How To Rig FX Like A Pro "Bandit", And Make Millions In The Process

We finally have the answer, courtesy of the FCA's partial and very much selective disclosure of FX rigging [8]findings by "The Cartel", the "Bandits [9]" and so on, as part of its wrist-slapping settlement, just how the big boys make millions in FX on every single fix. Hopefully one day the regulators, who are as corrupt and conflicted as the banks they quote-unquote police, will reveal all the documents in their possession and let the public decide what is important and what isn't. But in the meantime, for all those curious just why the Too Big To Fail are also Too Big To Prosecute, here is the blow by blow.

First, an example of JPMorgan manipulating the fix

An example of JPMorgan’s involvement in this behaviour occurred on one day within the Relevant Period when JPMorgan attempted to manipulate the WMR fix in the EUR/USD currency pair. On this day, JPMorgan had net buy orders at the fix which meant that it would benefit if it was able to move the WMR fix rate upwards. The chances of successfully manipulating the fix rate in this manner would be improved if JPMorgan and another firm or firms adopted trading strategies based upon the information they shared with each other about their net orders.

In the period between 3:41pm and 3:51pm on this day, traders at two different firms (including JPMorgan) inappropriately disclosed to each other via a chat room details about their net orders in respect of the forthcoming WMR fix in order to determine their trading strategies. The other firm is referred to in this Final Notice as Firm A. On the day in question, a third firm (Firm B) was a member of the chat room, but did not participate in the discussions. JPMorgan then participated in the series of actions described below in an attempt to manipulate the fix rate higher.

1. At 3:43pm, Firm A asked JPMorgan whether it would need to buy EUR in the market for the forthcoming WMR fix. JPMorgan responded that it had net buy orders for the fix, which it subsequently confirmed amounted to EUR105 million. It offered to transfer its net buy orders to Firm A.
2. At 3:44pm, Firm A replied “maybe” and went on to state that it had a buy order “for a top [account]” for EUR150 million at the fix.
3. At 3:46pm, Firm A then stated “i'd prefer we join forces”. JPMorgan responded “perfick…lets do ths…lets double team em”. Firm A replied “YESsssssssssss”. The Authority considers these statements to refer to the possibility of JPMorgan and Firm A co-ordinating their actions in an attempt to manipulate the fix rate higher. Since JPMorgan and Firm A each needed to buy EUR at the fix, each would profit to the extent that the fix rate at which it sold EUR was higher than the average rate at which it bought EUR in the market.
4. At 3:47pm and 3:51pm, JPMorgan informed Firm A that it had conducted trades with third parties that resulted in it needing to buy additional EUR at the fix. This is an example of “building”.
5. At 3:48pm, Firm A said that it was monitoring activity in relation to the forthcoming fix in the interdealer broker market (“i got the bookies covered”).

In the period leading up to the fix, JPMorgan “built” the volume of EUR that it needed to buy for the fix to a total of approximately EUR278 million via a series of transactions with market participants. Firm A had net buy orders associated with its client fix orders of EUR170 million in the period leading up to the fix. It increased this amount (or “built”) by EUR70 million.

From 3:52pm until the opening of the fix window at 3:59:30pm, JPMorgan and Firm A bought EUR on the EBS trading platform. In particular JPMorgan bought EUR57 million from 3:58pm onwards. These early trades were designed to take advantage of the expected upward movement in the fix rate following the discussions within the chat room described above.

In the first five seconds of the fix window, JPMorgan and Firm A each placed orders to buy EUR50 million and subsequently placed smaller orders to buy EUR throughout the remainder of the fix window. During the 60 second fix window, JPMorgan bought a total of EUR134 million and Firm A bought EUR125 million. Between them, they accounted for 41% of the volume of EUR/USD bought during the fix window.

The rate prevailing on EBS at the start of the fix window was 1.3957. Over the course of the window period, the rate rose and WM Reuters subsequently published the fix rate for EUR/USD at 1.39605.

The information disclosed between JPMorgan and Firm A regarding their order flows was used to determine their trading strategies. The consequent “building” by JPMorgan and its trading in relation to that increased quantity in advance of and during the fix window were designed to increase the WMR fix rate to JPMorgan’s benefit JPMorgan’s trading in EUR/USD in this example generated a profit of approximately USD33,000.

Subsequent to the WMR fix, the two traders discussed the outcome of their trading. At 4:03pm, Firm A stated “sml rumour we havent lost it”. JPMorgan responded “we…do…dollarrr”.

The following day Firm A stated to Firm B “we were EPIC at the [WMR] fix yest”. Firm B responded “yeeeeeeeeeeeeeeeeeeah”. Firm A added “i dragged [JPMorgan] in , we covered all the bases b/w us”. Firm B commented “so couldnt have been that $hit a week!!”

* * *

Here is UBS doing the same and making $513K in the process

An example of UBS’s involvement in this behaviour occurred on one day within the Relevant Period when UBS attempted to manipulate the ECB fix in the EUR/USD currency pair. On this day, UBS had net client sell orders at the fix which meant that it would benefit if it was able to move the ECB fix rate lower.10 The chances of successfully manipulating the fix rate in this manner would be improved if UBS and other firms adopted trading strategies based upon the information they shared with each other about their net orders.

In the period between 12:35pm and 1:08pm on this day, traders at four different firms (including UBS) inappropriately disclosed to each other via a chat room details about their net orders in respect of the forthcoming ECB fix at 1:15pm in order to determine their trading strategies. The other three firms are referred to in this Final Notice as Firm A, B and C. UBS then participated in the series of actions described below in an attempt to manipulate the fix rate lower.

1. At 12:36pm, Firm A disclosed that it had net sell orders for the fix. At 12:37pm, Firm A disclosed that these net sell orders were EUR200 million. At 12:40pm, Firm A updated this figure to EUR175 million.
2. At 12:36pm, UBS disclosed that it had net sell orders for the fix of EUR200 million. At 12:44pm, UBS disclosed that its net sell orders had increased to EUR250 million. Since UBS needed to sell Euros at the fix it would profit to the extent that the fix rate at which it bought Euros was lower than the average rate at which it sold Euros in the market.
3. At 12:36pm, Firm B disclosed that it had net sell orders for the fix of EUR100 million and that another of its offices also had net sell orders.
4. At 12:48pm, Firm A disclosed that its net sell orders had reduced to EUR100 million, but that it was “…hopefully taking all the filth out for u…”. The Authority considers that this statement referred to Firm A having netted off part of its net sell orders with smaller buy orders held by third parties, which might otherwise have traded in the opposite direction to UBS at the ECB fix. This is an example of Firm A “clearing the decks”.
5. At 1:02pm, Firm A disclosed that it had sold EUR25 million to a client in a transaction separate to the fix but would remain EUR25 million short (“lose… shet [i.e. 25 million] though natch dont buy”). The Authority considers that this statement referred to Firm A’s intention not to buy this amount of Euros in the market immediately, but to take advantage of the anticipated downwards rate movement at the fix by only buying when the rate had dropped.
6. In response, UBS disclosed that it had also sold EUR25 million to a client in a separate transaction. UBS inappropriately revealed the identity of the client to the chat room using a code known to the chat room participants. Firm B indicated that these short positions should be held for 12 minutes (i.e. until the ECB fix).
7. At 1:03pm, Firm A disclosed that it had been trading in the market and its net sell orders at the fix had been reduced to EUR50 million (“i getting chipped away at a load of bank filth for the fix… back to bully [i.e. 50 million]… hopefully decks bit cleaner”). The Authority considers this to refer to trades between Firm A and other market participants, whose buy orders might otherwise be traded in the opposite direction to UBS and Firm A at the fix. This is a further example of Firm A “clearing the decks”.
8. At 1:04pm, UBS disclosed that it still had net sell orders for EUR200 million at the forthcoming ECB fix. UBS also stated that it had a separate short position of EUR50 million. At 1:05pm, Firm B disclosed that it also had a short position of EUR50 million.
9. At 1:07pm, Firm C disclosed that it had net buy orders of EUR65 million at the forthcoming ECB fix. Firm C subsequently netted off with Firm A and Firm B, such that at 1:08pm Firm C disclosed that it only had EUR10 million left to buy in the opposite direction at the fix. This is an example of “leaving you with the ammo”. Firm B advised Firm C to “go late” (i.e. buy later when the rate would be lower).
10. At 1:14pm, Firm B copied into the chat a comment made by UBS at 12:04pm that day describing an earlier fix as “the best fix of my ubs career.” Firm B then said “chalenge [sic]” and Firm C added the comment “stars aligned”.

UBS’s net sell orders associated with its client fix orders were EUR86 million. During the period leading up to the ECB fix, UBS increased (or “built”) the volume of Euros that it would sell for the fix to EUR211 million through a series of additional trades conducted with other market participants, well above that necessary to manage UBS’s risk associated with net client orders at the fix.

From 12:35pm to 1:14pm, UBS sold a net amount of EUR132 million. At 1:14:59pm (i.e. 1 second before the ECB fix), UBS placed an order to sell EUR100 million at 1.3092, which was three basis points below the prevailing best market bid at that time.

This order was immediately executed and accounted for 29% of the sales in EUR/USD on the EBS platform during the period from 1:14:55 to 1:15:02pm.

The ECB subsequently published the fix rate for EUR/USD at 1.3092.

The information disclosed between UBS and Firms A, B and C, regarding their order flows was used to determine their trading strategies. The consequent “building” by UBS and its trading in relation to that increased quantity at the fix were designed to decrease the ECB fix rate to UBS’s benefit. UBS undertook the selling of Euros prior to the 1:15pm ECB fix in anticipation that the fix rate at which it would buy Euros would be lower than the average rate at which it had sold. The placing of a large sell order by UBS immediately prior to 1:15pm was designed to achieve this outcome. UBS’s trading in EUR/USD in this example generated a profit of USD513,000.

In the immediate aftermath of the ECB fix, UBS was congratulated on the success of its trading by Firms A, B and C (“hes sat back in his chaoir [sic]…feet on desk…announcing to desk…thats why i got the bonus pool” and “yeah made most peoples year”).

* * *

JPM triggering client stop loss orders

During its investigation, the Authority identified instances within JPMorgan’s G10 spot FX trading business of attempts to trigger client stop loss orders. These attempts involved inappropriate disclosures to traders at other firms concerning details of the size, direction and level of client stop loss orders. The traders involved would trade in a manner aimed at manipulating the spot FX rate, such that the stop loss order was triggered. JPMorgan would potentially profit from this activity because if successful it would, for example, have sold the particular currency to its client pursuant to the stop loss order at a higher rate than it had bought that currency in the market.

This behaviour was reflected in language used by G10 spot FX traders at JPMorgan in chat rooms. For example, a JPMorgan trader explained to other traders in a chat room that he had traded in the market in order “to get the 69 print” (i.e. to move the spot FX rate for that currency pair to the level (“69”) at which a stop loss would be triggered). On another occasion, the same trader disclosed the level of certain clients’ stop loss orders to other JPMorgan traders in a chat room and asked “shall we go get these stops?”

* * *

UBS tigerring stop loss orders

During its investigation, the Authority identified instances within UBS’s G10 spot FX trading business of attempts to trigger client stop loss orders. These attempts involved inappropriate disclosures to traders at other firms concerning details of the size, direction and level of client stop loss orders. The traders involved would trade in a manner aimed at manipulating the spot FX rate, such that the stop loss order was triggered. UBS would potentially profit from this activity because if successful it would, for example, have sold the particular currency to its client pursuant to the stop loss order at a higher rate than it had bought that currency in the market.

This behaviour was reflected in language used by G10 spot FX traders at UBS in chat rooms. For example, one UBS trader commented in a chat room “i had stops for years but they got sick of my butchering”. On a subsequent occasion, the same trader described himself as “just jamming a little stop here.”

UBS leaking confidential information

The attempts to manipulate the WMR and ECB fixes and trigger client stop loss orders described in this Notice involved inappropriate disclosures of client order flows at fixes and details of client stop loss orders.

HSBC is pretty good too. Here is the scandal-ridden bank rigging a GBP fix:

1. At 2:50pm, Firm A disclosed in a chat room (including to HSBC) that it had net sell orders for more than GBP100 million at the fix. At 3:25pm, Firm A indicated that the orders were for approximately GBP130 million.
2. At 3:25pm, HSBC disclosed to Firm A in a one-to-one chat that it had net client sell orders for GBP400 million at the fix. Since HSBC and Firm A each needed to sell GBP at the fix each would profit to the extent that the fix rate at which it bought GBP was lower than the average rate at which it sold GBP in the market.
3. Firm A informed HSBC that it now had net sell orders of GBP150 million at the fix. HSBC responded by saying “lets go”,11 to which Firm A replied “yeah baby”. The Authority considers these statements to refer to the possibility of HSBC and Firm A co-ordinating their actions in an attempt to manipulate the fix rate downwards.
4. At 3:28pm in a chat room which included HSBC, Firm A expressed the hope that other traders would also have sell orders at the fix (“hopefulyl a fe wmore get same way and we can team whack it”). At 3:36pm, Firm B, which was a participant in the chat room, confirmed to the other traders that he now also had net sell orders for GBP40 million at the fix.
5. At 3:28pm, HSBC informed Firm C via a one-to-one chat room that he had net client sell orders of around GBP300 million at the fix and asked the trader to do some “digging” to see if anyone else had orders in the same direction at the fix. Firm C replied at 3:34pm and disclosed to HSBC that it now also had net sell orders of GBP83 million at the fix.
6. At 3:36pm, Firm D asked Firm A in a chat room (which included HSBC), for an update on its net sell orders. Firm A disclosed that it had now increased to GBP170 million. Firm D noted that it did not have any fix orders at that time, but commented that he expected Firm A to “bash the fck out of it”.
7. At 3:38pm, HSBC commented simultaneously into chat rooms in which Firms A, C and D participated that it had net client sell orders at the fix for GBP in a “good amount”.
8. At 3:42pm, in a one-to-one chat Firm A warned HSBC that another firm which was not a participant in the chat room (Firm E) was “buidling” in the opposite direction to them and would be buying at the fix.
9. At 3:43pm, Firm A updated HSBC by indicating that it had netted some of its sell order off with Firm E and “taken him out… so shud have giot rid of main buyer for u…im stilla seller of 90… gives us a chance”. The Authority considers that this refers to Firm A’s belief that Firm E would no longer be transacting its orders in the opposite direction at the fix. It also confirmed that Firm A still held net sell orders for GBP90 million to trade at the fix and could still participate in the co-ordinated behaviour. This is an example of Firm A “clearing the decks”.

The information disclosed between HSBC and Firms A, B and C, regarding their order flows was used to determine their trading strategies. The consequent trading by HSBC during the fix window was designed to decrease the WMR fix rate to HSBC’s benefit. HSBC’s trading in GBP/USD in this example generated a profit of approximately USD162,000.

Subsequent to the fix, traders in the chat rooms congratulated one another by saying: “nice work gents…I don my hat”, “Hooray nice team work”, “bravo…cudnt been better” and “have that my son…v nice mate” and “dont mess with our ccy [currency]”. One of the traders commented “there you go … go early, move it, hold it, push it”. HSBC stated “loved that mate… worked lovely… pity we couldn’t get it below the 00”12 and “we need a few more of those for me to get back on track this month”.

Here is HSBC crucifying client stops:

For example, an HSBC trader in a chat room referred to “going to go for broke at this stop… it is either going to end in massive glory or tears”. On another occasion, the same trader refers in a chat room to the fact he is “just about to slam some stops”. When asked by a colleague whether a particular client’s stop loss orders were “a pain for you guys”, another HSBC trader replied “nah love them … free money” and “we love the orders … always make money on them”.

Because a stop-loss muppet is born every minute.

* * *

Here is Citi rigging the fix...

During the period from 1:14:29pm to 1:15:02pm, Citi bought EUR374 million which accounted for 73% of all purchases on the EBS platform. At 1:15:00pm, the bid (buying price) and the first trade for EUR/USD on the EBS platform was 1.3222. The ECB subsequently published the fix rate for EUR/USD at 1.3222.

The information disclosed between Citi and Firms A, B, C and D regarding their order flows was used to determine their trading strategies. The consequent “building” by Citi and its trading in relation to that increased quantity at the fix were designed to increase the ECB fix rate to Citi’s benefit. Citi bought EUR prior to the 1:15pm fix in anticipation that the fix rate at which it would sell EUR would be higher than the average rate at which it had bought. The placing of large buy orders by Citi immediately prior to 1:15pm was designed to achieve this outcome by improving the chance that the first trade on the EBS platform at 1:15:00pm, which it believed to be the basis for the ECB fix, was at a higher level. Citi’s trading in EUR/USD in this example generated a profit of USD99,000.

Subsequent to the ECB fix, Citi’s trading was variously described by other traders in chat rooms as “impressive”, “lovely” and “cnt teach that”. Citi noted “yeah worked ok”. When the fix rate was published to the market, Firm A commented “22 the rate” and Citi replied “always was gonna be."

* * *

And taking out client stops:

During its investigation, the Authority identified instances within Citi’s G10 spot FX trading business of attempts to trigger client stop loss orders. These attempts involved inappropriate disclosures to traders at other firms concerning details of the size, direction and level of client stop loss orders. The traders involved would trade in a manner aimed at manipulating the spot FX rate, such that the stop loss order was triggered. Citi would potentially profit from this activity because if successful it would, for example, have sold the particular currency to its client pursuant to the stop loss order at a higher rate than it had bought that currency in the market.

This behaviour was reflected in language used by G10 spot FX traders at Citi in chat rooms. For example, a Citi trader referred in a chat room to the fact he “had to launch into the 50 offer to get me stop done”. On another occasion, a trader at Citi described in a chat room how he “went for a stop”.

* * *

Finally, here is RBS:

In the period leading up to the 4pm fix, RBS increased (or “built”) the volume of GBP it would sell at the fix via a series of trades conducted with other market participants. RBS commenced this “building” after Firm A’s disclosure of its net sell orders at 3:22pm. Subsequently RBS received further client orders for the fix and briefly had net orders to buy GBP25 million before further “building” and client orders resulted in RBS having to sell GBP at the fix. Ultimately, RBS’s net sell orders associated with its client fix orders was GBP202 million; it “built” the volume of currency that it needed to sell at the fix to GBP399 million, well above that necessary to manage the risk associated with net client orders.

From 3:50:30pm to 3:52:10pm, RBS placed a series of sell orders in the GBP/USD currency pair on the Reuters platform. During this period, RBS sold GBP93 million and the GBP/USD rate dropped from 1.6276 to 1.6250. At 3:52pm, Firm C commented “nice job gents”.

In the period from 3:50:30pm to 3:59:30pm (i.e. immediately prior to the 4pm WMR fix window), RBS sold a total of GBP167 million and Firm A sold GBP26 million. Together they accounted for 28% of all sales on the Reuters platform during this period. The GBP/USD rate steadily dropped from 1.6276 to 1.6233. These early trades were designed to take advantage of the expected downwards movement in the fix rate following the discussions within the chat rooms described above.

During the 60 second fix window, RBS sold GBP182 million, which accounted for more than 32% of the sales in GBP/USD on the Reuters platform. RBS and Firm A together accounted for 41% of the sales in GBP/USD on the Reuters platform during the fix window. During this period, the GBP/USD rate fell from 1.6233 to 1.6213. Subsequently WM Reuters published the 4pm fix rate for GBP/USD at 1.6218.

The information disclosed between RBS and Firms A, B and C, regarding their order flows was used to determine their trading strategies. The consequent “building” by RBS and its trading in relation to that increased quantity in advance of and during the fix window, were designed to lower the WMR fix rate to RBS’s benefit. RBS’s trading in GBP/USD in this example generated a profit of USD615,000.

The trading was discussed by the participants in the chat rooms subsequent to the fix, with references to “I don my hat”, “welld one [sic] lads”, “what a job”, “bravo” and “[RBS] is god”. RBS commented when the 4pm WMR fix rate was published “1.6218…nice”, whilst Firm A commented later on ”we fooking killed it right… [Firm C], myself and RBS."