Financial Forensics: The Due Diligence Files

Every risk committee has a version of the same conversation. A desk is generating exceptional returns. The risk officer asks about the control environment. Nobody asks the question that would end the conversation: can we independently verify that the positions the desk reports are the positions the desk holds?

🔴 Every corporate failure leaves behind a pattern.

FFL Risk Pattern Scan provides access to a searchable library of documented corporate collapses, frauds and restructurings that can be filtered by geography, sector, collapse mechanism and fraud vector.

Compare live opportunities against historical cases using pattern matching and risk assessment tools designed for investors, lenders and deal teams.

All analysis runs locally and remains private.

⁠⁠⁠⁠⁠https://risk-pattern-scan.lovable.app/⁠⁠

At Barings, the answer was no. For three years. The 1994 internal audit had explicitly flagged the dual role as a material control weakness and recommended separation. The recommendation was not implemented. The funding requests from Singapore were inconsistent with a matched arbitrage strategy — matched arbitrage does not require one-way external margin. The profit anomaly and the control weakness appeared in the same audit document. Nobody combined them into a single finding requiring immediate resolution.

This episode applies the Barings mechanism to current fund and trading operation risk frameworks: the three trip wires for any market-neutral strategy, the independent position verification standard, the SIMEX margin concentration as a counterparty signal, and the three questions that would have ended the position at any point between 1993 and 1995 — none of which required a derivatives specialist to ask.

The revenue was real. The position generating the other side of the ledger was in account 88888. Every collapse has a pattern. We dissect it. Layer by layer.

In 1992, Nick Leeson was the most profitable trader at Barings Bank. In 1995, he was the reason Barings Bank no longer existed. The same person. The same desk. The same Singapore office. Three years apart.

What changed was not Leeson. What changed was the size of the position hidden in account 88888 — an internal error account he had the authority to use because Barings had given him an unusual mandate: run both the front office trading desk and the back office settlement operations. The person executing the trades and the person reconciling the trades were the same person.

🔴 Every corporate failure leaves behind a pattern.

FFL Risk Pattern Scan provides access to a searchable library of documented corporate collapses, frauds and restructurings that can be filtered by geography, sector, collapse mechanism and fraud vector.

Compare live opportunities against historical cases using pattern matching and risk assessment tools designed for investors, lenders and deal teams.

All analysis runs locally and remains private.

⁠⁠⁠⁠⁠https://risk-pattern-scan.lovable.app/⁠⁠

That single structural decision, made for cost efficiency reasons and documented in internal memos, was the condition that made everything that followed possible. By February 1995, account 88888 held losses of $1.4 billion. Barings' total capital was $900 million. The bank was sold to ING for one pound.

This episode dissects the mechanism layer by layer: the origin of the account in a 1992 junior trader error, the Kobe earthquake and the collapse of the short volatility position, the three signals that were in the internal record and never combined into a single finding, and what Peter Baring told the Bank of England after the collapse.

Every collapse has a pattern. We dissect it. Layer by layer.

The structure that allowed BCCI to operate for 19 years across 78 jurisdictions is not unusual. A holding company in a low-disclosure jurisdiction. Operating entities across multiple countries. Audit mandates divided by geography. Ownership partially obscured by nominee layers. Each piece reviewed by a different regulator. No single authority with a consolidated view.

🔴 Every corporate failure leaves behind a pattern.

FFL Risk Pattern Scan provides access to a searchable library of documented corporate collapses, frauds and restructurings that can be filtered by geography, sector, collapse mechanism and fraud vector.

Compare live opportunities against historical cases using pattern matching and risk assessment tools designed for investors, lenders and deal teams.

All analysis runs locally and remains private.

⁠⁠⁠⁠⁠https://risk-pattern-scan.lovable.app/⁠⁠

That is standard architecture for international capital deployment. The question a GP or LP almost never asks — and the question BCCI made permanently relevant — is whether the regulatory opinion from any single jurisdiction is an opinion on the institution, or only an opinion on the piece that jurisdiction can see.

This episode applies the BCCI mechanism to current institutional due diligence practice: the three-part consolidated counterparty framework, the Price Waterhouse qualified opinion as an actionable signal, the 1990 guilty plea as a counterparty risk event, and the compliance calculation that kept 35 institutions in a relationship with a criminal bank because each reviewed its own piece and found it acceptable.

The piece the regulators could not see was $13 billion. The split audit mandate — two firms, neither with global visibility — was disclosed in the annual reports. The regulatory structure — supervised in pieces across seventy-eight jurisdictions — was a matter of public record. The correspondent banking relationships that gave BCCI access to the global dollar clearing system were visible to any institution that traced the ownership chain. Three signals. The structure itself was the signal. This episode dissects the BCCI jurisdictional arbitrage mechanism, the consolidated supervision gap architecture, and the three institutional signals that indicated a global bank operating without consolidated oversight before $13 billion disappeared across three decades. GP/LP analysis. Global banking risk. Correspondent banking. Consolidated supervision. Counterparty jurisdiction risk. Financial Forensics Labs — GP/LP Analysis. Every collapse has a pattern. We dissect it. Layer by layer. 

In 1972, Agha Hasan Abedi incorporated a bank in Luxembourg, headquartered in London, with a holding company in the Cayman Islands, a majority shareholder from Abu Dhabi, and a split audit mandate across two firms that never shared consolidated access. By 1991, that structure had allowed the largest criminal bank in history to operate across 78 countries for 19 years — not by corrupting regulators, but by designing a corporate architecture that made consolidated supervision impossible.

🔴 Every corporate failure leaves behind a pattern.

FFL Risk Pattern Scan provides access to a searchable library of documented corporate collapses, frauds and restructurings that can be filtered by geography, sector, collapse mechanism and fraud vector.

Compare live opportunities against historical cases using pattern matching and risk assessment tools designed for investors, lenders and deal teams.

All analysis runs locally and remains private.

⁠⁠⁠⁠⁠https://risk-pattern-scan.lovable.app/⁠⁠

When the Bank of England directed the coordinated closure on July 5th, 1991, liquidators found $13 billion missing from a $20 billion deposit base. The books were a fiction. The loans were to nominees. The treasury losses had been reclassified for years. The auditors had each seen their piece. Nobody had seen the institution.

This episode dissects the BCCI mechanism layer by layer: the split audit design, the nominee acquisition of US banks, the ICIC shadow ledger, and the three signals that were in the public record before the collapse — including the 1990 criminal guilty plea that every correspondent bank processed and ignored.

he Bank of Credit and Commerce International operated in seventy-eight countries. It was audited by two different firms — each responsible for different parts of the global operation, neither with visibility over the whole. It was supervised by the Bank of England, the Luxembourg banking authority, and the central banks of every country it operated in — none of which had consolidated oversight. When it was shut down in 1991, $13 billion had disappeared across three decades of operation. The closure was coordinated across sixty countries in a single day. This episode dissects the BCCI collapse, the jurisdictional arbitrage mechanism, and the consolidated supervision gap that allowed a global bank to operate without any single regulator seeing the complete picture. BCCI. Bank of Credit and Commerce International. Banking fraud. Jurisdictional arbitrage. Bank of England. Global banking collapse. $13 billion fraud. Financial Forensics Labs — Every collapse has a pattern. We dissect it. Layer by layer.

The transaction structure was disclosed to tax authorities at the individual transaction level. The aggregate pattern — multiple refund claims on a single tax payment — was only visible by combining data across jurisdictions. The legal opinions supporting the trades came from the same law firms earning fees on the transaction volume. Three signals that required combining data sources nobody was combining. 

🔴 Every corporate failure leaves behind a pattern.

FFL Risk Pattern Scan provides access to a searchable library of documented corporate collapses, frauds and restructurings that can be filtered by geography, sector, collapse mechanism and fraud vector.

Compare live opportunities against historical cases using pattern matching and risk assessment tools designed for investors, lenders and deal teams.

All analysis runs locally and remains private.

⁠⁠⁠⁠⁠https://risk-pattern-scan.lovable.app/⁠⁠

This episode dissects the Cum-Ex legal arbitrage mechanism, the permissibility gap between legal structure and legal authorization, and the three systemic signals that indicated a €55 billion extraction scheme operating inside the defended perimeter of individual transaction legality. GP/LP analysis. Tax risk. Legal arbitrage detection. European financial regulation. Banking criminal liability. Financial Forensics Labs — GP/LP Analysis. Every collapse has a pattern. We dissect it. Layer by layer. 

€55 billion extracted from six European treasuries by 35 institutions — each holding a legal opinion that said the transaction was permissible. That is the Cum-Ex case. And the distinction at the center of it is the one most institutional compliance frameworks never formalize: the difference between a transaction that is permissible because the law affirmatively authorizes it, and one that is permissible because the law as written does not explicitly prohibit it. That gap — between authorization and the absence of prohibition — is where regulatory arbitrages live. It is also where they die, when the regulatory authority decides the intent of the rule is clear even if the text is ambiguous. We dissect the dividend withholding extraction mechanism, the compliance architecture that reviewed each transaction in isolation while the aggregate grew invisible across six jurisdictions, and the legal standard that sent the lawyers who wrote the opinions to prison. The practical framework: how to identify whether a yield strategy is authorized or merely not-yet-prohibited — and what the difference means for the institution holding the position when the gap closes.

The Cum-Ex scheme was simple in principle: a stock was traded rapidly around the dividend record date between multiple parties in a way that made it impossible for the tax authority to determine who actually owned the shares on the relevant day. Multiple parties filed refund claims for a withholding tax that had been paid only once. Thirty-five banks across six European countries participated. The total extraction from European public treasuries was approximately €55 billion. Each individual transaction was structured to appear legally defensible

🔴 Every corporate failure leaves behind a pattern.

FFL Risk Pattern Scan provides access to a searchable library of documented corporate collapses, frauds and restructurings that can be filtered by geography, sector, collapse mechanism and fraud vector.

Compare live opportunities against historical cases using pattern matching and risk assessment tools designed for investors, lenders and deal teams.

All analysis runs locally and remains private.

⁠⁠⁠⁠⁠https://risk-pattern-scan.lovable.app/⁠⁠

. This episode dissects the Cum-Ex dividend fraud, the legal arbitrage mechanism, and the systemic tax extraction architecture that turned a technical ambiguity in dividend withholding rules into the largest tax fraud in European history. Cum-Ex. Dividend stripping. Tax fraud. Germany. European banking. Withholding tax refund fraud. Legal arbitrage. Bank criminal liability. Financial Forensics Labs — Every collapse has a pattern. We dissect it. Layer by layer.

€55 billion. Six European treasuries. A decade. No forged documents, no hacked systems, no corrupted officials. Just a stock trade timed around the dividend record date — structured so that two parties simultaneously appeared as the registered owner of the same share, and both claimed a refund of the withholding tax that had only been paid once. The German tax authority kept processing the claims. The traders kept filing. The legal opinions kept saying it was permissible. By the time the gap was closed, it had become the largest theft of public money in European history — executed by 35 major banks, reviewed by the most prestigious law firms in Europe, and invisible to every regulator until investigative journalists forced a parliamentary inquiry. We dissect the Cum-Ex mechanism layer by layer: how a settlement timing window became a €55B extraction, why the structure was individually defensible and systemically indefensible, and what the courts decided when they finally looked at the aggregate.

Swedbank's Estonian branch had a functioning compliance department. It identified the risk, wrote the reports, and escalated correctly — for eight years. The override came from a revenue target, not a corrupt officer. That is the mechanism: compliance capture through economic dependence, where the cost of acting on the recommendation falls on a business unit whose performance is measured by the revenue the problem client generates. We dissect the institutional decision architecture — the specific incentive structure that makes the fee income always win over the regulatory cost until the regulatory cost arrives all at once — and the active version of the same structure running in non-bank financial intermediaries right now.v

🔴 Every corporate failure leaves behind a pattern.

FFL Risk Pattern Scan provides access to a searchable library of documented corporate collapses, frauds and restructurings that can be filtered by geography, sector, collapse mechanism and fraud vector.

Compare live opportunities against historical cases using pattern matching and risk assessment tools designed for investors, lenders and deal teams.

All analysis runs locally and remains private.

⁠⁠⁠⁠⁠https://risk-pattern-scan.lovable.app/⁠⁠

The Baltic branches were the highest-margin segment of Swedbank's retail network. The compliance reports were filed and archived. The correspondent banking relationships that processed the flows had standard KYC documentation. Three signals — all visible inside the institution. None acted on because the revenue protected the exposure from the question. This episode dissects the Swedbank compliance capture mechanism, the correspondent banking incentive failure architecture, and the three institutional signals that indicate when a compliance function has lost independence from the revenue it is supposed to monitor. GP/LP analysis. AML compliance risk. Correspondent banking due diligence. Compliance independence. Financial crime risk. Financial Forensics Labs — GP/LP Analysis. Every collapse has a pattern. We dissect it. Layer by layer.

A Swedish retail bank branch in Tallinn with fewer than 200 employees processed €200 billion in suspicious transactions over eight years. The compliance function flagged it. Escalated it. Documented management's response. The accounts stayed open because closing them would have created a revenue gap the branch manager would have had to explain to Stockholm.

🔴 Every corporate failure leaves behind a pattern.

FFL Risk Pattern Scan provides access to a searchable library of documented corporate collapses, frauds and restructurings that can be filtered by geography, sector, collapse mechanism and fraud vector.

Compare live opportunities against historical cases using pattern matching and risk assessment tools designed for investors, lenders and deal teams.

All analysis runs locally and remains private.

⁠⁠⁠⁠⁠https://risk-pattern-scan.lovable.app/⁠⁠

We dissect the correspondent banking mechanism — how fee income captures a compliance function without corrupting it, how the calculation to keep the accounts open is always wrong and always made anyway, and why the CEO learned the board had endorsed her performance two hours before she was fired. Between 2007 and 2015, approximately €200 billion in suspicious transactions passed through Swedbank's Baltic branches. The compliance function wrote the internal reports. Management overrode the escalations. The CEO was terminated the same day a documentary about the scheme aired on Swedish television. The mechanism was not that compliance failed to see the risk. It was that compliance had no authority to stop the revenue. This episode dissects the Swedbank Baltic money laundering case, the compliance capture through revenue dependence mechanism, and the correspondent banking incentive failure that made €200 billion in suspicious flows invisible to the institutions that were supposed to stop them. Swedbank. Baltic money laundering. Estonia. AML failure. Correspondent banking. Money laundering scandal. CEO fired. Financial Forensics Labs — Every collapse has a pattern. We dissect it. Layer by layer.

A bilateral investment treaty is not a diplomatic document. It is an irrevocable option the state sells to every covered investor — the right to sue when policy moves against them, in a tribunal no domestic court can override, with an award enforceable in 169 countries. Romania signed one in 1994. EU accession required a policy change the treaty prohibited. The collision produced a $250M award that EU law cannot extinguish outside its own borders.

🔴 Every corporate failure leaves behind a pattern.

FFL Risk Pattern Scan provides access to a searchable library of documented corporate collapses, frauds and restructurings that can be filtered by geography, sector, collapse mechanism and fraud vector.

Compare live opportunities against historical cases using pattern matching and risk assessment tools designed for investors, lenders and deal teams.

All analysis runs locally and remains private.

⁠⁠⁠⁠⁠https://risk-pattern-scan.lovable.app/⁠⁠

We dissect the mechanism — how BITs create sovereign exposure that outlasts every political transformation, how the enforcement architecture travels globally under the New York Convention, and what any institution with cross-border investment exposure needs to model before the next policy change triggers the next claim. The BIT obligation and the EU accession requirement were in direct conflict from the moment Romania joined. The arbitration award was public. The EU's prohibition on payment was a formal decision. The US enforcement order was in the federal court record. Three documents. Three jurisdictions. One state caught between obligations it cannot simultaneously honor. This episode dissects the ISDS BIT optionality mechanism, the treaty obligation permanence architecture, and the three jurisdictional signals that define the enforcement gap when a sovereign owes an arbitration award it is legally prohibited from paying. GP/LP analysis. Cross-border investment risk. Treaty arbitration. Sovereign enforcement. EU investment law. Financial Forensics Labs — GP/LP Analysis. Every collapse has a pattern. We dissect it. Layer by layer.

Romania guaranteed investment incentives in writing. An investor structured an entire operation around them. The EU required Romania to cancel them. The investor filed for arbitration under a 1994 bilateral investment treaty. The award: $250 million plus daily compounding interest. Romania has spent over a decade fighting enforcement in US courts, UK courts, Belgium, France, and Luxembourg — and losing. We dissect the ISDS mechanism: how a treaty creates promises that survive every political transformation around them, and why the Romanian state cannot escape an obligation it had every institutional reason to see coming.

🔴 Every corporate failure leaves behind a pattern.

FFL Risk Pattern Scan provides access to a searchable library of documented corporate collapses, frauds and restructurings that can be filtered by geography, sector, collapse mechanism and fraud vector.

Compare live opportunities against historical cases using pattern matching and risk assessment tools designed for investors, lenders and deal teams.

All analysis runs locally and remains private.

⁠⁠⁠⁠⁠https://risk-pattern-scan.lovable.app/⁠⁠

 Romania signed a bilateral investment treaty with Sweden in 2002. It provided state aid to attract foreign investors. When Romania joined the EU in 2007, EU rules required it to abolish the state aid — breaking the treaty. The Micula brothers, Swedish-Romanian investors who had relied on that aid, filed an arbitration claim. They won $250 million. The EU said Romania could not pay it. The US courts said it had to. The dispute is ongoing. This episode dissects the ISDS Micula v. Romania case, the bilateral investment treaty mechanics, and the sovereign policy risk that emerges when treaty obligations conflict with supranational legal requirements. ISDS. Bilateral investment treaty. Micula v Romania. EU state aid rules. Investment arbitration. Sovereign policy risk. Cross-border enforcement. Financial Forensics Labs —