SUPREME COURT OF THE STATE OF NEW YORK

COUNTY OF NEW YORK

and

,

Defendants.

Plaintiff, the People of the State of New York, by Eric T. Schneiderman, Attorney

General of the State of New York (the “Attorney General”), alleges the following against

Barclays Capital, Inc. and Barclays PLC (together, “Barclays”):

PRELIMINARY STATEMENT

additional wrongdoing by Barclays, as set forth in this Amended Complaint. This Amended

Complaint also includes further details relating to the allegations set forth in the initial

Complaint.

3. Beginning in 2011 and continuing at least until the filing of the initial Complaint

in this action, Barclays engaged in a broad effort to deceive clients and the investing public about

how, and for whose benefit, Barclays operated its electronic trading services.

4. In 2011, Barclays’ Equities Electronic Trading Division embarked on a business

plan to dramatically increase the revenues and market share of its three core, interrelated

products and services: (i) algorithmic trading services; (ii) smart order routing services; and (iii)

a private securities trading venue known as a “dark pool.” Each of those relates directly to the

Client Protection” from “aggressive,” “predatory,” and “toxic” high frequency traders.

5. Barclays represented that allowing it to control client trades from start to finish

would (i) increase the rates at which those orders were filled; (ii) decrease the likelihood that

prices would move against clients; and (iii) minimize the likelihood that high frequency traders

would act as counterparties. In short, Barclays represented that its electronic trading services and

products would have a positive impact on the likelihood that client orders would be filled and

would have a positive impact on the prices at which those client orders would be executed.

Those are core aspects of any securities transaction.

6. Barclays’ representations were contrary to what Barclays was, in fact, doing with

client orders. In fact, Barclays operated its electronic trading services in its own interests and at

the expense of its clients by funneling as many orders as possible to Barclays’ own dark pool,

knowing that those orders were exposed to the aggressive or predatory high frequency traders

that Barclays claimed it was protecting its clients from.

7. As discussed in further detail below, Barclays’ wrongdoing includes the

following:

counterparties that Barclays represented to clients its algorithms were designed to avoid;

b) Barclays falsely represented that its “Liquidity Profiling” service applied

to, and protected, orders routed to its dark pool by Barclays’ algorithms. In truth,

Liquidity Profiling did not apply to such orders;

c) Barclays falsely represented that its smart order router “treat[s] all venues

the same based on execution quality,” and was not biased in favor of any trading venue.

In fact, Barclays’ smart order router was biased in favor of its own dark pool, regardless

of whether an order was likely to be filled there, and secondarily to other venues based on

whether those venues were profitable for Barclays. When a detailed analysis of Barclays’

order routing practices was conducted for a major institutional investor (that manages

Barclays was routing and executing the vast bulk of this client’s sampled orders to

Barclays’ own dark pool, senior Barclays executives directed that a written presentation

to that client be falsified to mask Barclays’ biased order routing practices;

d) Barclays provided false analyses to clients, potential clients, and the public

to hide the extent and type of high frequency trading in its dark pool. For instance, in

June 2012, senior Barclays employees doctored a graphic representation of the trading in

its pool by removing its then-largest participant, a high frequency trading firm named

multiple internal documents obtained by the Attorney General show that Barclays knew

that at least 25% of the trading activity in the pool was “aggressive”;

e) Barclays made a series of material false representations to clients,

potential clients, and the public about its “Liquidity Profiling” service. Barclays claimed

that its Liquidity Profiling service allowed traders to limit the kinds of counterparties

with whom they would interact, “protect [clients] from predatory trading,” and that

Barclays would “continuously police . . . trading activity” in order to “maintain quality

flow” in the dark pool. Those representations were false for several reasons, including

that: (i) Barclays never removed a single trader – even known predatory traders – from

its dark pool; (ii) Barclays failed to regularly profile traders in its dark pool; (iii) Barclays

desk in order to make each appear less predatory than they really were; (iv) Barclays

failed to apply the protections of Liquidity Profiling to a significant portion of the trading

in its dark pool; (v) Barclays misled clients as to how Liquidity Profiling evaluated

traders; and

f) In an effort to demonstrate its commitment to protecting clients from

predatory traders, Barclays falsely represented that it had removed at least one predatory

trader from its dark pool. Documents obtained by the Attorney General confirm that that

with technical tools to exploit traditional investors.

9. In sum, Barclays represented to clients, potential clients, and the investing public

that its electronic trading products and services offered protection from predatory or aggressive

high frequency trading. In truth, Barclays’ electronic trading products and services – its

algorithms, its smart order router, and its dark pool – operated at odds with those representations.

10. As the detailed allegations herein demonstrate, the wrongdoing at issue was not

limited to a few isolated incidents, nor was it the result of a few rogue employees. As detailed

below, this was a broad course of conduct involving numerous Barclays employees.

11. Finally, despite representations to its shareholders and the investing public that

Barclays “will continue to cooperate with the New York attorney general,” Barclays has refused

instance, in the face of lawful subpoenas issued by the Attorney General demanding testimony

under oath of the two highest-ranking executives within Barclays’ Equities Electronics Trading

Division – both of whom were directly involved in, and oversaw, much of the fraud alleged –

Barclays has refused to produce those executives for questioning.

12. Plaintiff Eric T. Schneiderman is the Attorney General of the State of New York.

13. The State of New York has an interest in upholding the laws of the State, and the

or from this State.

15. Barclays PLC is registered in England. Barclays PLC functions in the United

States through Barclays Capital Inc., an affiliate of Barclays PLC. Barclays operates in fifty

countries with particularly large business operations in New York and London. Barclays Capital

Inc. is a registered broker-dealer and investment adviser with its primary offices at 745 Seventh

Avenue, New York, New York.

16. As part of its business in New York, Barclays operates as a registered broker-

dealer. Barclays’ brokerage clients, serviced out of Barclays’ New York offices, include many

institutional investors, mutual funds, pension funds, and others (“buy-side” clients).

17. Many of Barclays’ buy-side clients manage accounts for thousands (or more)

18. High frequency trading is a type of trading that uses sophisticated computer

programming to conduct securities transactions extremely quickly in order to take advantage of

small, momentary changes in stock prices. While not all high frequency traders use the same

kinds of trading strategies, certain characteristics are common. Those firms tend to trade

securities for their own account (as opposed to doing so on behalf of clients), and use high-speed,

19. Some estimates place high frequency trading activity at over fifty percent of the

total trading volume in U.S.-listed equities.

20. In order to execute their trading strategies effectively, high frequency traders seek

speed advantages. Some pay to “co-locate” or “cross-connect” their trading computers in the

same facilities as public exchanges and dark pools in order to reduce the amount of time it takes

to receive market information from those trading venues, and in order to rapidly place or cancel

orders. These firms pay a premium for “direct data feeds” from trading venues, which are high-

speed data feeds that travel faster and contain more information than market data available to

ordinary investors by other, less expensive means.

21. Those speed and technology advantages allow high frequency traders to profile

usually from institutional investors. This “information leakage” allows high frequency traders to

trade ahead of an anticipated stock purchase or sale or to have a negative impact on the price

obtained by the institutional investor in that securities transaction. Speed and technology

advantages also allow for strategies that seek to exploit the small, temporary pricing dislocations

in a security that occur because of differential and/or delayed access to market data. This

strategy is sometimes referred to as “latency arbitrage,” because the high-speed trader is seeking

to exploit the relative slowness, or “latency,” in the transmission of market data experienced by

worse prices.

24. Certain brokers, including Barclays, have developed electronic products and

services, supposedly to help traditional investors avoid interactions with predatory or aggressive

high frequency traders. Among the products and services marketed for this purpose are (i)

algorithms; (ii) smart order routers; and (iii) dark pools.

25. “Algorithms,” in the context of electronic trading, are automated computer

programs that enact complex trading instructions to automatically execute large client trades over

a period of time, usually over the course of a trading day. Algorithms are typically marketed as

tools for institutional investors to effectively place large stock orders, dividing up the order into

smaller portions, avoiding (as much as possible) interaction with aggressive high frequency

generally take into account the size of the order, the current price of the security on various

trading venues, and client objectives when determining how to execute the trade over a given

period of time.

26. “Smart order routers” are automated systems that, generally speaking, look for

liquidity on the various equities trading venues, making decisions about where to send orders so

as to maximize the likelihood that the order is filled. In many circumstances, a smart order

router functions after an algorithm; for instance, after the algorithm determines the general

executions are posted immediately, dark pools generally do not post pending orders. The lack of

visibility of pending orders – the “dark” aspect of such venues – is thought to help protect

traditional (non-high frequency) traders with large orders and long-term positions from the

strategies employed by high frequency trading firms on the public exchanges, as discussed

above. It is estimated that a significant percentage of all U.S. equities trades are executed in dark

pools.

TRADING DIVISION , BASED ON SUPPOSEDLY PROTECTING INVESTORS

28. Beginning in 2011, Barclays began a major effort to grow the revenues of its

Equities Electronic Trading Division, which operates several related lines of business, including

algorithms, smart order routing, and Barclays’ dark pool.

50. Barclays represented that instead of using a “static schedule” of venues, its

algorithms used “[d]ynamic order placement” logic in order to “eliminate a detectable pattern” in

the placement of trades.

51. Those representations were false. In reality, and contrary to Barclays’

representations, its algorithms were designed to preference Barclays’ own dark pool.

52. On January 31, 2011, William Libby, Barclays’ Vice President of Electronic

Trading Services, sent a document entitled “2011 Goals” to William White. In that document,

minutes of an August 30, 2012 “Best Execution Committee” meeting, Merrell Hora (then the

Head of Algorithms) told the Committee that Barclays had “chang[ed] the overall algo plant”

which “resulted more often in rebate taking and which often involved targeting LX more

frequently.” Further, Hora explained that one particular algorithm “had been decommissioned

because the core algos has been updated at this point to preference LX anyway” (emphasis

added).

54. Later in the meeting, Jacek Janczewski, head of dark pool operations, explained

that “a big focus for the remainder of the year was to identify and enable more flow to be

internalized that was currently not accessing” the dark pool.

55. On November 6, 2012, Nej D’jelal, a Barclays employee, wrote an email to senior

“As expected, LX volumes and our cross rate continue to grow . . . Deliverables this month saw

the introduction of key LX items in the form of Liquidity Profiling and new LX friendly algo

logic” (emphasis added).

56. In an April 19, 2013 instant messaging exchange between Antonio Trillo and

David Green, both Barclays sales personnel, Trillo stated that “we preference LX on the [smart

order router] portion of a strategy like hydra,” and noted that a high percentage of algorithmic

orders execute in Barclays’ own dark pool (emphasis added).

traders – which can impact the rates and prices at which client orders are executed – was the very

outcome that Barclays represented it would help clients avoid if they used Barclays’ algorithms.

2. Contrary to Barclays’ Representations, Client Orders Sent to the Dark

Service

58. Barclays also represented to clients and potential clients that orders routed to the

dark pool via Barclays’ proprietary algorithms would be subject to, and thus protected by, its

“Liquidity Profiling” service.

59. As described in greater detail in Section III(E), below, Barclays’ “Liquidity

Profiling” was a surveillance system that supposedly would allow Barclays’ clients to determine

the types of counterparties with whom they trade in Barclays’ dark pool. Barclays represented

that Liquidity Profiling monitored every trade in the dark pool, objectively and fairly graded

traders by how “toxic” or “aggressive” their trading activity was, and allowed clients to decline

to trade with such counterparties when trading in the pool.

60. As an example of how Barclays’ representations to clients, in a written

“Reference Guide” first given to Barclays sales staff in late 2011, Barclays set forth a number of

“Frequently Asked Questions” from clients regarding Barclays algorithmic offerings. In

offering adapted to efficiently facilitate execution with the increase in high frequency trading

over the last several years?” Barclays directed its sales staff to “[p]lease see question 1 regarding

liquidity profiling,” which is the response noted in the preceding paragraph.

62. The answers set forth in the “Reference Guide” remained in use by Barclays’

sales personnel, in various forms, through at least 2012.

63. Senior Barclays employees represented to the investing public that orders sent

through Barclays algorithms were protected by Liquidity Profiling. On June 6, 2013, the online

magazine Hedgeweek published an interview with William White, Head of Electronic Trading,

wherein White stated that Barclays “[has] an algorithm and a scoring methodology that basically

creates a liquidity spectrum . . . as we are able to restrict HFTs interacting with our clients we’re

64. Barclays made similar representations in response to numerous written client

requests seeking information about how Barclays’ algorithms operated.

65. For example, in February 2013, Barclays was sent a written list of questions by a

major public pension plan, which manages accounts on behalf of hundreds of thousands of

public employees. One of the questions asked “Do you categorize clients using your

algorithms?” In its written response to the pension plan, Barclays answered that while its

algorithms themselves do not categorize clients, in the dark pool “we employ our advanced

67. Barclays’ written representations to its clients and potential clients regarding the

applicability of Liquidity Profiling to algorithmic orders were false. In truth, Liquidity Profiling

did not apply to orders sent to Barclays’ dark pool via Barclays’ algorithms.

68. Senior Barclays personnel were aware that clients and potential clients were under

the mistaken impression that orders routed to the dark pool via Barclays’ algorithms were

protected by Liquidity Profiling, but did nothing to correct the false impression that Barclays

itself created.

69. In an August 19, 2012 email to several senior division employees (including

Anthony Pallone, Jacek Janczewski, and Sarah Naegele), David Johnsen, Head of Product

Development and second-in-command of the Equities Electronic Trading Division, wrote that

“what is a real issue for institutions is that the Profiling tag doesn’t get fed from the algos . . .

that’s just stupid and certainly something that they assume we are doing” (emphasis added).

70. In written marketing materials sent to clients and potential clients, and in

passive trading.”

71. In pitchbooks sent to clients and potential clients, and in statements posted on

Barclays’ website from 2011 until at least the filing of the initial Complaint in 2014, Barclays

represented to investors that, when routing client orders to various trading venues in the United

States, Barclays ranked trading venues “dynamically,” using “parallel routing to all venues based

on probability of fill.” Barclays claimed that its smart order router “[t]reat[s] all venues the

same based on execution quality.”

72. In response to written questionnaires sent from clients in 2013, Barclays stated

that its router did not preference any venue and did not route to venues based on where Barclays

received payments for orders. For instance, in a July 12, 2013 response to a client’s written

venues based on criteria other than execution quality?” Barclays answered: “No. The Dynamic

SOR is built to preference venues solely on performance metrics” (emphasis added). Barclays

made substantially similar representations in response to other client inquiries.

73. On March 28, 2014, Barclays distributed an internal document to employees that

included “Selling points on our dark pool and order handling practices.” Barclays instructed its

employees to tell clients and potential clients that, among other things:

Development and second-in-charge of the Equities Electronic Trading Division, responded on

behalf of Siffringer, telling the client that Barclays’ smart order router “dynamically adjusts its

behavior order-by-order based on where it is finding liquidity.”

76. With respect to in-person meetings between Barclays and its potential clients, one

former Director in Barclays’ Equities Electronic Division, who attended multiple sales meetings

with clients, recalls that in every sales presentation at which he was present, clients were told by

Barclays employees that “the router was routing based on best execution,” and that it was

79. Barclays’ representations about its how it routed clients’ orders for execution

were false and misleading for several reasons, including:

not work that way] because that was not economically advantageous for Barclays.” In the

words of this Director, Barclays routes trades to its own dark pool “whether it is right or not.”

81. As another former Barclays employee recalled, “what Barclays did was rather

than route [client orders] in a manner that would fulfill [their] needs, that would give best

execution . . . they jammed it all into [Barclays own dark pool].” Another former employee

recalls, “such a high [] internalization rate could suggest that there were other opportunities

being missed and [clients] weren’t receiving the true best executions.”

rebates, and this likely had an impact” on the rates at which client orders were being filled. In

other words, more client trades were not getting filled because Barclays stopped routing to

certain venues that did not offer sufficient financial incentives to Barclays.

84. In 2013, senior directors in the Equities Electronic Trading Division began a

broad analysis of Barclays’ order routing practices, gathering detailed trade data from over 100

million unique trades. Upon analyzing the data, these directors determined that Barclays had an

extremely high “internalization rate” – that is, a high rate of routing client orders into Barclays’

own dark pool. The analysis also determined that certain trading venues were disadvantaged by

Barclays’ routing procedures, either because Barclays was submitting orders that had no chance

of being accepted in that particular venue, or because those venues were not seen as financially

routed unfilled orders (after first having routed them to its own dark pool) tended to be venues

hosted by high-speed trading firms, “[n]one of which,” recalled one Director, “had a reputation

for being favorable to clients from an execution perspective.” Those venues are operated by

Knight Capital, Getco, and Citadel.

85. In October, 2013, Barclays prepared a trading analysis for a major institutional

investor that services millions of individual accounts both inside the United States (including for

many New Yorkers) and abroad (“Institutional Investor”). Barclays knows the identity of this

shared with the client because it reflected poorly on Barclays. Despite the pressure from senior

leadership, this senior director declined to agree to withhold the findings from the Institutional

Investor. The next day, and prior to the scheduled meeting with the Institutional Investor, this

senior director was fired by William White.

89. Another senior director was then instructed to change crucial figures in the

PowerPoint presentation, in order to make them more favorable to Barclays. Specifically, that

director was instructed to change Barclays’ internalization rate for all orders routed to dark

90. In the days after that employee was fired, a third Barclays employee, who was not

involved in the presentation to the Institutional Investor, was tasked with analyzing the relevant

data to determine what percentage of the Institutional Investor’s orders were in fact routed to

Barclays’ own dark pool. In other words, the third employee was tasked with confirming

whether the first two directors’ analysis was correct.

91. On February 6, 2014, the results of this third employee’s analysis were circulated

among various members of the division. As this employee noted in an email, “Data is attached.

[I] just want to note a few things . . .% executed in LX is very high for 2013. LX as a high % of

all dark remains true” (emphasis added). This employee found that for the relevant Institutional

Investors’ orders between 100 and 400 shares, over 90% were executed in Barclays own dark

was removed from the analysis: “Can someone tell me where Tradebot is in Figure 1? It looks

like some of those data points have been moved around.” Martin Lujan forwarded this email to

Sarah Naegele, saying only: “Reprising this old debate.” The analysis was not corrected.

121. On April 14, 2014, an employee in Barclays’ Equities Marketing Division was

tasked with reviewing certain marketing materials provided to institutional investors. With

regard to the document in question, this employee asked Sarah Naegele by email whether it was

in fact accurate that the analysis represented the “Top 100” traders in the dark pool. Rather than

Liquidity Profiling service as having aggressive trading behavior. That is because “Modified

Take %” (the x-axis of the chart) was not used by the Liquidity Profiling service to categorize

traders. Rather, the service used an altogether different metric. The chart therefore

misrepresents both the metrics used by Barclays to classify traders in the dark pool, and the kind

of trading taking place in the pool.

123. Immediately following the filing of the initial Complaint in this matter, Barclays

removed the analysis from its website.

2. Barclays Underrepresented the Amount of “Aggressive” Trading Activity

128. Those figures were false. Contrary to what Barclays represented to clients,

potential clients, and to the investing public, Barclays knew that there was significantly more

“aggressive” activity in Barclays’ dark pool than either 6% or 9%.

129. For instance, in August 2012, Barclays conducted an “Execution Aggressiveness”

analysis of the trading in its dark pool. That analysis, contained in a document circulated via

email to key personnel in the Equities Electronic Trading Division, found that between 25% and

30% of all activity in the dark pool was, in Barclays’ own terminology, “aggressive.”

volume of shares executed were known high frequency trading firms.

132. Those numbers remained consistent over late 2013 and early 2014. For instance,

in March 2014, Barclays was engaged in discussions with a prominent high frequency trading

firm wherein Barclays itself categorized approximately 25% percent of the orders taking

liquidity in its dark pool as aggressive. In an internal document collecting the information

received from Barclays, that firm summarized the data provided to it by Barclays, and concluded

that the trading activity in Barclays’ dark pool was “50% good, 50% aggressive.”

133. In May 2014, Barclays’ own analysis again determined that over 30% of all

activity in the dark pool was, in Barclays’ own words, “Aggressive.”

136. Barclays’ Liquidity Profiling service ostensibly worked by grouping traders in the

“4” and “5” categories were those traders conducting the safest, most passive, long-term-

investor-like trading activity. Participants in Barclays’ dark pool were told that they could

disable their orders from interacting with traders falling into any of the various categories – in

particular, clients could opt-out of trading with those counterparties identified by the Liquidity

137. In a July 31, 2012 “Sales Trader Education” seminar given to Barclays sales

personnel by Jacek Janczewski, head of dark pool operations, and Anthony Pallone, head of

Sales for the division, Barclays’ sales staff was instructed to tell clients and potential clients that

Liquidity Profiling uses “objective criteria” to “proactively monitor [the dark pool]” and that

Barclays “regularly evaluate[s] client profiles.” Janczewski and Pallone instructed sales staff to

represent to clients and potential clients that Barclays was “effectively ‘policing’ the pool.”

138. Janczewski and Pallone further instructed sales staff to represent to clients and

LX on a daily basis,” and “runs surveillance reports every week to make sure that there is no

toxic flow in the book” (emphasis added). Barclays made that particular representation in

written documents provided to clients on more than a dozen occasions in 2012 and 2013.

140. In a range of other documents and communications from 2012 until at least the

filing of the initial Complaint in 2014, Barclays represented to clients, potential clients, and the

public that Liquidity Profiling allowed it to “hold [traders] accountable” if their trading was

“aggressive,” “predatory,” or “toxic.” For instance:

• In pitchbooks provided to clients and potential clients in 2013 and 2014, as well as in

• On March 8, 2013, the online magazine Markets Media published an interview with

• On June 6, 2013, online magazine Hedgeweek published an interview with White,

141. Similarly, in a presentation made to an audience on June 18, 2012, and which is

Handling Practices.” The document, marked “For Internal Use Only,” notes that given “recent

press on market structure [and the] NY Attorney General’s comments on HFT . . . we have an

opportunity to engage with clients about how we think about electronic trading and best

execution” (emphasis omitted). Schlanger instructed employees to represent to clients that

Barclays “[i]nvented Liquidity Profiling to police trading behaviour” in the dark pool, and to tell

clients that Barclays “can deny access to predatory participants.” Schlanger instructed

employees to represent that “Liquidity Profiling encourages passive/benign liquidity provision,”

and that Liquidity Profiling “manages toxicity within the pool without limiting access to

potentially beneficial liquidity.”

143. That message was repeatedly conveyed to clients, potential clients, and the

investing public by senior Barclays personnel in written and oral communications from 2012 to

Head of Equities in the Americas distributed “Internal Q&A and Talking Points on Barclays

Electronic Trading and Order Handling Practices.” That document notes that clients might ask

“Why has LX grown so much over the last 3 years?” In response, Barclays urged its Sales staff

to note that “Liquidity Profiling has improved quality of flow, giving clients more confidence to

interact with LX.”

145. Barclays believed the above representations to be material to clients and potential

clients.

146. Barclays’ claims about Liquidity Profiling were false. In truth, the service offered

little or no benefit to Barclays’ clients, because (i) Barclays did not actually police or punish bad

trading behavior; (ii) Barclays failed to regularly update the profiles of traders in its dark pool;

the dark pool; and (v) senior Barclays employees knew that the service was of little benefit to

institutional investors.

147. Contrary to Barclays’ claim that it uses Liquidity Profiling to “police” its dark

pool and would “refuse a client access” if that trader’s activity became toxic, and contrary to its

representations to dozens of clients throughout 2012 and 2013 that it “monitor[s] client flow in

149. For example, on January 16, 2014, senior leaders in the Equities Electronic

trading firms engaged in significant trading activity in Barclays’ dark pool. That analysis

Hudson River Trading, was described in the analysis as “historically . . . very toxic.” Another

firm, Sun Trading LLC, was described as having “[trading activity that] is very toxic, and the

client is up-front about this.” Another firm, Getco, “is in [the most toxic] bucket 0 every week

and will very likely stay there for a long time.” GTS Securities LLC was described as having

“[k]nown latency arbitrage flow” in the pool (emphases added). Barclays never denied any of

those firms (or any others) access to its dark pool.

150. In its memorandum of law in support of its motion to dismiss the initial

Complaint in this action, Barclays asserted to the Court that White’s June 18, 2012 claim to have

“sat down” with multiple clients and “shut them off from the firm” was true. Barclays asserted

that in the days before White’s comments, Barclays had in fact shut off a high-frequency trading

firm for abusive trading practices. Barclays appended redacted documents to its brief, purporting

to establish that Barclays prohibited that firm from trading in the pool.

151. Barclays’ assertions in its brief were false.

by approximately 1/3” on the next trading day, Monday, June 11, 2012.

153. GTS never was never prohibited from trading in Barclays’ dark pool for any

period of time. According to an internal analysis conducted by Barclays in July 2012, GTS

continued to trade millions of shares in the dark pool on every day trading day following the

5

Barclays wrote: “[the NYAG’s] Opposition challenges a statement from [White’s June 8, 2012]

presentation that ‘where we see suspect activity we’ve gone out proactively to sit down with

analysis, showing that GTS traded millions of shares in Barclays’ dark pool on every trading day

during the month immediately following June 8, 2012:

154. Also, in a letter dated May 2, 2014, responding to a subpoena issued by the Office

of the Attorney General, Barclays stated that “Barclays has not prohibited any firm from

participating in LX.”

155. As such, White’s comments on June 18, 2012, and Barclays’ assertions in its brief

about removing GTS from the pool, were false.

156. In addition, senior Barclays employees admitted to each other privately that

Barclays was not in fact “policing” the dark pool in the manner represented to the public.

157. In an email dated August 22, 2012, Jacek Janczewski, head of Barclays’ dark pool

operations, stated to others in the division that while “the strongest selling point to institutions . .