Before the scandal had a name, foreign exchange already looked like a system built to conceal itself. It was the largest market in the world by volume, operating across time zones, with prices flashing on screens in London, New York, Tokyo, and Singapore. Yet one of its most important reference points—the daily WM/Reuters fix—was not a printout from a transparent auction. It was a window, a calculation, and a convention. That mattered because conventions can be gamed long before they are understood as vulnerable.
The benchmark at the center of the affair was not obscure. It was used by asset managers, pension funds, corporations, and sovereign institutions that needed a single reference point to value or execute foreign exchange transactions. The short sampling period around 4 p.m. London time made the fix efficient, but also exposed. In the minutes around that window, a trader with knowledge of client flow could anticipate pressure in the market; a trader with allies could amplify it. The opportunity was not hypothetical. It was embedded in the way the market worked.
The traders at the center of this story were not outsiders trying to break into finance. They were insiders at major banks, moving in a culture that prized speed, relationships, and informal coordination. The public record, including later enforcement releases, shows the alleged conduct concentrated among desks at Citigroup, JPMorgan, Barclays, Royal Bank of Scotland, and UBS. Each bank had its own compliance regime, but the FX market’s structure blurred the edges: client orders arrived throughout the day, the benchmark was widely used, and the banks that traded around it also helped shape the flow that determined it.
That overlap between client service and proprietary advantage is where the story begins to darken. A trader holding a large order could see, or infer, which way the market might move when the fix window opened. If a rival on another desk knew the same thing, then, in theory, the traders could help each other by concentrating orders, steering the market, and sharing the timing of the trade. According to later regulatory findings, what began as awareness of shared incentives became coordination, first in small exchanges and then in a private chat room that gave the conspiracy its public shorthand: “The Cartel.” The significance of that label was not that it sounded theatrical, but that it described a practical arrangement: a space where banks’ traders could compare information and act in parallel at a moment when the benchmark was most vulnerable.
The conditions that enabled it were structural. FX was decentralized; there was no single exchange with a central order book. Supervision was fragmented across jurisdictions. Compliance teams were often focused on outright spoofing, market abuse in listed products, or obvious conflicts, while benchmark trading in spot currency sat in a softer zone of custom and practice. In London, where much of the conduct took place, the City’s trading culture still prized the old virtues of the dealing room: loyalty, discretion, and the unspoken understanding that what mattered was not whether one trader had an edge, but whether the edge could be defended as market knowledge.
A concrete scene captures the atmosphere. In a glass office tower in London’s financial district, traders sat before multiple monitors showing spot rates, order tickets, and instant-message windows. The room was noisy even when nobody raised a voice: keyboards clattering, phones clicking back into their cradles, the low hum of market data. A benchmark order did not look like a crime scene. It looked like business. Yet the very ordinary appearance of the desks was what made the behavior difficult to detect. Large orders could be discussed, split, timed, and re-timed without any single moment standing out as obviously improper.
Another scene played out far from the trading floor, in the administrative machinery that made the benchmark possible. The WM/Reuters fix was based on a short sampling period around 4 p.m. London time, and banks could measure their own client flows against that interval. The very simplicity of the calculation was its weakness. If enough large players knew the window, and enough of them cared to act together, the fix could be nudged by coordinated buying or selling. Nothing about that required a grand conspiracy in the cinematic sense. It required only enough traders, enough information, and enough confidence that the chance of detection was low.
The tension was built into the day. Traders who were supposed to represent customers’ interests were also judged by their desks on profitability. If a large client order could be traded less expensively by using information shared with peers, there was an incentive to rationalize the practice as harmless market color. The legal line, however, was not obscure to everyone. The later settlements and prosecutions make clear that regulators viewed this as conduct that crossed from aggressive trading into collusion.
One surprising fact helps explain the scale of the opportunity: the foreign exchange market was so large that even a tiny movement at the benchmark could affect vast sums. The settlement documents and later reporting repeatedly described the benchmark as influencing trillions in transactions globally, which is why a few basis points of movement could matter so much to bank revenues and client outcomes. A slight push at the wrong time could change the economics of a corporate hedge, a pension portfolio rebalancing, or a fund’s attempt to convert currency at a fair rate. That scale turned a technical benchmark into a point of enormous leverage.
At first, the arrangement was operational before it was infamous. Orders were still being filled. Messages were still being sent. Traders were still making money. The first money flowing in did not announce itself with sirens. It arrived as better execution, better timing, and incremental profit, hidden in the narrow space between what clients thought they were paying and what the market was made to show. The harm was diffuse and therefore easy to miss in real time. A client saw a price that seemed plausible. A desk saw a day that closed profitably. The gap between the two was the scandal’s initial hiding place.
The apparatus also benefited from its own ordinariness. Because the benchmark was part of the normal rhythm of the market, there was no single dramatic alert when it was manipulated. There were no flashing warnings on the screens in the London dealing rooms, no immediate collapse in liquidity, no public sign that the fix had become a focal point for coordination. What existed instead were patterns: repeated timing around the window, repeated communication among traders, repeated efforts to align interest where market participants were supposed to compete.
This is why the first stage of the story is so important. The scandal did not begin with a confession or a whistleblower’s file dropped on a desk. It began with a structure that made coordination possible, a benchmark that concentrated attention, and a culture that treated intimate market knowledge as professional competence. Those ingredients mattered long before regulators assembled the record. They explain how a practice that might once have seemed like sharp dealing could harden into something more organized.
The documents that later brought the affair into view—regulatory releases, settlement papers, and court filings—would show that the conduct was concentrated among named desks at some of the world’s largest banks. But before those papers existed in public, the market itself had already supplied the conditions. The fix window was known. The flows were visible. The incentives were misaligned. And in the private channels where traders gathered, the idea of collaboration could be made to sound less like conspiracy than routine.
By the time anyone outside the desks understood that the benchmark itself might be bent, the machinery was already working smoothly enough to feed on its own logic. And once the traders realized they could speak in code, privately, and with apparent impunity, the next step was not to stop. It was to recruit the room.