume that later fails confirmation or remains financed through repeated extensions.

The same architecture must extend to outsourced technology and servicing. A platform may automate onboarding, invoice matching, eligibility and allocation, but automation can reproduce a wrong classification at scale. Access controls should separate creation, approval, amendment and write-off. Logs should record the source of buyer confirmations and protect them from silent replacement. Business-continuity plans should let purchasers and administrators export asset and payment histories in a usable format if the originator stops operating.

A dark platform cannot become the reason that legal ownership, payment instructions or policy coverage cannot be reconstructed.

Assurance should use outcome samples. Reviewers can select matured assets and trace them backward from cash receipt to the legal obligor, invoice evidence, insurance and investor allocation. They can select outstanding assets and obtain fresh external confirmation. They can compare stated due dates with actual collection and identify extensions that make a short-duration portfolio behave like longer-term corporate credit. They can examine limit exceptions to see whether promised reductions occurred. These tests show whether controls work under normal pressure, rather than merely confirming that policies contain the right words.

Finally, governance needs a credible wind-down mode. Growth-stage assumptions often treat continuing origination as the normal source of fee income and portfolio replacement. A wind-down plan should model zero new business, preserve servicing staff and systems, identify who can collect each asset, maintain insurance claims, communicate with purchasers and protect client or vehicle money. Funding providers should know the triggers for activating that plan. Practising the plan before distress is a stronger signal of resilience than a board statement that diversified relationships will always supply replacement capital.

Evidence from those exercises should reach the board as exceptions, not just completion rates. A failed data export, an unreachable obligor, an ambiguous payment instruction or a policy document that cannot be matched to assets is a control failure requiring repair. Repeating the exercise after remediation creates proof that the problem was closed. Without that second test, management has recorded an activity but not established resilience.

What durable supervision must prove

Greensill is often described as a regulatory-perimeter failure, but simply expanding one regulator's jurisdiction would not answer every weakness. Commercial lending, asset management, insurance, banking, accounting, public procurement and insolvency involve different objectives and legal powers. The stronger response is a system that gives each risk a named owner and makes cross-boundary dependencies visible soon enough for coordinated action.

At origination, durable proof means sampled independent confirmations, stable asset identifiers, explicit labels for prospective assets, related-party mapping and cash reconciliation. At portfolio level, it means connected-group, insurer and purchaser concentrations with binding limits and recorded exceptions. At funding level, it means stress tests joining obligor default, insurance non-renewal and investor redemption. At distribution, it means mandate testing, conflict escalation, valuation challenge and liquidity gates that do not depend on uninterrupted new sales.

For regulators, proof includes accurate perimeter maps, information-sharing triggers and principals with capacity to oversee representatives. The closure of one investigation must be reflected accurately; reform cannot be justified by implying a breach the regulator did not find. At the same time, a no-action outcome for one firm does not erase market-wide evidence supporting stronger oversight. The test is whether supervisors now receive timely information about representative scale, activity and risk, and whether they intervene before an arrangement exceeds the principal's control capacity.

For government, proof is a retrievable decision trail. Contacts are logged, interests declared, substantive requests routed formally, alternatives compared and records retained. Emergency accreditation states its assurance gaps and compensating limits. Monitoring resolves connected borrowers and moves concerns to the guarantor. Public-service contracts demonstrate exit and data portability. These controls protect both public money and fair access without treating every conversation with a former office-holder as improper.

For creditors and the public, proof is versioned recovery reporting and disciplined legal language. Administrator estimates are dated; allegations are attributed; investigations are not convictions; committee findings are not court judgments; regulatory actions stay within their named respondents and periods. Gross exposure, admitted claims, cash realisations and distributions are not collapsed into one number. That precision is not caution for its own sake. It prevents accountability from being built on figures or conclusions that later evidence cannot support.

The lasting test is whether comparable models can now withstand the removal of one assumption. If an insurer refuses renewal, assets should remain identified and an orderly funding plan should exist. If a large obligor weakens, connected exposures should be visible. If investors redeem, valuation and liquidity controls should protect equal treatment. If a provider fails, public services should transition. If a supervisor lacks jurisdiction, another named authority or an explicit policy decision should own the gap. Confidence comes from those demonstrated controls, not from the claim that the exact Greensill structure will never recur.