The market was enormous. The missing controls were often very small.
Foreign exchange is not a niche corner of finance. The Bank for International Settlements measured average over-the-counter FX turnover at $9.6 trillion per day in April 2025. That scale creates deep liquidity and professional infrastructure, but it also creates a dangerous assumption: that a market this large must automatically be transparent. The cases in this investigation show the opposite. A global benchmark could be influenced inside a private chat room; a customer account could be emptied while forged bank confirmations passed through one mailbox; a glossy “robot” could display profits without placing the trades it claimed to execute.
The word “Forex scandal” therefore covers several different failures. Some cases were criminal frauds. Some were regulatory findings about undisclosed conflicts. One involved a conviction later vacated on appeal. The Swiss-franc shock was not fraud at all, but a market discontinuity that exposed how quickly capital and liquidity can disappear. Keeping those categories separate is not legal caution for its own sake. It is the central lesson: a trader cannot evaluate protection until the actual failure mode is identified.
This report follows two records through ten cases. The public record is what customers were shown—segregated money, audited accounts, no dealing desk, stable returns or automated execution. The control record is who could move cash, prepare statements, confirm bank balances, route orders and change the customer’s ability to exit. The distance between those records is where the loss accumulated. Fines, restitution, missing customer money and transaction notional are not added into one sensational total because they measure different things.
How to read the legal record
Established
A guilty plea, conviction, final civil order or formal regulator finding supports the description.
Changed outcome
An earlier conviction or theory was later altered on appeal; the latest legal status takes priority.
Contrast case
A failure is included to explain market and capital risk even though no fraud finding is asserted.
Fraud, misconduct and failure are not the same
This report separates criminal convictions, regulatory findings, civil allegations and insolvency. A broker collapse is not automatically fraud, and a charge is not a conviction. Figures are not directly comparable: some are money raised, some missing customer funds, some fines or court judgments, and one is the size of a client transaction.
Ten cases at a glance
A documentary map, not a ranking by loss. Historical amounts are rounded.
| Case | Category | Reported scale | Legal status |
|---|---|---|---|
| FX benchmark cartel | Bank benchmark manipulation | >$2.5B | Parent-level guilty pleas |
| Refco | Accounting fraud · broker collapse | ≤ $2.4B | Guilty plea · 16-year sentence |
| PFGBEST | Theft of customer funds | $215.5M | Guilty plea · 50-year sentence |
| Trevor Cook / Oxford | Managed-FX Ponzi scheme | $194M | Criminal convictions · 25 years for Cook |
| Black Diamond | Fictitious Forex platform · Ponzi | $35M | Court orders · permanent bans |
| FXCM U.S. | Misleading execution claims | $7M | CFTC order · U.S. withdrawal |
| Capital Blu | Forex fund fraud | $17M | Jury verdict · court judgment |
| RoFX | Fake automated Forex trading | >$225M | Default judgment · court findings |
| HSBC / Mark Johnson | Client FX order · changing legal outcome | $3.5B | 2017 conviction vacated in 2025 |
| EUR/CHF / Alpari UK | Broker insolvency · not fraud | EUR/CHF | Special administration · no fraud finding |
Three decades of hidden deficits, rigged fixes and courtroom reversals
The scandals did not form one continuous scheme. They reveal how the same absence of independent verification resurfaced inside banks, brokers, funds and online platforms.
The deficits begin
Russell Wasendorf later admitted stealing from PFG customer accounts from the early 1990s. Refco’s concealed related-party debt dated from at least 1998.
Refco collapses
Two months after its public offering, hidden debt is disclosed and a major broker enters bankruptcy. Trevor Cook’s currency programme begins attracting investors.
Three new fault lines
The FX benchmark cartel period begins. Black Diamond solicits for a fictitious platform, while Capital Blu reports gains despite trading losses.
Smooth returns break
Trevor Cook’s operation collapses; court cases expose fabricated statements, lulling payments and money routed away from trading.
PFGBEST’s mailbox fails
A failed suicide attempt and confession expose a customer-account shortfall hidden through forged statements and intercepted confirmations.
Cartel pleas and the franc shock
Four banking groups agree to FX-market guilty pleas. In the same year, the Swiss National Bank removes the EUR/CHF floor and Alpari UK becomes insolvent.
Execution claims meet the order book
FXCM’s U.S. entities settle CFTC findings over undisclosed interests behind the promoted No Dealing Desk model.
The robot era
RoFX uses a polished website, internal balances and claimed automation to collect at least $57.5 million, according to later court findings.
RoFX judgments
A federal court orders more than $56 million in restitution and roughly $169 million in civil penalties against nine defendants.
The record is corrected
An appeals court vacates Mark Johnson’s 2017 conviction. The broader lesson remains current: legal status and broker controls must be checked, not remembered from old headlines.
Court record
Documentary photography · source and licence in each caption
Three systems, three places where trust replaced verification
The diagrams reduce complex cases to the control point that mattered most. They do not imply that a bank cartel, customer-fund theft and online Ponzi operation were legally identical.
The FX fix: how a minute of coordination could touch a global benchmark
A benchmark fix is supposed to compress observable market activity into a reference rate. Asset managers and corporations can use that rate to value portfolios or meet large currency-conversion needs without negotiating every price individually. The design creates a predictable window: dealers know when customer demand will be aggregated and when trading will influence the calculation. That window is useful for legitimate execution, but it is also sensitive to coordination.
The Justice Department said euro-dollar traders at Citicorp, JPMorgan, Barclays and RBS used an exclusive chat room and coded language to share order information and coordinate around the 1:15 p.m. ECB and 4:00 p.m. WM/Reuters fixes. They could align trading, withhold bids or offers and protect one another’s positions. The problem was not that one dealer had a market view. It was that supposed competitors exchanged information and restrained competition at a price-setting moment used beyond their own desks.
The four banking groups agreed in 2015 to parent-level guilty pleas and more than $2.5 billion in criminal fines. Related U.S. and international actions took penalties for the broader conduct close to $9 billion. Scale did not prevent manipulation; it made a small distortion consequential. The practical defence is institutional rather than personal: transaction-cost analysis, independent benchmarks, time-stamped execution records and policies that separate customer information from proprietary risk-taking.
Control failure: competitors exchanged confidential order information at the very moment the market relied on them for price formation.
Case files · FX benchmark cartel ↓PFGBEST: segregation on paper, one-man control in practice
Customer segregation is often described as though it were a physical vault. In reality it is a legal and accounting system: money must sit in designated accounts, the broker must calculate what it owes customers, banks must confirm balances and regulators must receive accurate reports. Each link matters. At PFGBEST, the public rule existed, but one executive controlled too many links in the evidence chain.
Russell Wasendorf Sr. admitted secretly withdrawing customer money, forging bank statements, submitting false reports and using a post-office box to intercept verification forms intended for the bank. The same person could create the false balance and control the channel through which that balance was supposedly confirmed. The fraud continued for almost two decades because the verification was not truly independent. A control that asks the subject of an audit to deliver the auditor’s confirmation is not a control.
Wasendorf admitted stealing more than $215 million from over 13,000 victims and received a 50-year sentence, a $100 million forfeiture judgment and restitution exceeding $215.5 million. The case changed regulatory practice, but its lesson is broader than one jurisdiction. “Segregated” must be followed by operational questions: which bank holds the funds, who receives confirmations, whether balances are reconciled directly and whether any founder can override the process.
Control failure: the executive who created the shortfall also controlled the evidence that was supposed to reveal it.
Case files · PFGBEST ↓RoFX: when a dashboard performed the role of a broker statement
RoFX looked familiar to a retail trader: account tiers, an automated strategy, smooth performance and a dashboard that translated technical complexity into a reassuring balance. Automation was central to the pitch because it moved the burden of proof away from identifiable human decisions. Customers did not need to see an investment committee, a dealing desk or a verifiable order history; the “robot” was presented as the explanation for the returns.
The 2024 CFTC judgment found that the enterprise accepted at least $57.5 million from U.S. and international customers, did not conduct the claimed Forex trading and sent money to offshore entities and personal accounts. An internal dashboard could therefore show profitable trading because it was not constrained by an external broker ledger. As long as withdrawals and new deposits continued, the screen functioned as social proof and accounting record at the same time.
The court ordered more than $56 million in restitution and a civil monetary penalty above $169 million against the nine defaulting defendants, with separate sanctions for another participant. Those numbers must not be mistaken for cash already recovered. A judgment establishes liability; payment depends on locating assets. For a customer, the decisive pre-deposit question is whether the balance can be matched to an account at a regulated executing broker—not whether the web interface looks technically sophisticated.
Control failure: the promoter’s interface was treated as proof of both trading and custody, although neither was independently visible.
Case files · RoFX ↓FX benchmark cartel
How it worked
Traders at major banks used private chat rooms to exchange order information and coordinate activity around important currency benchmarks. The conduct blurred the line between executing client business and moving the market for the banks’ own advantage.
What happened next
In 2015 Citicorp, JPMorgan, Barclays and Royal Bank of Scotland agreed to plead guilty to conspiring to manipulate U.S. dollar and euro prices in the spot FX market. Their criminal fines exceeded $2.5 billion; other regulatory settlements made the wider scandal still larger.
What traders should learn
A liquid market can still be manipulated around a benchmark. Large institutions, regulation and familiar names do not remove conflicts of interest.
Refco
How it worked
Refco was a major futures and commodities broker. Its chief executive used transactions between related entities to hide hundreds of millions of dollars in debt, presenting a stronger balance sheet before the company’s public offering.
What happened next
The hidden exposure surfaced in 2005 and Refco entered bankruptcy. Phillip Bennett pleaded guilty and received a 16-year sentence; the court ordered forfeiture of assets up to $2.4 billion in the related criminal prosecution. That order was separate from the SEC's civil consent resolution and does not establish the amount ultimately recovered.
What traders should learn
A recognised broker can fail because of risks outside the trading platform. Audited accounts, related-party exposure and control over cash matter.
PFGBEST
How it worked
PFGBEST appeared to be an established U.S. futures and forex broker. Russell Wasendorf Sr. intercepted verification mail, forged bank documents and concealed a customer-account shortfall for almost two decades.
What happened next
After the fraud was exposed in 2012, Wasendorf admitted stealing more than $215 million from over 13,000 customers. A federal court sentenced him to 50 years and ordered about $215.5 million in restitution. An order to pay is not evidence that victims have already recovered that amount.
What traders should learn
Segregation is meaningful only when balances are independently confirmed. One executive must never control the records, mail and bank verification chain.
Trevor Cook / Oxford
How it worked
Trevor Cook and associates sold a foreign-currency programme promising double-digit returns with little or no risk. Investors were told their money was segregated and liquid, while fabricated statements and payments created the appearance of success.
What happened next
The group collected about $194 million. Court evidence showed only part reached trading firms, while tens of millions funded lulling payments and expenses. Cook received a 25-year sentence and several associates were also convicted.
What traders should learn
A managed account must produce statements directly from an independent regulated custodian. Promoter-made reports and smooth monthly returns are not proof.
Black Diamond
How it worked
Black Diamond claimed to operate a Forex trading platform and accepted at least $35 million from at least 240 people. The CFTC found the platform never existed and no Forex trading took place through it.
What happened next
Funds were routed through related entities and Ponzi-style payments. Courts imposed permanent trading bans and, in the 2015 orders, a combined $76 million in civil monetary penalties against key defendants.
What traders should learn
Verify the account at the executing broker, not inside a manager’s private portal. A dashboard can be entirely fictional.
FXCM U.S.
How it worked
FXCM promoted a “No Dealing Desk” model as free from conflicts while routing a substantial share of retail orders to an important market maker. The CFTC found that the relationship and economic interest behind that market maker were concealed from customers.
What happened next
In 2017 the U.S. entities and principals agreed to a $7 million penalty, withdrawal from CFTC registration and permanent bars from registration. This is a historical U.S. regulatory case, not a blanket claim about every present FXCM entity.
What traders should learn
Execution labels such as ECN, STP and NDD are marketing until order routing, counterparties, slippage policy and conflicts can be verified.
Capital Blu
How it worked
Capital Blu raised roughly $17 million for a Forex fund. When trading produced losses, investors received statements reporting positive returns; fund money also covered business and personal expenses.
What happened next
A jury found fraud and a federal court ordered restitution and civil penalties exceeding $17 million. One principal was later sentenced to 17 years in a related criminal case.
What traders should learn
Compare administrator and broker records with the manager’s statement. Repeated gains during real trading losses are a classic warning sign.
RoFX
How it worked
RoFX marketed an automated “robot forex trading” service with a polished website, account tiers and an internal dashboard. The CFTC alleged that no real trading supported the displayed returns and that customer funds were misappropriated.
What happened next
In 2024 a Miami federal court entered default and consent judgments exceeding $225 million in restitution and penalties against the companies and individuals covered by the orders.
What traders should learn
A trading robot is not evidence of trading. Demand the regulated broker, independently verifiable account history, withdrawal proof and ownership of the receiving account.
HSBC / Mark Johnson
How it worked
U.S. prosecutors said HSBC traders traded ahead of a client’s $3.5 billion currency conversion and benefited as the pound moved. Mark Johnson was convicted in 2017 and served a prison sentence.
What happened next
In July 2025 a U.S. appeals court vacated the conviction after later Supreme Court precedent changed the applicable wire-fraud analysis and the panel expressed grave doubt about an alternative theory. The case demonstrates why legal status must be updated.
What traders should learn
Charges, convictions and final outcomes are different facts. Responsible financial reporting records appeals and reversals rather than freezing a case at its most dramatic headline.
EUR/CHF / Alpari UK
How it worked
On 15 January 2015 the Swiss National Bank unexpectedly removed the CHF 1.20 floor against the euro. Prices moved violently, liquidity thinned and some clients accumulated losses beyond their account balances.
What happened next
Alpari UK entered special administration after negative client balances created substantial losses. The event belongs in this report as a contrast: a broker can fail because of market and capital risk without the failure itself being fraud.
What traders should learn
Check negative-balance protection, margin close-out rules, capital strength and treatment of extreme gaps. Regulation reduces risk; it does not abolish market discontinuity.
The control map: what failed in each case
A mark means the weakness was central to the documented event or legal theory. It does not classify every row as the same offence.
| Case | Opaque custody | False records | Hidden conflict | Return promise | Exit failure | Key-person risk |
|---|---|---|---|---|---|---|
| FX benchmark cartel | — | — | ● | — | — | — |
| Refco | ● | ● | ● | — | ● | ● |
| PFGBEST | ● | ● | — | — | ● | ● |
| Trevor Cook / Oxford | ● | ● | — | ● | ● | ● |
| Black Diamond | ● | ● | — | ● | ● | ● |
| FXCM U.S. | — | — | ● | — | — | — |
| Capital Blu | ● | ● | — | ● | ● | ● |
| RoFX | ● | ● | — | ● | ● | ● |
| HSBC / Mark Johnson | — | — | ● | — | — | — |
| EUR/CHF / Alpari UK | — | — | — | — | ● | — |
What the ten cases have in common
The customer saw a label; the risk lived in the legal entity
Broker, bank, fund, robot and platform are descriptions, not protections. Rights attach to a contracting company in a jurisdiction. FXCM showed that an execution label can hide an economic relationship; Alpari showed that a regulated broker can still fail under extreme market stress. The first due-diligence step is always to identify the exact entity holding the obligation.
Independent evidence quietly became promoter-controlled evidence
PFGBEST’s forged bank statements, Cook’s investor reports, Capital Blu’s positive statements and RoFX’s dashboard differed in form but shared one weakness: the party asking for trust also produced the proof. A genuine verification channel must bypass the promoter and reach the bank, custodian, administrator, regulator or executing broker directly.
Smoothness was marketed as safety
Double-digit returns with little risk and automated gains appear attractive because currency prices are liquid and continuously quoted. But a liquid underlying market does not make a manager’s strategy liquid or solvent. Unrealistically steady returns can be created by fabricated reporting, hidden leverage, unpriced tail risk or new deposits funding old withdrawals.
Conflicts mattered most when customers could not see order routing
The cartel cases and FXCM were not Ponzi schemes. Their lesson is about agency. A dealer may execute a customer order while also managing inventory, paying a related liquidity provider or earning from the way an order is routed. Conflict is not automatically fraud, but undisclosed conflict prevents the customer from evaluating price and execution quality.
The final trigger was often smaller than the hidden problem
Refco collapsed after a related-party receivable became public. PFGBEST unravelled when its founder could no longer maintain the confirmation system. The Swiss-franc shock lasted minutes but exposed capital weaknesses built over years. Scandals appear sudden because the trigger is visible; the vulnerability usually predates it.
After the verdict: why a fine is not the same as repayment
Criminal punishment, civil penalties, restitution and insolvency distributions answer different questions. A headline can be legally important without telling a victim how much money will arrive.
PFGBEST
The criminal court ordered more than $215.5 million in restitution and a bankruptcy trustee and receiver gathered assets for creditors. The order established the amount owed; actual distributions depended on assets available, claim status and administration costs. It did not recreate the missing segregated balance.
RoFX and Black Diamond
Large civil judgments combined restitution with penalties. Restitution is directed toward victim loss; civil monetary penalties punish and deter. Default judgment also does not guarantee collection, especially where money moved through offshore entities or personal accounts.
Bank benchmark cases
Corporate fines punished market manipulation and related misconduct, while affected clients could have separate civil routes. A market-wide fine cannot be divided mechanically into a per-trader loss because benchmark effects, transactions and counterparties differ.
Alpari and the general rule
Insolvency distributes available property under a statutory ranking; it is not a fraud-recovery programme. Traders may face delays, set-off questions and valuation rules even when nobody stole the money. Anyone offering to “unlock” a court award for an advance crypto payment is creating a new fraud around an old loss.
The safest question is not “Is this broker famous?” but “Who can falsify the evidence?”
The ten cases range from global banks to a one-man control failure, from fabricated managed accounts to a genuine market shock. Treating them as proof that every Forex broker is dishonest would erase the distinctions the evidence requires. The useful conclusion is operational: trust becomes dangerous when custody, reporting, execution and verification converge inside the same person or corporate group without an independent route for the customer to test them.
Before funding, map six elements: contracting entity, licence, custodian, source of statements, order-routing conflict and exit conditions under stress. A licence is important, but it is not a substitute for that map. If a company cannot explain where money sits, who confirms it and how a balance can be verified outside its own interface, the investigation should happen before the deposit—not after the withdrawal fails.
Editorial method
MyProfitFactor reviewed court and regulator releases, central-bank material, insolvency reporting and the latest available appellate outcome. We separate criminal pleas and convictions from civil findings, allegations, regulatory settlements and market-driven insolvency. Amounts are rounded, and fines, transaction notional, customer shortfalls and restitution are not treated as interchangeable measures.
This is a documentary synthesis, not a claim of original interviews or unpublished evidence and not legal or investment advice. The selected sources below support the central status of each case without turning the article into an unfiltered archive. Readers should use the linked regulator or court record when a legal distinction affects a decision.
Information and legal status are current as of 20 September 2026.
The recurring anatomy of a Forex fraud
Technology and branding change. The pressure tactics and missing controls repeat.
- Guaranteed or unusually smooth monthly returns
- Statements produced only by the promoter
- Deposits sent to a different company or personal account
- A “regulated” claim without the exact legal entity
- ECN, STP or NDD labels without execution evidence
- Withdrawal delays explained by audits, taxes or upgrades
- Secret robot or strategy that cannot be independently verified
- Pressure to add funds after losses or recruit other investors
A ten-minute defence before funding a broker or manager
- Identify the exact contracting legal entity and jurisdiction.
- Check the licence in the regulator’s own register and match the domain.
- Confirm who holds client money and what segregation legally means.
- Ask whether the account is in your name at the executing broker.
- Compare spread, commission, swap, slippage and conversion costs.
- Read negative-balance, margin close-out and insolvency terms.
- Test support and a small withdrawal before increasing capital.
- Save statements, agreements, payment records and correspondence.
Frequently asked questions
Were all ten cases Forex scams?
No. The report deliberately separates fraud, regulatory misconduct, a changing criminal outcome and a market-driven insolvency. That distinction is one of its main lessons.
Does a licence guarantee client money?
No. A licence establishes rules and supervision, but protection depends on the exact entity, product, segregation regime and jurisdiction.
Can an ECN or STP broker trade against clients?
The label alone proves very little. Traders should examine the legal disclosures, execution policy, counterparties and conflict-management rules.
What should a victim do first?
Stop further payments, preserve evidence, contact the relevant bank or payment provider and report to the regulator or law-enforcement channel. Be wary of recovery services demanding advance fees.
Selected court and regulatory records
We keep the source list compact and use the latest material outcome available for each case. Historical allegations, findings, convictions and reversals are labelled separately.