BAILII
British and Irish Legal Information Institute


Freely Available British and Irish Public Legal Information

[Home] [Databases] [World Law] [Multidatabase Search] [Help] [Feedback]

United Kingdom Upper Tribunal (Tax and Chancery Chamber)


You are here: BAILII >> Databases >> United Kingdom Upper Tribunal (Tax and Chancery Chamber) >> Gonzalez & Ors v Financial Conduct Authority (FINANCIAL SERVICES - market abuse - placing and cancellation of "large" orders in BTP Futures on the Eurex Exchange which overlapped with small orders placed by the applicants on the opposite side - whether "large" orders give false or misleading impression or signal to other market participants - whether placed as part of an abusive scheme to facilitate execution of small orders or a legitimate trading strategy being pursued) [2025] UKUT 214 (TCC) (01 July 2025)
URL: http://www.bailii.org/uk/cases/UKUT/TCC/2025/214.html
Cite as: [2025] UKUT 214 (TCC)

[New search] [Contents list] [Printable PDF version] [Help]

Neutral Citation Number [2025] UKUT 214 (TCC)

 

 

UT (Tax & Chancery) Case Number: UT-2022-000134
UT-2022-000135
UT-2022-000137

 

Upper Tribunal

(Tax and Chancery Chamber)

 

FINANCIAL SERVICES - market abuse - placing and cancellation of “large” orders in BTP Futures on the Eurex Exchange which overlapped with small orders placed by the applicants on the opposite side - whether “large” orders give false or misleading impression or signal to other market participants - whether placed as part of an abusive scheme to facilitate execution of small orders or a legitimate trading strategy being pursued - held - references against prohibition orders dismissed - amount of penalties determined

 

Hearing venue: The Rolls Building

London

EC4A 1NL

                                                                                                           

                                                                                                                       

 

Heard on: 27 to 31 January 2025, 3 to 7, 11 to 14 February 2025

Judgment date: 01 July 2025

 

 

 

Before

 

JUDGE JEANETTE ZAMAN

MEMBER GARY BOTTRIELL

MEMBER PETER FREEMAN

 

Between

 

(1)  MR JORGE LOPEZ GONZALEZ

(2) MR POOJAN SHETH
(3) MR DIEGO URRA

 

 

Applicants

and

 

THE FINANCIAL CONDUCT AUTHORITY

 

Respondent

 

 

Representation:

 

For Mr Urra, Andrew George KC and Ava Mayer, counsel, instructed by Mishcon de Reya LLP

 

For Mr Lopez, Ben Jaffey KC and Celia Rooney, counsel, instructed by Stephenson Harwood LLP

 

For Mr Sheth, Alex Bailin KC and Jason Mansell, counsel, instructed by Bryan Cave Leighton Paisner LLP

 

For the Authority:  Sharif Shivji KC, Lara Hassell-Hart and Nicholas Wright, counsel, instructed by The Financial Conduct Authority

 


Table of Contents

 

Table of Contents. 1

Introduction and summary. 3

Decision Notices and Authority’s amended statements of case. 4

Recklessness. 10

Traders’ Replies and outline of trading strategies relied upon. 11

Relevant legal principles. 11

Market Abuse. 12

Dishonesty. 15

Role of the Tribunal 15

Non-disciplinary references. 16

Disciplinary references. 17

Burden and Standard of proof 17

Evidence including witnesses who had not been called, information that is no longer available and relevance of delay. 18

Outline of evidence before the Tribunal 18

Pace of Authority’s investigation and particularisation of its case. 20

Lack of information that would have been available to the Traders during the Relevant Period. 22

Passage of time, memory and witness evidence. 27

Potential witnesses who were not called by the Authority. 30

Approach of the Tribunal 36

Findings of fact 37

EGBs, market making, BTPs and BTP Futures. 37

The Traders - roles at MHI and experience. 39

MHI and the EGB Trading Desk. 41

Eurex Letter 48

Interviews with Compliance. 50

Investigation by MHI Compliance. 51

MHI disciplinary process. 55

Interviews by the Authority. 55

Traders’ explanations of rationale for the Large Orders. 56

Trading Activity of the Traders in the Relevant Period. 59

Trading Activity of the Traders outside the Instance Pool 64

Trading Activity of other participants in the market 65

Market abuse. 66

Evaluation - Whether Large Orders are likely to impact the market 66

Tribunal’s assessment of the Experts. 67

Summary of evidence of Mr Creaturo. 69

Market liquidity. 70

Whether Large Orders may influence other market participants. 73

Conclusions on market impact 75

Evaluation - Whether traders committed market Abuse. 77

Criteria used to identify the Instance Pool 80

The Trading Strategies - contemporaneous explanations. 84

Mandate. 87

Information Discovery Strategy - plausibility. 89

Information Discovery Strategy - operation. 95

Conclusions on the Information Discovery Strategy. 104

Anticipatory Hedging Strategy - plausibility. 104

Anticipatory Hedging Strategy - operation by Mr Lopez. 113

Conclusions on the Anticipatory Hedging Strategy. 117

Placing of concurrent Large Orders. 118

Collaboration. 121

Plausibility of Authority’s case that the Traders conducted an abusive scheme. 130

Trading Activity of the Traders in the Relevant Period. 138

Conclusions on Market Abuse. 147

Prohibition orders. 160

Penalties. 161

Step 1: Disgorgement 162

Step 2: The seriousness of the breach. 162

Step 3: Mitigating and aggravating factors. 163

Step 4: Adjustment for deterrence. 164

Step 5: Settlement Discount 164

Authority’s determination of the penalties to be imposed. 164

Assessment of the financial penalty. 164

Mr Urra. 165

Mr Lopez. 168

Mr Sheth. 169

Disposition. 172

Directions. 172

Appendix 1. 174

Statement of agreed background facts. 174

The Cash BTP Market 174

RFQs and cash trades. 175

Hedging and trading BTP futures on EUREX.. 175

MHI and the EGB Desk. 176

Glossary. 178

Appendix 2. 182

 

 


 

 

DECISION

Introduction and summary

1.       On 31 October 2022 the Financial Conduct Authority (the “Authority”) through its Regulatory Decisions Committee (“RDC”) issued Decision Notices to each of Diego Urra (“Mr Urra”), Jorge Lopez Gonzalez (“Mr Lopez”) and Poojan Sheth (“Mr Sheth”) (together, the “Traders”).  In those Decision Notices the Authority decided to make orders prohibiting each of the Traders from performing any function in relation to any regulated activity carried on by an authorised person, exempt person or exempt professional firm pursuant to s56 Financial Services and Markets Act 2000 (“FSMA 2000”), and imposed penalties on each of them pursuant to s123(1) FSMA 2000.

2.       Each of the Traders referred their respective Decision Notices to the Tribunal.  This decision concerns the subject matter of those references, which were heard together.

3.        The Traders were traders at Mizuho International Plc (“MHI”).  The Authority’s case is that the Traders pursued an abusive trading strategy in relation to Long-Term Italian Government Bond Futures (“Futures” or “BTP Futures”) on 233 occasions (between the three of them) between 1 June 2016 and 29 July 2016 (the “Relevant Period”).  This abusive strategy is alleged to have involved placing orders of at least 200 lots (the “Large Orders”) which gave the false or misleading impression and/or signal that the Traders wanted to buy or sell a specified number of Futures, when in fact their objective was to facilitate the execution of other (genuine) smaller orders (generally of fewer than 200 lots) on the opposite side of the order book (the “Small Orders”).  The Authority’s case is that this constituted market abuse within s118 FSMA 2000 and market manipulation within Article 12 of Regulation (EU) No 596/2014 (the “Market Abuse Regulation”).  The Tribunal uses “market abuse” to refer to both market abuse and market manipulation within these regimes.  

4.       Each of the 233 occasions is referred to as an “Instance”, ie that is the term used to capture the trading activity on which the Authority relies.  The 233 Instances are together the “Instance Pool”, those Instances involving orders placed by one Trader are “Single Trader Instances”, and those where orders (whether Small Orders or Large Orders) are placed by more than one of the Traders are “Multi Trader Instances”.  There are a total of 341 Large Orders in the Instance Pool.

5.       The Traders do not dispute the details of the activity within the Instances.  They deny they committed deliberate market abuse, and submit they did not act dishonestly or lack integrity, as alleged by the Authority.  The Traders each say that they placed each large order (ie the Large Orders and other large orders outside the Instance Pool) in pursuit of a legitimate trading strategy, they intended to trade their Large Orders and there was no collaboration (such that the Authority’s approach of identifying and relying on Multi Trader Instances is misguided).  They have put forward their explanations for their trading strategy (for Mr Urra and Mr Sheth, the Information Discovery Strategy, and for Mr Lopez, the Anticipatory Hedging Strategy) and sought to explain their trading by reference to available evidence (and drawing attention to information that is no longer available). 

6.       The parties have produced a Statement of Agreed Background Facts (“SOABF”) which is set out at Appendix 1 and forms part of this decision.  We use the terms as defined in the SOABF in this decision.  The SOABF covers the cash BTP market, client RFQs, hedging and trading in Futures on the Eurex Exchange (“Eurex” or “the Exchange”), MHI and the EGB Desk, and includes a glossary.     

7.       On terminology in relation to the orders placed by the Traders, the parties have at various times used different defined terms for what we have referred to above as the Small Orders and Large Orders (which was adopted by the Authority and Mr Bailin by the time of the hearing, with Mr George and Mr Jaffey referring to the Large Orders as “Larger Orders” and “Relevant Orders” respectively).  The Authority had previously defined these as the “Genuine Orders” and the “Misleading Orders”, and that is the language used in their statements of case (and subsequent amendments thereto).  The Tribunal recognises that the terms Small Orders and Large Orders are not neutral - the Traders’ position is that the orders which are relied upon by the Authority as having given a false impression and/or signal, whether of 200, 450 lots or 500 lots, are not “large” but medium-sized. 

8.       The Tribunal has recognised throughout that whilst the Authority’s case is that the Traders individually and acting together committed market abuse by placing orders they did not wish to trade but whose purpose was to facilitate the execution of the Small Orders, we need to reach a conclusion in relation to each of the Traders.  We have taken account of and assessed the activity of each trader individually as well as alongside that of the other Traders in the Instances when assessing their explanations for their trading. 

9.       For the reasons set out below, the Tribunal has concluded that the Traders each deliberately engaged in market abuse, and that this conduct was dishonest and lacked integrity.  The Traders’ references against the prohibition orders are dismissed.  The Tribunal determines that the appropriate action for the Authority to take is to impose a financial penalty of the following amount on each of the Traders and the references are remitted to the Authority with the direction that effect be given to our determinations:

(1)        Mr Urra - £223,400;

(2)        Mr Lopez - £100,000; and

(3)        Mr Sheth - £57,600.

10.    The Tribunal is grateful to the parties for their written and oral submissions.  We have not found it necessary to refer expressly in this decision to all of the evidence and submissions but we have taken them all into account.   

11.    There was, as is to be expected, some overlap between the submissions made on behalf of each of the Traders, albeit that some submissions were distinct (eg the difference in the trading strategies between Mr Urra and Mr Sheth on the one hand and Mr Lopez on the other, the size and number of Large Orders placed by Mr Lopez, the issue of Mr Sheth’s level of experience and the submissions in relation to Mr Sheth’s placing of multiple overlapping Large Orders).  We do specifically refer to some of the submissions as having been made by counsel for one of the Traders; we do not lose sight of the fact that the same or a similar submission was also made on behalf of another Trader. 

Decision Notices and Authority’s amended statements of case

12.    The Authority issued Decision Notices to each of the Traders on 31 October 2022. 

13.    Whilst the Authority relied on a different number of Instances for each Trader, the basis of the Authority’s decision was the same for each Trader, being that:

(1)        The Traders undertook an abusive trading strategy in Futures on the Eurex Exchange, alone and in collaboration with each other.  They would place a large sized order on one side of the order book for the purpose of creating the impression of increased supply or demand, with the objective of assisting the execution of a smaller genuine order they wished to trade on the opposite side of the order book.  Once the smaller genuine order had been executed, they would cancel the large order.  Through the placement of these large misleading orders the Traders falsely represented to the market an intention to buy or sell when their actual intention was the opposite.  The only purpose of the large orders was to assist the execution of the smaller genuine orders that the Traders wanted to trade.  The abusive trading strategy was such that it was unlikely the large misleading orders would themselves trade; notably, they were placed away from the touch and were quickly cancelled.

(2)        This conduct gave false and misleading signals (there being no material difference between this concept and that of a false and misleading impression) to the market as to demand and supply.  It amounted to market manipulation.  The Traders frequently repeated this pattern of abusive conduct during the Relevant Period. 

(3)        The Traders knew that placing large orders on the opposite side of the book to assist the execution of other orders they or another Trader genuinely wanted to trade would result in false and misleading signals to the market; and they knew that this would be likely to impact the trading activities of other market participants.  Their conduct constituted deliberate, intentional and repeated market manipulation and was dishonest.    

14.    The Authority considered that the conduct of each Trader in deliberately engaging in market manipulation was dishonest and lacked integrity.  This dishonest conduct was highly likely adversely to impact other market participants and was repeated many times over a period of two months.  As a result, they were each not a fit and proper person to perform any function in relation to any regulated activity carried out by an authorised person, exempt person or exempt professional firm.

15.    The Authority decided to make orders prohibiting each of the Traders from performing any function in relation to any regulated activity carried on by an authorised person, exempt person or exempt professional firm, and imposed penalties on each of them: £395,500 on Mr Urra, £100,000 on Mr Lopez and £100,000 on Mr Sheth. 

16.    The Authority’s amended statements of case for each Trader are dated 6 December 2024 (together, the “Amended Statements of Case”) and are substantially similar to each other, in that the Authority relies on the same features as supporting its allegations of an abusive trading strategy in respect of the Large Orders in the Instance Pool and then identifies the number of occasions on which each Trader is alleged to have committed market abuse.    

17.    The number of occasions on which each Trader is alleged to have committed market abuse, and the number of Large Orders they placed on which the Authority relies are as follows:

 

Single Trader Instances

Multi Trader Instances

No of Large Orders in Instance Pool

Mr Urra

31

97

135

Mr Lopez

41

84

55

Mr Sheth

47

57

151

 

18.    The Amended Statements of Case proceed by identifying the “Facts and Matters relied upon by the Authority” and starts by setting out matters under the sub-headings of “The BTP and BTP Futures Market”, “The EGB Desk and its Mandate”, “BTP Trading and the EGB Desk’s process”, “Trader Remuneration” and “Training and Awareness of Market Abuse”.  The Amended Statement of Case in respect of Mr Urra’s reference then sets out the following:

“Abusive Trading Strategy

58.  During the Relevant Period, Mr Urra undertook an abusive trading strategy, both alone, and in collaboration with Mr Lopez and/or Mr Sheth.  The abusive strategy involved placing orders for the purpose of giving the false or misleading impression and/or signal that the Traders wanted to buy or sell a specified number of BTP Futures lots (the “Misleading Orders”), when in fact the Traders did not intend to trade these orders, but instead intended or hoped to facilitate the execution of other genuine orders on the opposite side of the order book (the “Genuine Orders”). Misleading Orders which appeared to increase supply would encourage other market participants who wanted to sell to cross the spread and trade with existing buy orders in anticipation of the market moving lower (and vice versa for Misleading Orders which appeared to increase demand and might prompt the market to move higher).

…

61.  That the trading carried out by Mr Urra (alone or in collaboration with Mr Sheth and/or Mr Lopez) gave or was likely to give a false or misleading impression and/or signal as to the supply of, demand for, or price of, BTP Futures, and was therefore abusive, is to be inferred from one or more of the following features:

61.1  The Misleading Orders were orders of BTP Futures showing 200 lots or more on the Exchange (“Large Orders”). Large Orders were rarely placed on the Exchange (see further paragraph 63.1 below) and were large compared to the Genuine Orders. They therefore gave or were likely to give an impression and/or signal of significant supply/demand to other market participants on the opposite side of the order book from the Genuine Orders.

61.2  None of the Misleading Orders were placed as Iceberg Orders, whereas some of the Genuine Orders (despite typically being significantly smaller) were placed as Iceberg Orders. By showing the full size of the Misleading Orders, this gave or was likely to give an impression and/or signal of significant supply/demand to other market participants on the opposite side of the order book from the Genuine Orders.

61.3  The vast majority of Misleading Orders were placed near enough to the Best Bid or Best Offer to be visible to other market participants (and thereby increase the pressure on the order book) but not so close that they were likely to be traded.

61.4  None of the Misleading Orders fully executed, and only 3 partly executed. In those 3 cases, only a very small minority of the order traded (under 10%) before being cancelled. Moreover, in each instance where the Misleading Orders began to trade, they were cancelled under 10 seconds later. By contrast, in each trading instance particularised in Amended Annex I, the Genuine Orders together mostly or fully executed prior to the cancellation of the Misleading Orders; the overall execution rate across the total volume of Genuine Orders before the cancellation of the Misleading Orders was 95%.

61.5  The Misleading Orders always overlapped with at least one Genuine Order on the other side of the book. Save for one instance, the last Misleading Order(s) on the order book in a particular instance were cancelled after the majority of the Genuine Order had filled, on average 5 seconds after.

61.6 The Traders have not matched executed BTP trades or RFQs against the placing of their Misleading Orders in the BTP Futures market.

61.7  The volumes of the Misleading Orders were particularly large sizes, frequently placed in round numbers and/or repeated numbers of lots. Of the 36 Misleading Orders placed by Mr Urra acting individually across 31 instances, for example, 13 were lots of 444, 9 were lots of 450, 5 were lots of 490, and 3 were lots of 400. It is inherently unlikely when placing genuine orders for hedging purposes that there would be so many identical orders, since market-makers will seek to hedge the precise residual risk on their books from time to time.

61.8  The Misleading Orders were not netted off against the Genuine Orders on the other side of the book (for example, if the Genuine Order was to buy 10 lots and the Misleading Order was to sell 500 lots, they could have placed a sell order of 490 lots, but did not do so). It was more costly to trade on both sides than to net off, and in addition there was potential additional risk from movements in the price from failing to net off.

61.9  The volume and repeated pattern of these types of orders over the Relevant Period by all the Traders (alone and in conjunction with one another) indicates a deliberate strategy.

62.  In relation to each of the occasions where Mr Urra is averred to have carried out the strategy in collaboration with Mr Lopez and/or Mr Sheth…, it is inferred from the contemporaneous activity of the other Traders in each of the instances specified and the matters pleaded at paragraph 55 above that they were working together to a common strategy.

63. That the Traders’ conduct was abusive is also to be inferred by comparing their behaviour with the trading and order placement of BTP Futures on the Exchange by other market participants during the Relevant Period. The former was markedly different from the latter:

63.1  Large Orders of BTP Futures were rarely placed on the Exchange by other market participants. Including MHI, 47 market participants placed Large Orders during the Relevant Period, accounting for 0.02% of the total number of orders placed on the Exchange during the Relevant Period.

63.2  Despite MHI being a small market player, trading less than 0.43% of the total traded volume of BTP Futures, the Traders placed more Large Orders than any of the other market participants and accounted for 23.24% of the total volume of Large Orders placed across the Relevant Period. However, the Traders rarely executed BTP trades or received client orders in BTPs that they could have wanted to hedge with these Large Orders.

63.3  While … the Traders placed significantly more Large Orders than other market participants, they had much lower execution rates of their Large Orders. The Traders partially or fully executed only 1.5% of their Large Orders, cancelling 98.5% without them having begun to execute (Mr Urra himself executed 0.8% of the total volume of Large Orders that he placed). By comparison, other market participants partially, or fully executed 72.28% of their Large Orders, cancelling only 27.72% of their Large Orders without them having begun to execute. It would likely have been possible for the Traders to execute a larger proportion of their Large Orders if it had been their intention for them to execute.

63.4  When the Traders placed Large Orders, they rarely priced them competitively, placing only 1.93% of them at the Best Bid or Best Offer price. In contrast, other market participants placed 80.34% of their Large Orders at Best Bid or Best Offer prices, or at improved prices. By placing their Large Orders away from the Best Bid or Best Offer price, the Traders were less likely to execute them.

…”

19.    We refer to the features set out by the Authority at [61] in its Amended Statement of Case above as the “Criteria”.  These Criteria were used by the Authority to identify the Instance Pool from the totality of the trading activity of the Traders during the Relevant Period. 

20.    The Authority then disputes the explanation provided by Mr Urra for his trading strategy pleading that it is “inherently improbable” and particularises its case using F150 as an example of a Single Trader Instance, and F31 and F209 as Multi Trader Instances before then setting out its position on the alleged breaches:

“Section 118 of FSMA

…

104. The Misleading Orders were not placed for legitimate reasons, nor did they conform with accepted market practice. In placing the Misleading Orders, Mr Urra gave or was likely to give a false or misleading impression as to the supply of, or demand for, the BTP Futures to which the Misleading Orders related. This was because in placing the Misleading Orders, Mr Urra signalled that he wanted to buy or sell a specified number of BTP Futures. In fact, he did not wish to trade in that manner and the purpose of placing the Misleading Orders was to facilitate the execution of Genuine Orders at a more advantageous price, or on a more timely basis, than would otherwise have been achieved but for his having misled other market participants by the Misleading Orders.

…

Articles 12 and 15 of the Market Abuse Regulation

…

107. Mr Urra’s misleading orders were not placed for legitimate reasons, nor did they conform with an accepted market practice as established in accordance with Article 13 of the Market Abuse Regulation. In placing the Misleading Orders (alone or in concert with Mr Lopez and/or Mr Sheth), Mr Urra gave or was likely to give a false or misleading signal as to the supply of, or demand for, the BTP Futures to which the Misleading Orders related. This was because in placing the Misleading Orders (alone or in concert with Mr Lopez and/or Mr Sheth), Mr Urra signalled that he wanted to buy or sell a specified number of BTP Futures. In fact, he did not wish to trade in that manner and the purpose of placing the Misleading Orders was to facilitate the execution of Genuine Orders at a more advantageous price, or on a more timely basis, than would otherwise have been achieved but for his having misled other market participants by the Misleading Orders.

…

Fitness and Propriety

109. Mr Urra’s conduct in deliberately engaging in market manipulation was dishonest and lacked integrity. This dishonest conduct was highly likely to adversely impact other market participants and was repeated many times over a period of two months. As a result, he is not a fit and proper person to perform any function in relation to any regulated activity carried out by any authorised person, exempt person or exempt professional firm.”

21.    The Authority has pleaded its case on a substantially similar basis in respect of Mr Lopez and Mr Sheth, and in particular has relied on the Criteria and the comparison with trading and order placement by other market participants.  In its Amended Statements of Case for Mr Lopez and Mr Sheth the Authority has used examples of Instances relevant to each Trader.  We do not set those out here.

22.    In addition to the number of Instances relied upon against each Trader, the only point of substantive difference in the Authority’s decision and pleadings relates to Mr Sheth, who had placed multiple concurrent Large Orders on the same side in some of the Instances (the “Multiple Large Orders”).  The Authority alleged that each of these are Large Orders and its Amended Statement of Case in respect of Mr Sheth, having said his explanation for his trading strategy is “inherently improbable” sets out the following:

“64. …Mr Sheth has also suggested that in some cases, where he placed multiple, overlapping Large Orders, these were in fact intended to be amendments to an existing Large Order.  Whether or not this is true does not affect the Authority’s assessment of the purpose of the placement of a Large Order on the opposite side of the order book to a small order.”    

23.    The 233 Instances have been set out in Annex 1 to the Amended Statements of Case as lists of the trading activity which occurred in each Instance (this Annex being the same for each Trader, albeit that the Authority does not rely on each Instance against each Trader). 

24.    Whilst the Authority has amended its statements of case since they were first served on 24 February 2023, and significant changes have been made to the level of detail provided in relation to the specific trading activity relied upon, the allegations set out above, in particular as to an “abusive trading strategy”, “deliberately engaging in market manipulation” and “dishonest and lacked integrity” have been pleaded throughout.

25.    The Authority has further particularised its case since it first served its statements of case, and, since February 2024, has particularised each of the Instances.  We have set out in Appendix 2 the Authority’s particularisation of one instance (F7, the first Specified Instance) to illustrate the data provided (but it should be noted that the narrative was only provided for the “Specified Instances”, namely those Instances which the Authority was directed by the Tribunal in June 2024 to identify to the Traders before the hearing and on which it would cross-examine the Traders).  Appendix 2 also includes the daily price graph, Electronic RFQs submitted to the Desk and the executed cash bond transactions for that day, 7 June 2016. 

26.    By way of further explanation, for each Instance the Authority provided:

(1)        overview table - A summary of the relevant Instance, identifying the Trader(s) involved and the date and time of the Instance;

(2)        table of all trading activity by the Traders between the identified start and end time of the Instance relied upon by the Authority - Where an order had been iceberged, the placing of the subsequent slices are shown as a new order (and we refer to an order of, eg, nine lots which is iceberged to show slices of three lots as “iceberged to three”).  Times of day in the trading activity are set out in Continental European Time (“CET”), and that is used throughout this decision notice;

(3)        “Spread graph” - This shows the Best Bid and Best Offer, highlighting the Traders’ orders and trades.  These graphs were produced from Eurex data using the Authority’s data analysis system.  Best Offer is shown by a red line, Best Bid by a green line, order entry is a circle, order amendment a diamond and order cancellation a cross.  Where an order trades, that is shown by a solid circle.  The grey shaded columns are levels of traded volumes; and

(4)        “Replay graph” or “stack” - This shows the (visible) orders on Eurex at the point of entry of the Large Order, showing the size of orders and the Bid/Offer prices.  Where an Instance includes more than one Large Order, there is a separate Replay graph for each Large Order entered.  These graphs were also produced from the Eurex data using the Authority’s data analysis system.  The offers on the buy side from other market participants are shown in green on the left-hand side, with the Best Bid at the top.  The offers on the sell side from other market participants are shown in red on the right-hand side, with the Best Offer at the top.  The Large Order on which the Authority relies is grey.  Some Replay graphs had bars in blue (orders to buy) or yellow (orders to sell), and those are existing orders which had been placed by MHI on the book at that time.  Where the Authority relied on more than one Large Order in an Instance, the most recent would be shown in grey but existing Large Orders would then be shown in blue or yellow (as applicable).

Recklessness

27.    The basis on which the Authority has pleaded its case against the Traders is set out above.  The Authority has not pleaded in its Amended Statements of Case that the Traders’ conduct was reckless, nor has it applied to the Tribunal to amend further its Amended Statements of Case to plead recklessness in the alternative.

28.    The Authority served its written skeleton argument on 20 January 2025 in accordance with the Tribunal’s directions, and that included the following:

“42.  In order to establish market abuse, the Authority does not need to go further than that and show that the Traders knew that they were committing market abuse or that they were acting in a way which was dishonest or reckless.

43.  Whether the Traders behaved recklessly or dishonestly is however a consideration which arises in relation to the prohibition order … as well as being relevant to the level of penalty…”

29.    Then, in a footnote to the Authority’s summary of the principles relevant to the imposition of a penalty, the skeleton included:

“FN147.  For the avoidance of doubt, the Authority does not positively say that the Applicants’ conduct was reckless in circumstances where it is alleged to have been deliberate and dishonest. If, however, under its broad powers under s. 133(5), the Tribunal finds that market abuse is made out against the Applicants, but that their conduct fell short of being dishonest, the Tribunal will nevertheless have to address whether to apply a penalty against the Applicants and, if so, at what level. In order to address that question, it will have to consider the nature of their conduct, including whether it was reckless, on any view.”  

30.    In their opening submissions Mr George, Mr Jaffey and Mr Bailin each objected to this inclusion of a reference to recklessness in the Authority’s skeleton on the basis that the Authority has throughout imposed sanctions and pleaded its case on the basis that the Traders’ conduct was deliberate and dishonest.  The parties made submissions in closing as to the jurisdiction of the Tribunal by reference to the decision of the Court of Appeal in Financial Conduct Authority v Bluecrest Capital Management (UK) LLP [2024] EWCA Civ 1125 and the Authority’s obligations under The Tribunal Procedure (Upper Tribunal) Rules 2008 (the “Upper Tribunal Rules”), in particular under paragraph 4(2)(c) of Schedule 3 that the Authority’s statement of case “set out all the matters and facts upon which the respondent relies” as well as to the principles relevant to the exercise of the Tribunal’s discretion to allow amendments and the Tribunal’s role under s133(4) to (7) FSMA 2000 on a reference.

31.    On the basis of our decision on the Authority’s pleaded case, the Tribunal has not needed to address these issues further and does not do so.

Traders’ Replies and outline of trading strategies relied upon

32.    The Traders have each denied being engaged in an abusive scheme.  The Traders submit that they intended to trade each of the orders they entered and that the Large Orders (as well as all of the large orders which are outside of the Instance Pool) were placed pursuant to a legitimate trading strategy.  

33.    The strategies relied upon by the Traders were, in outline, as follows (this description being based on their Replies and the Traders’ subsequent witness statements):

(1)        Mr Urra’s evidence was that he was pursuing what we refer to as an “Information Discovery Strategy” - MHI was a small market player, with less visibility over market activity and had an information disadvantage.  He believed that clients might split a large order across more than one market maker and not tell each one that a larger order was being worked.  By withholding the size of the total order, clients could benefit from more favourable pricing, before the market makers recognised the true extent of the trading activity.  The risk to the EGB Desk was that it would not have a complete understanding of the price request and the market, and as a result the EGB Desk would misprice the trade.  In this situation, other market makers, dealing with a possibly much larger part of the same client order, might now be exposed to significant risk and be looking to hedge their position promptly.  He tried to test ways to exploit that situation, and theorised that what he described as a medium-sized order for Futures offered at a premium might be tempting for such a market maker.  He had practical experience of such trades having occurred previously.  In anticipation of demand, he placed a medium-sized order for Futures, with a direction based on the direction he expected the market maker would want to trade, and selected its price at a level he thought would be of interest to the market maker but represented a good deal for him if the order was transacted.

(2)        Mr Sheth’s evidence was that he was also pursuing this “Information Discovery Strategy”, which he said had been shown to him by Mr Urra in May 2016 when Mr Urra had placed an order (on 12 May 2016) of 450 lots in Futures which filled.  Mr Urra had explained to him that there were two reasons for the strategy - to obtain information as to whether there was hidden liquidity in the cash market and to create profitable positions - and had encouraged him to try it.  Mr Sheth explained the Multiple Large Orders he placed as being “poor practice” and a mistake, arising as a result of his inexperience and lack of training in operating Bloomberg Escalator. 

(3)         Mr Lopez’s evidence was that he was pursuing an “Anticipatory Hedging Strategy”.  He was looking ahead to what he predicted to be likely client demand, ie RFQs from clients, with particular emphasis on those of medium-size, and seeking to position his trading book and inventory of bonds to service that demand and be able to offer competitive prices.  This was a constant process of assessing the levels at which anticipatory hedges would be attractive, relative to the RFQs he expected the Desk to receive.

34.    We refer to the Information Discovery Strategy and the Anticipatory Hedging Strategy together as the “Trading Strategies”, recognising that these are strategies that the Traders say there were pursuing and that the Authority does not accept that they were pursuing any such strategies:  the Authority submitted they are fabrications designed to fit the facts which are now known.

Relevant legal principles

35.    We set out here the relevant legal principles in relation to market abuse and dishonesty.

Market Abuse

36.    The Relevant Period straddles two legislative regimes.  Prior to 3 July 2016, s118 FSMA 2000 contained the relevant statutory provisions relating to market abuse.  The Market Abuse Regulation came into force with effect from 3 July 2016. 

37.    Until 2 July 2016, s118(1) FSMA 2000 provided:

“For the purposes of this Act, market abuse is behaviour (whether by one person alone or by two or more persons jointly or in concert) which - 

(a) occurs in relation to – 

(i)  qualifying investments admitted to trading on a prescribed market, 

(ii) qualifying investments in respect of which a request for admission to trading on such a market has been made, or 

(iii) in the case of subsection (2) or (3) behaviour, investments which are related investments in relation to such qualifying investments, and 

(b) falls within any one or more of the types of behaviour set out in subsections

(2) to (8). …”

38.    “Prescribed market” and “qualifying investments” were defined by Articles 4 and 5 respectively of the Financial Services and Markets Act 2000 (Prescribed Markets and Qualifying Investments) Order 2001/996.  The Eurex Exchange was a prescribed market and Futures were a qualifying investment.

39.    Sections 118(2) to (8) of FSMA made provision as to the types of behaviour that constituted market abuse.  Section 118(5) provided:

“(5) The fourth is where the behaviour consists of effecting transactions or orders to trade (otherwise than for legitimate reasons and in conformity with accepted market practices on the relevant market) which -

(a)         give, or are likely to give, a false or misleading impression as to the supply of, or demand for, or as to the price of, one or more qualifying investments …”

40.    The Market Abuse Regulation came into effect on 3 July 2016 and directly applied to the UK for the remainder of the Relevant Period. 

41.    Article 15 of the Market Abuse Regulation provided: “A person shall not engage in or attempt to engage in market manipulation.”  Market manipulation was defined in Article 12 and Article 12(1) provided:

“For the purposes of this Regulation, market manipulation shall comprise the following activities:

(a) Entering into a transaction, placing an order to trade or any other behaviour which:

(i) gives, or is likely to give, false or misleading signals as to the supply of, demand for, or price of, a financial instrument …; 

…

unless the person entering into a transaction, placing an order to trade or engaging in any other behaviour establishes that such transaction, order or behaviour have been carried out for legitimate reasons, and conform with an accepted market practice as established in accordance with Article 13”.

42.    Article 12(2)(c) provided that the following behaviour shall, inter alia, be considered as market manipulation:

“the placing of orders to a trading venue, including any cancellation or modification thereof, by any available means of trading, including by electronic means, such as algorithmic and high-frequency trading strategies, and which has one of the effects referred to in paragraph 1(a) or (b), by:

…

(ii) making it more difficult for other persons to identify genuine orders on the trading system of the trading venue or being likely to do so, including by entering orders which result in the overloading or destabilisation of the order book; …”

43.    Futures were financial instruments and the Eurex Exchange was a regulated market for this purpose (Article 2 of the Market Abuse Regulation).

44.    Annex I of the Market Abuse Regulation sets out, without prejudice to the forms of behaviour set out in Article 12(1), further indicators of manipulative behaviour relating to false or misleading signals and to price securing which were to be taken into account by the Authority when transactions or orders to trade are examined.  These included, at A(f) of Annex I:

“the extent to which orders to trade given change the representation of the best bid or offer prices in a financial instrument…or more generally the representation of the order book available to market participants, and are removed before they are executed”

45.    Article 4 and Section 1 of Annex II of the Commission Delegated Regulation (EU) 2016/522 set out the practices which specify indicators A(a) to A(g) of Annex I of the Market Abuse Regulation.  Paragraph 6 of Section 1 of Annex II provided that such practices included:

“(a) Entering of orders which are withdrawn before execution, thus having the effect, or which are likely to have the effect, of giving a misleading impression that there is demand for or supply of a financial instrument […] at that price - usually known as ‘placing orders with no intention of executing them’…

(g) The practice set out in Point 5(e) of this Section, usually known as ‘layering’ and ‘spoofing’;”

46.    Point 5(e) was set out as:

“(e) Submitting multiple or large orders to trade often away from the touch on one side of the order book in order to execute a trade on the other side of the order book. Once the trade has taken place, the orders with no intention to be executed shall be removed - usually known as layering and spoofing…”

47.    The Authority’s case here is that the Traders were spoofing.

48.    The essential ingredient of market abuse under both s118 FSMA 2000 and the Market Abuse Regulation (for present purposes) is the giving of (or likelihood of giving) a false or misleading impression or signal as to supply, demand or price. 

49.    The question of what impression or signal has been given (or is likely to have been given) is an objective question.  The Authority is not required to establish that a trader intended to manipulate the market, or to give any particular signal to others in the market (Financial Conduct Authority v Da Vinci Invest Ltd [2015] EWHC 2401 (Ch) (“Da Vinci”) at [107]  and [156] and Winterflood Securities Ltd v Financial Services Authority [2010] EWCA Civ 423 at [17] and [25]).

50.    The question of whether the signal that is given (or is likely to be given) is false or misleading may, however, entail a subjective analysis, as explained by Andrew Baker J in Burford Capital Ltd v London Stock Exchange Group plc [2020] EWHC 1183 (Comm) (“Burford”) in the context of the Market Abuse Regulation:

“50. Thus, the essential ingredient is the giving of (or likelihood of giving) false or misleading signals as to supply, demand or price. Contrary to a submission by Mr Dhillon QC, that cannot always be an entirely objective enquiry, as it depends upon the signal given out (or likely to be given out) by a particular activity or behaviour. Depending on what that signal is (an objective question, to be sure), falsity might involve some subjective enquiry. The point of substance, therefore, is whether, as Mr Dhillon contended, there is no subjective element in the supply, demand or price signalling involved in this type of case (or, therefore, in considering whether there was any false or misleading signal).

51. I do not accept that contention. A seller wishing to sell at £10 who offers to sell at £10 but, finding no takers at that price, withdraws his offer because he does not want to sell for less, and a seller who has no intention to sell at £10 but who offers to sell at £10 to initiate or exacerbate a price trend, then withdraws his offer, appear to the outside observer to have behaved identically. If their behaviour fell to be judged entirely by that appearance, they would have to be either both guilty or both innocent of a charge of market manipulation. But in reality, surely the former is innocent, the latter guilty, and that is because although the signals sent out were the same (eg their initially signalled intention to sell at £10), the truth or falsity of those signals turns on their actual intentions, which differed radically.

52.  It is no defence for the guilty seller to say that he did not appreciate he was sending out a false signal, or in some other way that he did not intend to manipulate. In that sense, the manipulation does not have to be deliberate; and that is the relevant proposition to derive from Financial Conduct Authority v Da Vinci Invest Ltd [2015] EWHC 2401 (Ch), [2016] 3 All ER 547, [2016] Bus LR 274, per Snowden J at [104]–[108], decided under s 118 of the Financial Services and Markets Act 2000 (‘FSMA 2000’) implementing the EU Market Abuse Directive (European Parliament and Council Directive 2003/6/EC), which was the EU law predecessor to MAR. That is a different point, however, and the conclusion in that case, that there had been market abuse, required (and was justified by) the finding as to the actual purpose behind the relevant order activity (ibid, at [163]): ‘The purpose of the Traders was not in truth to sell or buy as indicated by the orders they placed. Instead, the Traders intended to cause a movement in the market price of the share … and to induce other market participants to place similar larger orders, with which the Traders could then trade aggressively in the opposite direction’.

53.  Again, in my simple example, the time at which and market circumstances in which the initial offer to sell at £10 was placed might afford an argument that a seller looking to sell but unwilling to sell below £10 would have assessed that there was no sensible prospect an offer to sell at £10 might be matched and so would not have placed the offer. That might be an aspect of assessing the genuineness of the offer in fact placed, but it could not be a conclusion that foreclosed the inquiry. That MAR adopts that approach is confirmed by Pt A of Annex I, with its list of ‘non-exhaustive indicators’ to be ‘taken into account’ when examining transactions or orders to trade for the purpose of applying art 12(1)(a).”

51.    Having considered the expert evidence, Andrew Baker J then emphasised:

“80.  Then finally, to be clear, and applying what I said in paras [50] to [51] above, it is not the case that to find a pattern of repeated placement, cancellation and replacement is to find spoofing or layering, because the question remains one of intention to trade. For example, a trader wishing to sell off or reduce a long position, in response to the Muddy Waters tweets or simply in response to the price falling, might generate such a pattern as part of a best-execution strategy that aimed to maximise his average price traded while still selling the volume he wishes to sell. That execution strategy might or might not hit both of those targets; but if that is what he is doing, though he generates one of Prof Mitts’ patterns, he never gives out a false supply or pricing signal to the market since each of his sell orders is placed intending it to trade, each cancellation is a response to the lack of matching demand, and each replacement order is again placed intending it to trade.

81.  Before getting into the detail, therefore, I am clear that Prof Mitts is wrong to opine that ‘repeatedly placing and cancelling orders in a very short time is strong evidence of manipulative intent’, or that ‘a wave of abnormal order cancellations [ie sell-side cancellations in statistically atypical volumes] at or above the best offer … indicates intentional manipulation of Burford’s share price’. Mitts 1 set out the logic for those claims, but it immediately betrays their error. The logic set out is that there is no economic justification for a short seller to place a large volume of sell orders above the best offer. The opinion that logic might justify, a more limited opinion than that expressed by Prof Mitts, is that a short seller repeatedly placing and cancelling orders in a very short time, above the best offer, is strong evidence of manipulative intent.”

Dishonesty

52.    Dishonesty is not part of the applicable test for determining whether a person has committed market abuse.  It was the basis of the Authority’s pleaded case in relation to the imposition of the prohibition orders, with the Authority alleging that the Traders were dishonest and lacking integrity.  

53.    The judgment of the Supreme Court in Ivey v Genting Casinos Ltd [2017] UKSC 67 (“Ivey”) is the leading authority on the meaning of dishonesty, and Lord Hughes set out the approach as follows:

“74 …When dishonesty is in question the fact-finding tribunal must first ascertain (subjectively) the actual state of the individual’s knowledge or belief as to the facts. The reasonableness or otherwise of his belief is a matter of evidence (often in practice determinative) going to whether he held the belief, but it is not an additional requirement that his belief must be reasonable; the question is whether it is genuinely held. When once his actual state of mind as to knowledge or belief as to facts is established, the question whether his conduct was honest or dishonest is to be determined by the fact-finder by applying the (objective) standards of ordinary decent people. There is no requirement that the defendant must appreciate that what he has done is, by those standards, dishonest.”

Role of the Tribunal

54.    Section 133(4) FSMA 2000 provides that, on a reference, the Tribunal may consider any evidence relating to the subject matter of the reference whether or not it was available to the decision-maker at the material time.  This is not an appeal against the Authority’s decision on each of the references but a complete rehearing of the issues which gave rise to the decision and which the applicant wishes the Tribunal to consider.  Section 133(5) to (7) provides as follows:

“(5) In the case of a disciplinary reference or a reference under section 393(11), the Tribunal must determine what (if any) is the appropriate action for the decision-maker to take in relation to the matter, and on determining the reference, must remit the matter to the decision-maker with such directions (if any) as the Tribunal considers appropriate for giving effect to its determination.

(6) In any other case, the Tribunal must determine the reference or appeal by either -

(a) dismissing it; or

(b) remitting the matter to the decision-maker with a direction to reconsider and reach a decision in accordance with findings of the Tribunal.

(6A) The findings mentioned in subsection (6)(b) are limited to findings as to -

(a) issues of fact or law;

(b) the matters to be, or not to be, taken into account in making the decision; and

(c) the procedural or other steps to be taken in connection with the making of the decision.

(7) The decision-maker must act in accordance with the determination of, and any direction given by, the Tribunal.”

55.    Pursuant to s133(7A), “disciplinary reference” includes a decision to take action under s66 FSMA 2000, ie to impose a financial penalty on a person.  The term does not include a decision to impose a prohibition order under s56.  Thus, these references are disciplinary references in respect of the penalties imposed and non-disciplinary references in respect of the prohibition orders.

Non-disciplinary references

56.    In the case of a non-disciplinary reference, the Tribunal must determine the reference by either: “(a) dismissing it; or (b) remitting the matter to the decision-maker with a direction to reconsider and reach a decision in accordance with the findings of the Tribunal” (s133(6)).  It does not itself determine what the appropriate action is for the Authority to take.

57.    Otherwise than in cases where the Tribunal dismisses the reference, its powers are confined to giving directions to the Authority to remake its decision in light of the matters within the scope of s133(6A) which the Tribunal considers appropriate (Page v Financial Conduct Authority [2022] UKUT 124 (TCC) (“Page”) at [113]).

58.    Accordingly, as set out in in Carrimjee v Financial Conduct Authority [2016] UKUT 447 (TCC) at [39] to [40] and Page at [114] to [115], under s133(6), the Tribunal has various options when dealing with a “non-disciplinary” reference:

(1)        It will dismiss the reference if either:

(a)        having reviewed all the evidence and factors taken into account by the Authority in making its decision, and having made findings of fact in relation to that evidence and such other findings of law that are relevant, it concludes that the decision to prohibit is one which was reasonably open to the Authority; or 

(b)        it concludes that the seriousness of the Tribunal’s own findings would lead inevitably to the Tribunal reaching the same decision as the Authority in the event the decision was remitted back to the Authority, even where the Tribunal has not accepted all the factors relied upon by the Authority in reaching its original decision to make a prohibition order, such that it can be said to have taken into account irrelevant considerations. 

(2)        It will remit the decision to the Authority with directions if it is not satisfied that the decision made is one which, in all the circumstances, is within the range of reasonable decisions open to the Authority.

59.    The Tribunal may have regard to any evidence, whether or not it was available to the Authority at the time. 

Disciplinary references

60.    A reference of a decision to impose a financial penalty is a “disciplinary reference” and the Tribunal has the power to determine at its discretion what (if any) is the appropriate action for the Authority to take, including a determination as to whether or not to impose a financial penalty and, if so, the amount of such penalty.

61.    As the Tribunal indicated in Tariq Carrimjee v Financial Conduct Authority [2015] UKUT 79 (TCC), the Tribunal is not bound by the Authority's policy when making an assessment of a financial penalty on a reference, but it pays the policy due regard when carrying out its overriding objective of doing justice between the parties.  In so doing the Tribunal looks at all the circumstances of the case.

62.    Similarly, in Westwood Independent Financial Planners v Financial Conduct Authority [2013] 11 WLUK 630, the Tribunal held at [181]:

“In considering the appropriate level of a penalty we are not bound by the Authority’s tariff for particular misconduct, or even the factors the Authority takes into account, but may reduce or increase a penalty which is the subject of a reference on any grounds we think fit, within the parameters of the proper exercise of judicial discretion. In practice, the Tribunal respects the Authority’s tariff, in the interests of consistency between applicants, while departing from it in an appropriate case.”

63.    This approach was adopted by the High Court in Da Vinci where Snowden J said in the context of the imposition of a penalty for market abuse at [201]:

“It was the FCA's submission, and I accept, that in determining any penalty under section 129, the starting point for the court should be to consider the relevant DEPP penalty framework that was in existence at the time of commission of the market abuse in question. To do otherwise would risk introducing an inequality of treatment of defendants depending upon whether the proceedings were taken against them under the regulatory route or the court route and depending upon how long the proceedings had taken to come to a conclusion. By the same token, however, in common with the Upper Tribunal, the court is not bound by that framework, or by the FCA's view of how it should be applied. But if the court intends to depart from the framework in a particular case, it should explain why it considers it appropriate to do so. It occurred to me that in this regard there is some analogy with the approach of the criminal courts to the application of the sentencing guidelines produced by the Sentencing Council.”

Burden and Standard of proof

64.    The burden of proof is on the Authority to establish the matters on which they rely, namely that the Traders deliberately engaged in market abuse and their conduct was dishonest and lacked integrity. 

65.    The standard of proof is the civil standard of the balance of probabilities.

Evidence including witnesses who had not been called, information that is no longer available and relevance of delay

66.    The Traders made submissions in relation to the witnesses who had not been called by the Authority to give evidence before this Tribunal, the information that would have been available to the Traders during the Relevant Period but which is no longer available or has not been disclosed to them and the time which has passed since the Relevant Period (more than 8.5 years by the time of the hearing).  They submitted that the Tribunal should draw adverse inferences from the Authority’s failure to call witnesses, and that the delay and absence of information that would have been available to the Traders at the time was unfair and prejudicial to the Traders, particularly in circumstances where the Authority’s case was based on detailed trading activity within the Instance Pool and they were cross-examined on the reasons for placing particular orders within that Instance Pool.  

67.    We first set out a summary of the evidence which was produced, before then addressing these submissions and setting out the Tribunal’s approach to the evidence.

68.    When considering the parties’ submissions in relation to the evidence the Tribunal keeps in mind throughout that in regulatory proceedings the Authority is not an ordinary litigant but is instead engaged in a “common enterprise” with the court in ensuring that the objects of the legislation are achieved and that public confidence is maintained in the integrity of the financial markets (Financial Conduct Authority v Seiler [2024] EWCA Civ 852 at [51]). 

Outline of evidence before the Tribunal

69.    We heard evidence from the Traders, each of whom had provided witness statements and were cross-examined by Mr Shivji on behalf of the Authority and, in the case of Mr Urra, also by Mr Bailin on behalf of Mr Sheth.  The latter cross-examination was confined to whether Mr Urra recalled either showing the Information Discovery Strategy to Mr Sheth and/or a subsequent discussion with Mr Sheth in relation to which Mr Sheth’s evidence was that he expressed his frustration that the strategy was not working. 

70.    In cross-examination Mr Shivji put the Authority’s case to each of the Traders, both in general terms by reference to the Authority’s allegations that their Trading Strategies were fabricated, and specifically by reference to some of the Instances and other trading activity to which the Traders had referred in their witness statements.  The Tribunal does not at this stage set out its conclusions on the credibility of each of the Traders.  They each denied market abuse, giving evidence that they had an intention to trade all of their Large Orders and did not collaborate to commit an abusive scheme.  The Tribunal’s assessment of that evidence is made in the light of all of the evidence.   

71.    The Tribunal also heard evidence from Mr Fernando Fernandez-Maquieira, a broker in the EGB market, who had provided a witness statement and was cross-examined by Ms Hassell-Hart on behalf of the Authority.  Mr Fernandez-Maquieira had worked with traders for 20 years and Mr Urra had been a client of his.  He had first met Mr Urra whilst Mr Urra was at Credit Suisse, and Mr Urra remained a client during his time at MHI between 2012 and 2019.  Mr Fernandez-Maquieira’s role was to help the traders find buyers and sellers for their orders.  He gave evidence in relation to order size (both in the EGB futures market generally and sizes transacted by MHI), the splitting of orders and his opinion on the plausibility of Mr Urra’s Trading Strategy.  The Tribunal accepts that Mr Fernandez-Maquieira gave his evidence honestly and sought to assist the Tribunal.  However, the Tribunal has not relied on his evidence when making findings of fact as his evidence was given by reference to the long period of time during which he worked with Mr Urra and he was not able to be specific about order sizes by reference to particular periods of time as he no longer has access to the data relating to his transactions with MHI during the Relevant Period. 

72.    Mr Simon Virciglio, who is a former broker in the EGB market, had provided a witness statement in the context of Mr Urra’s reference and was willing to attend and be cross-examined but the Authority, Mr Lopez and Mr Sheth agreed he did not need to attend for cross-examination.  Mr Virciglio’s evidence was essentially a character reference for Mr Urra, referring to him being held in high regard, straightforward, fair, meticulous and focused on his trading strategies.  He also described both Mr Lopez and Mr Sheth as straightforward and honest.  It was common ground between the parties that none of the Traders had previously been subject to any compliance investigation or disciplinary proceedings and, with no disrespect intended to Mr Virciglio, the Tribunal places no weight on his evidence.

73.    The Tribunal also had the benefit of expert evidence from:

(1)        Mr Stephen Creaturo, who is a Financial Markets Advisor at Valere Capital Partners LLP, and had been instructed as an expert by the Authority; and

(2)        Mr Andrew Kasapis, who is a Senior Director at Kroll Advisory Ltd, and had been jointly instructed by the Traders as an expert.

74.    Mr Creaturo and Mr Kasapis both produced expert reports for the Tribunal and the Tribunal was also taken to the expert reports which Mr Kasapis had prepared before the RDC.  Mr Creaturo and Mr Kasapis both addressed the matters identified in the agreed list of issues, namely:

(1)        What impact, if any, would the Instances identified in the Instance Pool have on the market (considered by reference to the impression and/or signal as to the supply or demand for, or price, of BTP Futures that the trading instances would give or be likely to give to the market, if any)?

(2)        Is the trading activity (both in the individual trader and the multi-trader Instances) identified in the Instance Pool explained by/consistent with an intention to pursue the alleged trading strategies identified by each Trader?

(3)        Whether the large orders outside the Instance Pool are explained by/consistent with an intention to pursue the alleged Trading Strategies.

(4)        Are the Specified Instances representative of the Instance Pool as a whole?

75.    The evidence in the hearing bundles included:

(1)        trading activity in relation to the 233 Instances in the Instance Pool as illustrated in Appendix 2 (albeit that the narrative was only provided for the Specified Instances);

(2)        daily price graph (London time) for Futures on each trading day in the Relevant Period (which had been exhibited by Mr Lopez) - the price movements for each day were shown on a single page, with 15 minute intervals marked;

(3)        list of trades in BTPs entered into by the Desk between 1 June and 31 July 2016;

(4)        Electronic RFQs received by the Desk in BTPs between 1 June and 31 July 2016;

(5)        a list of the 93 RFQs that came into the Desk during the Relevant Period which would have required hedging with a Futures order of 200 lots or more (which had been exhibited by Mr Lopez); 

(6)        floor plans for the relevant part of MHI’s offices during the Relevant Period;

(7)        records of training undertaken by the Traders, with copies of slides from some of that training;

(8)        the MHI Compliance Report (as defined below);

(9)        transcripts of the Authority’s interviews with other employees of MHI namely Christian Heiberg (Head of Fixed Income Trading), Dinesh Joshi (Head of Compliance), Anthony Algeo (Executive Director, Compliance), Graham Halliday (COO Front Office), Mehdi Barouti and James Hill (the other traders on the Desk); and

(10)     materials produced by the Authority, including summaries of data obtained by the Authority from the Exchange in relation to orders placed by other market participants (the “SMARTS data”).

Pace of Authority’s investigation and particularisation of its case

76.    The Traders drew attention to the length of time since the Relevant Period and submitted that the delays by the Authority in its investigation and in particularising its case have resulted in unfairness and prejudice. 

77.    On 26 July 2016, the Exchange sent a letter to Mizuho Securities USA Inc (“MSUSA”) (which was the member firm on the Exchange) querying the trading rationale for two identified episodes of trading which had taken place on 29 June 2016 (the “Eurex Letter”).  MSUSA forwarded this letter to MHI on 28 July 2016.  MHI then interviewed Mr Urra (on 29 July and 5 August 2016) and Mr Sheth (on 3 and 9 August 2016) before responding to Eurex on 22 August 2016.

78.    MHI’s Compliance team continued its own review of the trading activity (as part of which it interviewed Mr Lopez on 29 September 2016) and produced a report on “A Review of European Government Bond Desk Dealing Practice on the EUREX Exchange” in October 2016 (the “MHI Compliance Report”).  That report was sent to the Traders on 13 October 2016 and they were informed by MHI that they would be the subject of a disciplinary investigation in relation to orders for Futures placed on Eurex.  MHI sent this report to the Authority (which had already been informed of the Eurex Letter).

79.    The matter was referred to MHI’s Head of Human Resources (“HR”) for a disciplinary investigation.   MHI held further meetings with each Trader - Mr Urra and Mr Lopez were each interviewed on 18 October 2016, Mr Sheth was interviewed on 14 October 2016, and there were disciplinary meetings with each Trader on 2 November 2016.

80.    On 10 November 2016, MHI’s CFO wrote to each of the Traders to inform them of the outcome of the disciplinary process.  MHI took no disciplinary action against Mr Lopez.  MHI issued a first warning to Mr Sheth in respect of his “poor practice” in placing multiple orders instead of amending existing orders and a final warning to Mr Urra in respect of his management of Mr Sheth.  There were no findings of market abuse against any of the Traders.

81.    Mr George submitted that the Authority’s investigation then progressed at an exceptionally slow pace: 

(1)        The Authority had been sent the MHI Compliance Report in October 2016 but it did not appoint investigators until 16 March 2017.

(2)        The Authority then requested the Eurex trading data on 25 April 2017, with the assistance of the Federal Financial Supervisory Authority of Germany (“BaFin”), which was received on 19 September 2017.

(3)        The Authority didn’t interview the Traders until the first half of 2018 - Mr Lopez on 18 and 19 January 2018, Mr Sheth on 15 March 2018 and Mr Urra on 16 April 2018.

(4)        It was not until April 2019 that the Authority indicated that it intended to engage an independent expert.  They instructed their first expert, Mr Santacana, in June 2019, whose expert report was then produced on 30 September 2020.

(5)        The Authority served Annotated Warning Notices on the Traders in October 2020. 

(6)        The Traders each provided a response in December 2020, but the Enforcement and Market Oversight Division (“EMO”) didn’t serve its Enforcement Submissions Document on each of the Traders until 19 May 2021. 

(7)        The Warning Notices were issued to each Trader on 16 July 2021.

(8)        The RDC hearings took place in March 2022. 

(9)        The Authority issued its Decision Notices for each of the Traders on 31 October 2022. 

82.    The Traders referred these matters to the Tribunal on 28 November 2022, and the references were heard from 27 January to 14 February 2025.  Mr George referred to the decision of the Tribunal in Seiler, Whitestone and Raitzin v Financial Conduct Authority [2023] UKUT 133 (TCC) (“Seiler UT”) where the Tribunal had (at [59]) described the delays as unsatisfactory and recorded that in that case the lapse of time between the events in question and the hearing of the references was longer than any other comparable proceedings in the experience of that Tribunal, noting that such record has now been surpassed in these proceedings.

83.    Alongside their submissions in relation to what they described as the slow progress of the Authority towards the issue of Decision Notices, the Traders drew the Tribunal’s attention to how the Authority’s case has been particularised and submitted that the Authority’s failure to particularise the Instance Pool at an early stage has also hindered the Traders’ ability to recollect the specific trading activity on which the Authority now relies:

(1)        The Eurex Letter asked about two Instances, which are now referred to as F174 and F176.

(2)        The Traders were interviewed by the Authority in 2018, but at that time the Authority put a limited number of Instances to them - they had been provided with bundles containing trading data for approximately 11 Instances each, which were the subject of the interviews.

(3)        The Traders submitted that the first time the Traders were told in any detail of the allegations against them was when the Authority issued the Annotated Warning Notices in October 2020.

(4)        The Authority’s first statements of case were served in February 2023, which annexed a table listing the 233 Instances, of which only a small number were particularised.

(5)        On 2 February 2024 the Authority revised the number of Instances relied upon against each Trader and provided “further particulars” for each alleged Instance of market abuse.

(6)        Following a contested application which was determined at a case management hearing on 20 June 2024, the Tribunal ordered the Authority to: (1) identify ten individual instances per Trader and ten Multi Trader Instances (these being the Specified Instances); (2) provide narratives detailing each Specified Instance; and (3) limit cross-examination of the Traders to these Specified Instances and any further Instances or trading activity outside of the Instance Pool identified by the Traders.

(7)        The Authority identified the Specified Instances on 4 July 2024 and filed narratives on 18 July 2024.

84.    Mr Shivji submitted that this is a particularly technical and complex matter and the Authority had sought to conduct a thorough and fair investigation.  It was inevitable that it was always going to take some time for the Authority to analyse the data - the Authority received approximately 16.5m lines of data from Eurex alone.  The Authority submitted that the relevance of the delay must be set in its proper context, saying the Traders’ central complaint was understood to be that the passage of time meant that they have difficulty recalling what happened and why; this was described as “uncontroversial” by the Authority.  On the facts, Mr Shivji submitted that even if the Authority had been able to move much more quickly - which was not accepted - this would not have improved the evidence before the Tribunal.  The Traders acknowledged that their recollection of the specific trades was poor even in September and October 2016; and on any view the hearing before the Tribunal would be taking place years after the end of the Relevant Period.

85.    The hearing took place over a period of three weeks.  Having heard the submissions of the parties, the evidence of the Traders and the two experts, during the course of which the Tribunal was taken to a large amount of data in relation to the trading activity, the Electronic RFQs received by the desk and the cash bond transactions which were executed - which was itself only a small portion of that which was in the hearing bundle - the Tribunal accepts that a thorough investigation would take time.  However, whilst we have not heard from any witnesses from the Authority who could have explained the progress of the various steps in the investigation (an approach which was itself criticised by the Traders in their opening submissions), the Tribunal considers that there were some delays that are unexplained and are not a consequence of the complexity of the underlying activity.  Those which potentially had the most significance are those which occurred in the early stages, in particular:

(1)        the delay between the Authority receiving the MHI Compliance Report in October 2016 (having already been informed of the receipt of the Eurex Letter at the end of July or beginning of August 2016) and appointing investigators in March 2017, which in turn delayed the request for Eurex trading data and contributed to the Traders only being interviewed by the Authority in 2018; and

(2)        the delay in deciding to engage an independent expert and the period of time which followed until production of the first expert report in September 2020, ie more than four years after the end of the Relevant Period.

86.    Progress thereafter was not swift.  But it is these delays which meant that it was more than four years after the Relevant Period that the Authority served the Annotated Warning Notices on the Traders, and whilst the Authority has subsequently amended the number of Instances on which it relies against each Trader, it was not until this time that the Traders would have understood the breadth of the case being made against them and, importantly, that the Authority was not only challenging whether the Traders were pursuing a legitimate trading strategy but also alleging that they had each (individually and together) engaged in market abuse on a large number of occasions throughout a period of two months. 

Lack of information that would have been available to the Traders during the Relevant Period

87.    The Traders have each emphasised in their witness statements and during cross-examination that they no longer have access to the full range of information that would have been available to them during the Relevant Period and which would have formed part of the rationale for their trading activity and that this, as well as the delays by the Authority in disclosing the information which is now available, has hindered their ability to recall matters.

88.    The Traders have referred to having requested information from the Authority at various times during the Authority’s investigation, eg Mr Urra asked for specified categories of information on 4 April 2018.  Similar requests have been made by Mr Lopez and Mr Sheth.  Some, but not all, information that has been requested has been provided by the Authority.  The Traders have not made an application to the Tribunal for disclosure by the Authority.    

89.    The Authority’s disclosure obligations are set out in the Upper Tribunal Rules.  Paragraph 4(3) of Schedule 3 requires that the Authority provide a list of any documents on which the Authority relies in support of the referred action and any further material which in the opinion of the Authority might undermine the decision to take that action; and the secondary disclosure obligation in paragraph 6(1) then requires that the Authority disclose any further material which might reasonably be expected to assist the applicant’s case as disclosed by the applicant’s reply. 

90.    In late 2020, the Authority provided to the Traders:

(1)        Electronic RFQs received by the Desk for BTPs in the Relevant Period;

(2)        a record of all orders for Futures placed by the Traders during the Relevant Period - this includes orders which filled or partially filled as well as those which were cancelled, and identifies the Trader who placed the order, timing, direction (ie buy/sell), size and amendments made to the order; and

(3)        records of executed cash trades in BTPs entered into by the Desk in the Relevant Period.

91.    On 1 August 2023, the Authority provided to the Traders a copy of the full Eurex dataset it had received from BaFin.  This dataset listed all orders, including enters, fills, partial fills, amendments and cancellations, in Futures by all market participants during the Relevant Period. 

92.    In their written skeleton arguments and in their opening submissions the Traders referred to information that is not available.  This includes the following categories: 

(1)        records of RFQs received over the phone (“Voice RFQs”);

(2)        records of RFQs sent electronically through Bloomberg message/chat functionality (“Bloomberg/Chat RFQs”);

(3)        records of the information broadcast to the Traders on the squawk box;

(4)        records of the cash bond market, including interdealer market order books from Euro MTS (multiproduct), MTS Spain (Spanish bonds), Brokertec (multiproduct) and SENAF (Spanish bonds);

(5)        content of the Traders’ books of inventory at MHI; and

(6)        records of communications, including telephone calls, emails and messages, including with the sales team.

93.    In their written closing submissions the Traders adopted different positions from each other in relation to the disclosure which had been made by the Authority:

(1)        Mr George set out a chronology of Mr Urra’s requests for information, summarising responses and what had been received, and, when outlining the “considerable amount” of information that is not before the Tribunal, said it is “not clear” whether any of this information is available to the Authority and has not been shared.

(2)        Mr Jaffey submitted that the Authority has never produced a range of relevant secondary disclosure that is “understood to be in its possession” including evidence of voice orders placed with the Desk, the squawk box data and the relevant part of MTS (most relevantly the Eurobond book), which would have provided information on a number of issues, including the cash position at any given period of time. 

(3)        Mr Bailin submitted that there is now a question as to whether the Authority has properly discharged its obligations of disclosure, referring to the Enforcement Submissions Document which had been submitted to the RDC which confirmed that various communications were provided by MHI and reviewed by the EMO, namely telephone calls on fixed office lines, audio recordings from a dealer board system, emails and Bloomberg messages.  EMO’s review of these had been aimed at finding any communications that explained the rationale behind the placing of the Large Orders.  None of this information has been disclosed.  Mr Bailin submitted that it seems clear that the communications review has not been undertaken in respect of all of the Instance Pool, or all of the Specified Instances, and did not look at the rationale for the Small Orders.  The current position, he submitted, was that the Authority “simply do not know with regard to the communications review whether there is material which meets the statutory test for disclosure”. 

94.    In his oral closing submissions (which were delivered the afternoon after the exchange of lengthy written closing submissions), Mr Shivji submitted it had not been said by the Traders that the Authority has material and has failed to comply with its disclosure obligations.  Mr Shivji acknowledged that it was not possible to recreate the exact trading environment, referring to the difficulties experienced by MHI in seeking to retrieve information.

95.    The Tribunal considers that these differences between the Traders, and Mr Shivji’s submission at the hearing that it was not said by the Traders that the Authority has failed to comply with its disclosure obligations, whereas both Mr Bailin and Mr Jaffey did in fact come close (at the very least) to taking such a position in their written closing submissions, illustrate the difficulty which is faced by this Tribunal where an applicant has not made an application for disclosure before the start of the hearing.  Such an application would require the applicant to set out the basis on which it considers that the Authority had material which was required to be disclosed pursuant to the Upper Tribunal Rules, and the Authority would then be expected to make representations and potentially adduce witness evidence from those who had conducted the investigation and could speak to the disclosure exercise which had been undertaken, which would include identifying material which has not been obtained by the Authority, and, separately, explaining why the Authority considers that it is not required to disclose material which is in its possession.    

96.    In the absence of an application for disclosure having been made which would have enabled the matters to which the Traders now refer (so far as they comprise allegations or suggestions of non-compliance rather than drawing attention to material to which none of the parties have access) being assessed with the benefit of submissions and witness evidence, the Tribunal is not persuaded that it would be procedurally fair to the Authority for the Tribunal to reach a conclusion that the Authority has failed to comply with its disclosure obligations in these proceedings. 

97.    The Tribunal does, however, address the range of information that is not in fact available to the Traders and its potential relevance. 

98.    The Tribunal finds that the categories of information at [92] above would have been available to the Traders at the time of the Instances.  The Tribunal also finds that this information was potentially relevant to their trading activity and to the decisions which they made at that time.  This is apparent from the evidence of the Traders, where they each made repeated references to this information and in relation to which they were cross-examined:

(1)        Mr Urra described his trading as being multidimensional.  Mr Urra said he only had access to a fraction of the relevant information he would have had access to at the time - specifically identifying his general book positioning strategy, his views on the market (direction and curve views), what was taking place in the voice broker markets, who were the clients making enquiries or orders to work on and the MTS and Brokertec bond trade data. 

(2)        Mr Lopez frequently referred in his evidence to market movements in the cash bond market on MTS, explaining that this was vital to his assessment of the market position at the time and would have informed his placing and cancellation of orders.  During the day he would use the Bund futures market as a reference point as there is more liquidity and typically less price movement in the Bund futures market than in Futures.

(3)        Mr Sheth explained that cash bonds were his primary exit strategy, and referred to information from brokers and the sales team as indicators of client activity.

99.    The Authority has produced an explanation from MHI for the absence of Voice RFQs and Bloomberg/Chat RFQs, set out in the Enforcement Submissions Document for each Trader its review of communications data (which the Tribunal infers forms the basis of the Authority’s conclusion that disclosure was not required) and, in submissions, referred to the fact that the Authority does not have the Traders’ books of inventory as they existed in real-time throughout the Relevant Period. 

100. The Electronic RFQs from the Relevant Period had been retained in searchable form by MHI.  They were provided by MHI to the Authority, have been disclosed and were in the hearing bundle.  However, in response to an information request from the Authority, MHI had responded:

“Where a client submits RFQs to Sales staff by voice (on the phone) or through Bloomberg message/chat functionality, MHI responds to these RFQs by voice or Bloomberg message/ chat. Whilst these client interactions are recorded and records maintained (e.g. phone records, Bloomberg message/ chat records), they are not maintained centrally in an easily searchable format, and as such we have not included these in the attachment.

On occasion, the EGB desk may be asked to ‘work an order’ (for example, where the price quoted by MHI is rejected by the client, the client may ask the desk to trade if the price reaches a certain level). This will usually be done over the phone or through Bloomberg message/ chat. The desk has previously informed MHI Compliance Department that orders represent 1-3% of all trading activity.

In order to prepare a list of RFQs and orders received from clients by voice or through Bloomberg message/ chat during the period requested, a detailed review of electronic communications and telephone recordings between Sales staff and MHI clients would need to be undertaken.”

101. On the basis of this response, the Tribunal finds that MHI did not provide records of Voice RFQs or Bloomberg/Chat RFQs to the Authority.  The Authority was not therefore able to review such data itself or to disclose this data to the Traders; and such records are not otherwise available to the Traders.  (The communications that MHI provided to the Authority are identified at [104] below.) 

102. It is not known how many Voice RFQs or Bloomberg/Chat RFQs may have been sent to the Desk.  Mr George referred to Mr Sheth’s evidence that the records of executed trades in cash BTPs record “n/a” in the “venue” column if they were in response to a Voice RFQ or Bloomberg/Chat RFQ.  There were 446 such trades in the Relevant Period, out of 1,675 recorded entries.  The total number of Electronic RFQs received by the Desk in Futures during the Relevant Period was 6,972.  Mr George submitted that, assuming the execution rates for Electronic, Voice and Bloomberg/Chat RFQs were broadly similar, there may have been approximately 2,500 Voice or Bloomberg/Chat RFQs during the Relevant Period.  The Tribunal accepts this is a reasonable approach to estimating the number of RFQs which were sent to the Desk and in relation to which the parties (and the Tribunal) have no further information. 

103. Save for the communications referred to at [104] below, the Authority has not provided any communications data from the Relevant Period to the Traders, whether emails, phone records, or WhatsApp/text messages.  As identified by Mr Bailin, it is clear from the Enforcement Submissions Documents that the Authority did have some communications data from MHI, and MHI’s ability to capture this data is also evident from the list of material which was reviewed by Compliance in its preparation of the MHI Compliance Report. 

104. The Enforcement Submissions Document for each Trader listed the communications provided by MHI which were reviewed by EMO as being telephone calls made on fixed office lines, audio recordings from a dealer board system, emails and Bloomberg messages.  MHI had identified a technical fault with the telephone call recording system which resulted in a number of calls not being recorded, and this meant the majority of Mr Sheth’s calls were not recorded as well as calls of some members of the sales team.  The review conducted by EMO was “aimed at finding any communications that explained the rationale behind the Traders’ large order placement”, and they reviewed all available communications from 30 minutes before the placement of the first Large Order in each Instance to the cancellation of the last Large Order in each Instance.  This review was only conducted for some of the Instances.  The EMO stated that no communications were found in which the Traders discussed their rationale for the placement of the Large Orders.  They reviewed the communications for anything considered relevant to the placement of the Large Orders, as a result of which 31 communications were provided to the expert for his consideration, and he confirmed that his view of the Traders’ behaviour remained unchanged.

105. As regards this review by EMO, Mr Bailin drew attention to the aim being to find any communications that explained the rationale behind the placing of the Large Orders.  He submitted that the Authority had not reviewed communications for any explanation of the placing of the Small Orders.  This submission is relevant to the Information Discovery Strategy said to have been pursued by Mr Urra and Mr Sheth (where the Small Orders and the Large Orders were prompted by the same market activity).  However, the Tribunal does not accept that the description of the scope of its review in these terms necessarily means that the EMO, in conducting the review of communications, would have rejected as irrelevant any communications in relation to the rationale for the Small Orders.  Whilst Mr Urra and Mr Sheth have both provided further and more detailed explanations of the Information Discovery Strategy since they served their Replies, their explanation that there was a link between the client orders they were receiving and the placing of the Large Orders is reasonably clear from those Replies, and the Authority, we infer, concluded that there was no further material which might reasonably be expected to assist the Traders’ case as disclosed by the Replies.  The Authority’s conclusion was that no secondary disclosure was required in respect of its communications review.  This was not said to be incorrect by the Traders at that time.

106. The Authority submitted that the Traders, having seen the Eurex Letter and received the MHI Compliance Report, could have checked their own trading and secured the evidence on which they wish to rely whilst they were at MHI (seeking such assistance as they required from Compliance) - Mr Urra had continued to be employed by MHI until September 2019, Mr Lopez until February 2020 and Mr Sheth until October 2020.  During cross-examination, Mr Shivji criticised the Traders for not having obtained this evidence themselves. 

107. The Tribunal would expect Mr Lopez and Mr Sheth (whose trading activity in Futures was involved in F174 and F176) to have obtained all possible information that was potentially relevant to their trading in those Instances as soon as they became aware of the Eurex Letter.  The specific activity was referenced by the Exchange, they were asked about it by Compliance, they knew they needed to provide an explanation and the events were relatively recent such that records should have been available.

108. However, we are not persuaded that it would have been feasible for the Traders to secure evidence in relation to all of their trading activity during the Relevant Period at that time or whilst they were at MHI, nor that they could have anticipated that this would be required:

(1)        Whilst the Traders knew their trading was under scrutiny, and the interviews with the Authority in 2018 would have reinforced the potential severity of the situation, we consider that the breadth of the Authority’s case would only have become apparent to the Traders once they received the Annotated Warning Notices in October 2020.  Those notices were critical in setting out how the Authority was making its case and making clear that the Authority was relying on such a large pattern of trading activity.

(2)        To the extent that the Traders sought information from MHI as to Voice RFQs and Bloomberg/Chat RFQs which were received by the Desk, we anticipate that MHI would have given the same response to the Traders as it did to the Authority.

(3)        The Authority has not made submissions or provided evidence as to how it might have been possible for the Traders to capture or recreate data in relation to the cash BTP market from the time, or as to how they could have recreated their books of inventory showing live positions.

109. The Tribunal has regard to what it considers to be the practical reality, namely that the Traders only have some of the information available to them to assist with explaining their trading activity and it is not possible now to recreate the trading environment within which the Traders were working during Relevant Period.  Furthermore, the Tribunal is mindful throughout that the burden of proof is on the Authority.

Passage of time, memory and witness evidence

110. The Traders submitted that the delays by the Authority in conducting its investigation, and its failure to particularise fully its case at an early stage or put more of the allegations to the Traders in interview and to provide information that would have been available at the time of their trading has prejudiced the Traders and their ability to recall the rationale and motivations for particular trading activity which occurred several years ago. 

111. The Authority submitted that the Traders have known since the Eurex Letter that these were allegations that concerned potential market abuse.  The Authority’s case has, it submits, been consistent since the beginning of the investigation - the allegations relate to the placing of large orders opposite small orders which were then cancelled after the small order traded.  Mr Shivji submitted that the impact of the passage of time as a result of any delays (which were denied) in the investigation was likely to have been modest. 

112. The Traders took different approaches in their witness statements to explaining their trading activity.  They were each clear that they do recall and can explain their Trading Strategy.  They generally said they could not recall the rationale for specific trading activity in the Instances (although Mr Lopez did, whilst giving evidence, occasionally say that he had some actual recollection, eg in relation to F30).  Mr Urra addressed two Instances (F47 and F150) as they were two of the trades put to him by the Authority in interview in 2018 and, as a result, he said he had a better recollection of these trades as it had been necessary for him to consider them throughout the investigation, whereas Mr Lopez and Mr Sheth sought to reconstruct the Specified Instances as well as some activity outside the Instance Pool based on the information that is available to them, whilst also drawing attention to the information to which they do not have access. 

113. The submissions of the Traders understandably linked together what they submit are the delays in the investigation, lack of information and fallibility of memory.

114. The parties took the Tribunal to the leading authorities on the approach to be adopted when assessing the reliability of witness evidence.  The Tribunal found the following guidance on the importance of contemporaneous documentation, the task of making findings of fact based upon all of the evidence and the position where there is a “documentary lacuna” to be particularly helpful in the context of the present references:

(1)          In Simetra Global Assets Ltd & Anor v Ikon Finance Ltd & Ors [2019] EWCA Civ 1413, Males LJ stated the following at [48]:

"48. In this regard I would say something about the importance of contemporary documents as a means of getting at the truth, not only of what was going on, but also as to the motivation and state of mind of those concerned. That applies to documents passing between the parties, but with even greater force to a party's internal documents including emails and instant messaging. Those tend to be the documents where a witness's guard is down and their true thoughts are plain to see. Indeed, it has become a commonplace of judgments in commercial cases where there is often extensive disclosure to emphasise the importance of the contemporary documents.  Although this cannot be regarded as a rule of law, those documents are generally regarded as far more reliable than the oral evidence of witnesses, still less their demeanour while giving evidence. The classic statement of Robert Goff LJ in The Ocean Frost [1985] 1 Lloyd's Rep 1 at p.57 is frequently, indeed routinely, cited:

"Speaking from my own experience, I have found it essential in cases of fraud, when considering the credibility of witnesses, always to test their veracity by reference to the objective facts proved independently of their testimony, in particular by reference to the documents in the case, and also to pay particular regard to their motives and to the overall probabilities. It is frequently very difficult to tell whether a witness is telling the truth or not; and where there is a conflict of evidence such as there was in the present case, reference to the objective facts and documents, to the witnesses' motives, and to the overall probabilities, can be of very great assistance to a judge in ascertaining the truth. I have been driven to the conclusion that the Judge did not pay sufficient regard to these matters in making his findings of fact in the present case."

(2)        The Court of Appeal in Kogan v Martin [2020] EMLR 4 said this at [88]:

"88. …First, as has very recently been noted by HHJ Gore QC in CBX v North West Anglia NHS Trust [2019] 7 WLUK 57, Gestmin is not to be taken as laying down any general principle for the assessment of evidence. It is one of a line of distinguished judicial observations that emphasise the fallibility of human memory and the need to assess witness evidence in its proper place alongside contemporaneous documentary evidence and evidence upon which undoubted or probable reliance can be placed.  Earlier statements of this kind are discussed by Lord Bingham in his well-known essay The Judge as Juror: The Judicial Determination of Factual Issues (from The Business of Judging, Oxford 2000). But a proper awareness of the fallibility of memory does not relieve judges of the task of making findings of fact based upon all of the evidence. Heuristics or mental short cuts are no substitute for this essential judicial function. In particular, where a party's sworn evidence is disbelieved, the court must say why that is; it cannot simply ignore the evidence."

(3)        The Court of Appeal addressed the position where there are documentary lacunae in NatWest Markets PLC and others v Bilta (UK) Limited (In Liquidation) and others [2021] EWCA Civ 680 at [50] and [51]:

“50. ….it is important to bear in mind that there may be situations in which the approach advocated in Gestmin will not be open to a judge, or, even if it is, will be of limited assistance. There may simply be no, or no relevant, contemporaneous documents, and, even if there are, the documents themselves may be ambivalent or otherwise insufficiently helpful. The case could be one about an oral promise which turns entirely on the word of one person against another's, and the uncontested facts may well not point towards A's version of events being any more plausible than B's. Even in a case which is fairly document-heavy (as this one was) there may be critical events or conversations which are completely undocumented. The CarbonDesk dinner is a good example. Whilst there are documents from which inferences might be drawn about what was or was not said at that dinner, there are no notes of the discussions and no memoranda or emails sent afterwards which appear on their face to record or report what was said on that occasion.

51. Faced with documentary lacunae of this nature, the judge has little choice but to fall back on considerations such as the overall plausibility of the evidence; the consistency or inconsistency of the behaviour of the witness and other individuals with the witness's version of events; supporting or adverse inferences to be drawn from other documents; and the judge's assessment of the witness's credibility, including his or her impression of how they performed in the witness box, especially when their version of events was challenged in cross-examination. Provided that the judge is alive to the dangers of honest but mistaken reconstruction of events, and factors in the passage of time when making his or her assessment of a witness by reference to those matters, in a case of that nature it will rarely be appropriate for an appellate court to second-guess that assessment.”

115. The difficulty in these references is not primarily about conflicts between witness evidence and documentary evidence (the exception being in relation to challenges made to the accuracy of the notes of the Traders’ interviews with Compliance); it is the absence of contemporaneous documentary evidence, or evidence of electronic communications.  Indeed, all parties relied on the absence of such evidence - the Authority relied on the absence of evidence of the Traders documenting their Trading Strategies, explaining them in advance to Mr Heiberg and/or Compliance, keeping records of their use and outcomes; and the Traders relied on there being no contemporaneous documentary evidence, or evidence of electronic communications, of collaboration during the Instances or of a conspiracy to commit market abuse.

116. It is clear that, even within a few weeks of the Relevant Period, the Traders were not able to remember the details of their specific trading activity.  In September 2016 Mr Urra had produced a note for Compliance and in his witness statement he said that at that time his recollection of the specific trades in question was limited, as they had occurred several months earlier.  Mr Lopez tried to explain his activity in F174 in an email to MHI on 19 October 2016 (following his disciplinary interview) stating that he was providing details “to the best of what I can remember” but at that time wrongly identified a particular RFQ as the trigger for his Large Order (which had not in fact been received at that time, a confusion which arose as a result of cash trades generally being recorded by reference to UK time whereas the Futures data from Eurex is on CET) on the basis of an attempted reconstruction of events.  By the time the Traders produced their witness statements, it was, as Mr Sheth said,  “simply impossible” to remember the details of the rationale for their trading activity.  The Tribunal is very sceptical of any Trader stating that they could recall particular trades out of hundreds (or even thousands) that had taken place months, let alone years, previously.

117. Mr Shivji referred the Tribunal to the speech given by Popplewell LJ on 7 November 2023 “Judging Truth from Memory”, in which Popplewell LJ focused on what the current state of scientific research tells us about when and why our memories don’t serve us well and to consider what lessons we can learn when it comes to the exercise of fact finding in commercial dispute resolution.  Popplewell LJ looked at what he termed the “Architecture of Memory”, including the stages of memory, describing them as (i) encoding (feeding the information into the memory system); (ii) storage (the preservation of information in the system); and (iii) retrieval.  On storage of long-term memory, Popplewell LJ referred at [60] and [61] to a recent study which showed that there was rapid forgetting of both flashbulb and event memories within the first year, but the “forgetting curves” levelled off after that, not significantly changing even after a ten-year delay

118. The Tribunal accepts the Traders’ submissions that they cannot reasonably be criticised for the absence of recollection as to the specifics of any particular order or trade in the Relevant Period; but also considers that this lack of recollection is almost inevitable.

Potential witnesses who were not called by the Authority

119. Mr George, Mr Jaffey and Mr Bailin were critical of the Authority’s failure to call any witnesses of fact, whether from its own enforcement team or from MHI, with Mr George referring to them in opening as “witnesses who are not here”.

120. The following individuals were identified by the Traders:

(1)        the Authority’s enforcement division;

(2)        Mr Hill and Mr Barouti;

(3)        Mr Heiberg, the Head of Fixed Income, who was Mr Urra’s direct line manager; and

(4)        Mr Joshi, the Head of Compliance, or other members of Compliance.

121.   In opening, reference was also made to Mr Halliday, who had been interviewed by the Authority in 2017, and “other persons engaged by MHI” in the Relevant Period.  However, none of the Traders made detailed submissions as to what material evidence such persons might be able to give, or what inferences should be drawn from their absence, and, having heard all of the evidence and submissions, the Tribunal does not consider that such persons would have material evidence to give as to issues of fact and does not address their absence further.

122. The Tribunal has on a number of occasions been critical of the Authority for failing to call witnesses who may have had relevant evidence to give.  In Forsyth v Financial Conduct Authority [2021] UKUT 162 (TCC) the Tribunal summarised the position as follows:

“75. Furthermore, as Mr George submitted, the principle enunciated in Wisniewski v Central Manchester Health Authority [1998] 1 PIQR 324 is relevant in this regard. As was stated at page 340 of the judgment in that case, in certain circumstances the court may be entitled to draw adverse inferences from the absence of a witness who might be expected to have material evidence to give on an issue in action. In circumstances where the reason for the absence of the witness satisfies the court, then no such adverse inference may be drawn but in circumstances where it might have been expected that a party would call a particular witness then such an inference may be drawn. If the court is willing to draw such inferences, they may go to strengthen the evidence adduced on that issue by the other party or to weaken the evidence, if any, produced by the party who might reasonably have been expected to call the witness.

76. We received no explanation as to why other witnesses who may have given relevant evidence as to the documents from which we were asked to draw inferences were not present, as mentioned at [65] above, and in those circumstances, we are entitled to draw adverse inferences from their absences. We have done so to the extent that we have given more weight to Mr Forsyth’s evidence as regards the documents in question and less weight to the evidence that the Authority sought to rely on in that regard.”

123. In Seiler UT the Tribunal recorded at [113] “it is the Authority who has the burden of proof in these proceedings and should be seeking to assist the Tribunal by bringing to it all relevant evidence”.  The Tribunal did not accept in those proceedings that Mr Campeanu’s participation in the transactions was peripheral - “he was involved at various critical points in the narrative”.

124. The Authority’s position in these references was that none of the Traders had previously suggested the Authority should have taken steps to call these persons as witnesses (the issue only being raised in written skeleton arguments), there is no evidence that they attempted to do so themselves, and, moreover, these persons would not have material evidence to give as the Authority’s case is not about the Traders misleading management or compliance, and it is accepted that Mr Hill and Mr Barouti did not hear anything suspicious.  In this situation, Mr Shivji submitted that there is no basis to draw adverse inferences against the Authority.  Mr Shivji reiterated that the Authority takes the criticisms by the Tribunal in Seiler UT seriously; the Authority would have helped if the Traders wanted assistance in calling evidence. 

125. The Tribunal has considered whether it should draw adverse inferences from the Authority’s decision not to call these persons as witnesses.  The potential relevance of the various individuals is considered below, but a theme of the Traders’ submissions was that the Tribunal should infer that they were not called as witnesses as they would not support the Authority’s case.   The Tribunal was taken to various parts of the transcripts of the interviews conducted by the Authority with Mr Hill, Mr Barouti, Mr Heiberg and Mr Joshi.  The Tribunal has, as invited by the parties, read the whole of those transcripts.  

Authority’s Enforcement Division

126. Mr George referred to what he submitted were the various procedural shortcomings in the Authority’s investigation, and that the Authority had not explained why it had not sought to call any witness from its Enforcement Division to answer these matters.  Mr George submitted that in circumstances where the Authority has offered no witnesses who can speak to these events, there can be no good reason for the excessive delay or the absence of relevant evidence in this case. 

127. The Authority’s position was that there is no factual dispute over the investigation and it is impossible to see how such a witness would have materially assisted the Tribunal.

128. The Tribunal agrees that the factual chronology of the investigation is not in dispute, and it can itself identify the periods of time which have elapsed at various stages.  During the course of the hearing it did become apparent to the Tribunal that there is a potential issue as to the steps which had been taken by the Authority to secure documentation and other evidence and review the evidence which it had obtained.  This arises most obviously from Mr Bailin’s submissions in relation to the communications data, but a question could also be asked as regards the Authority’s acceptance of MHI’s explanation in relation to Voice RFQs and Bloomberg/Chat RFQs.  However, whilst the Traders had in correspondence made requests for information, and such information had either been provided or refused with an explanation, the Tribunal does not consider that the Authority should reasonably have expected that it would have been of assistance to the Tribunal to call witnesses who could explain the evidence-gathering aspects of its investigation, particularly in the absence of an application for disclosure having been made. 

129. In these circumstances, no adverse inference is drawn from the Authority’s decision not to call witnesses from its own Enforcement Division.  That approach is not to ignore the potential significance of information which the Tribunal has already identified as not being available; rather that the Tribunal does not consider that issues arising from its absence should necessarily be resolved in favour of the Traders, or that evidence which supports their case should be assumed to have existed.

Other traders on the EGB Desk - James Hill and Mehdi Barouti

130. Mr Hill and Mr Barouti were the other two traders on the Desk.  They traded in Core EGBs rather than Peripheral EGBs.  They were both interviewed by the Authority in June 2018 and were asked about whether they heard anything suspicious between the Traders. 

131. The Traders submitted that Mr Hill and Mr Barouti made it clear that they could overhear conversations between Mr Urra, Mr Lopez and Mr Sheth, but at no point heard anything suspicious which would suggest that the coordinated market abuse alleged by the Authority occurred.  We were taken to parts of the transcripts of the interviews which supports this:

(1)        Mr Barouti said the Traders were “talking constantly” and was regularly party to or overheard the discussions that Mr Urra had with the other Traders about the Futures market, but everything he heard discussed was “very vanilla, you know, market conversation”; and

(2)        Mr Hill said he could overhear the conversations between the Traders and witnessed their interactions throughout the day but did not overhear any coordination between the Traders about the placement or cancellation of orders or anything that indicated they were engaging in spoofing.

132. The Authority’s position was that there is no factual dispute about what Mr Hill and Mr Barouti saw and heard.  It was, they said, common ground between the Authority and the Traders that they saw and heard nothing which led them to suspect that the Traders were engaged in anything untoward. 

133. The Tribunal accepts that Mr Hill and Mr Barouti would be able to give evidence relevant to issues relating to collaboration between the Traders and what was seen and heard by others.  The relevance of such evidence is not disputed by the Authority - instead the Authority says there is no relevant disagreement of fact.  On that basis, the Tribunal finds that Mr Hill and Mr Barouti heard nothing which led them to suspect that the Traders were engaged in anything untoward. 

134. Mr George also submitted that Mr Hill could give evidence in relation to the plausibility of the Information Discovery Strategy which was said to be pursued by Mr Urra (and Mr Sheth).  That overlaps with submissions made in relation to Mr Heiberg and Mr Joshi, and is addressed alongside those individuals below.

Management and Compliance at MHI

135. The Traders referred the Tribunal to various parts of the transcripts of the interviews conducted by the Authority to support their submissions that Mr Heiberg, Mr Joshi and Mr Hill had relevant evidence to give, including in relation to:

(1)        the 15% hit ratio introduced by Mr Heiberg as an informal target;

(2)        Mr Heiberg was satisfied with Mr Urra’s personal performance;

(3)        the recruitment of Mr Lopez and the medium-sized RFQs received by the Desk;

(4)        Mr Sheth’s junior status;

(5)        the Traders’ understanding that the Eurex Letter was about the placing of Multiple Large Orders was reasonable;

(6)        the placing of Multiple Large Orders was the kind of mistake that might be made by a junior trader;

(7)        the plausibility of the Trading Strategies, this being relied upon in respect of both the Information Discovery Strategy and the Anticipatory Hedging Strategy, eg:

(a)        Mr Hill had said “…if you actually show some good size on one of the less liquid contracts, somebody else is out there desperate to trade this thing, and they will take the opportunity to access that liquidity, which is great and valuable information, yes, so if you've got a client that -- you know, who wants to buy it, then you can show them a better price, because you know now that that actually, there is liquidity at that point”; and

(b)        Mr Heiberg had said he understood the argument of “trying to build a basis position, hoping to be hit at a lower level if someone needs the liquidity”, and that there might be anticipatory trading, as the only way you can have a competitive desk is if you have a portfolio ready to sell to clients; and

(8)        the Trading Strategies were within the mandate which applied to the Desk, titled “European Government Bond Trading Desk Mandate” (the “Mandate”), this also having been the conclusion of the MHI Compliance Report.

136. The Traders made submissions as to the adverse inferences that should be drawn, including that the Tribunal should infer that these individuals were not called as their evidence would not support the Authority’s case. 

137. The Authority accepted that the 15% hit ratio was an informal target which had been introduced by Mr Heiberg, and that Mr Heiberg was satisfied with Mr Urra’s personal performance.  The Tribunal finds as facts accordingly (which is consistent with the other evidence before the Tribunal, including the records of Mr Urra’s appraisals).

138. The Authority’s position was that the remaining matters were largely matters of opinion.  The Tribunal has the Eurex Letter, the Mandate and the MHI Compliance Report and can reach its own conclusions as to the enquiry raised by Eurex, the Traders’ understanding of that letter (on which they were cross-examined), the requirements of the Mandate and the conclusions reached by Compliance.  The Authority submitted that opinions as to the plausibility of the Trading Strategies can best be addressed by the evidence of the independent experts.

139. The Tribunal acknowledges that it has the Eurex Letter, the Mandate and the MHI Compliance Report (with no further review having been undertaken by Compliance since the date of that report) and agrees that it can reach its own conclusions on each.   

140. With one exception, the Tribunal agrees with the Authority that Mr Heiberg, Mr Joshi and Mr Hill do not have material evidence to give in relation to facts (rather than matters of opinion) which are in dispute - they were not trading in Peripheral EGBs in the Relevant Period, or otherwise involved in the trading within the Instance Pool, Mr Heiberg and Mr Joshi did not sit with the Traders, and the Traders had not discussed their Trading Strategies with them with a view to obtaining approval in accordance with the terms of the Mandate.  The exception is that Mr Heiberg did in his interview with the Authority provide further background to the recruitment of Mr Lopez and said that his intention was that Mr Lopez would focus on medium-sized RFQs.  It became clear during Mr Shivji’s cross-examination of Mr Lopez that the Authority did not accept Mr Lopez’s evidence that he had been recruited to focus on medium-sized RFQs.  The Tribunal has taken account of Mr Heiberg’s explanations given to the Authority during this interview when making its findings of fact. 

141. The Tribunal agrees with the Authority that the further matters identified by the Traders are primarily matters of opinion on which the Tribunal has the benefit of expert evidence.  Furthermore, the Tribunal rejects the Traders’ submissions that these persons were not called as they would not have been supportive of the Authority’s case.  That submission is contradicted by the evidence including the totality of those interviews.

142. At an early stage Mr Heiberg had expressed concerns about the explanations being given by the Traders for their trading, telling Compliance on 6 October 2016 that it was “not fully convincing”.  The Tribunal takes this to refer to the Information Discovery Strategy, as Mr Heiberg had been speaking primarily to Mr Urra and this was his explanation.  Mr Heiberg was then interviewed by the Authority on 17 July 2017. 

143. During the course of his interview Mr Heiberg said:

(1)        He had brought in Mr Lopez and Mr Barouti to focus on the medium-sized tickets and wanted to have a clear division of roles and responsibilities.

(2)        The Mandate was that the Desk did business with its clients and did not take their own positions.

(3)        Asked if they were only trading stuff off the back of an indication that comes from a client, he added “or something we could sell to a client”.  He agreed there might be anticipatory trading, as the only way you can have a competitive desk is if you have a portfolio ready to sell to clients - they try and guess what would be interesting for an asset manager.  Being asked further about anticipatory trading, he referred to the trading history of the particular client, talking to sales about what they think, whether they have anyone looking to buy something, watching the enquiries from clients.

(4)        The 15% hit ratio was fairly modest and would allow them to do more business with clients without forcing them to do unprofitable business.  Mr Urra had been quite optimistic; he was performing quite well in meeting the hit ratios.

(5)        He became aware of the Eurex Letter whilst he was on holiday.  Mr Joshi had emailed and called him; Mr Urra had also called him and explained that this was human error, a one-off and a storm in a teacup; but on the Monday he got more detail from Mr Joshi and got very concerned again.  He did get more explanations from Mr Urra that day but mainly discussed with Mr Joshi after that as he didn’t find the explanations completely convincing.  He found the pattern and number of incidents problematic.  He said that MHI’s response to the Exchange was portrayed as the Traders’ response, not necessarily the way it was viewed by MHI, saying the explanation was plausible but they were not fully convinced and it was “a big dilemma”.

(6)        He referred to trying to build a basis position, hoping to be hit at a lower level if someone needs the liquidity, as making sense.  It wouldn’t make sense using an iceberg for that because then you wouldn’t show the liquidity to the market.  He said he did understand the argument of showing a larger bid that could be hit at an attractive price and it could be useful.  It was important for the market maker to be in the market, and an important part of that is the futures market; but he reiterated that he was troubled by the frequency.  Asked as to whether he would have approved this strategy, he initially said probably yes, but then would have said no to the frequency, extent and multiples and would definitely have said no to having a smaller order on the other side.  He found it a little bit striking that Mr Urra did not discuss this with him.

(7)        On multiple orders, he has seen this type of mistake made before. 

144. Viewed as a whole, we do not accept the Traders’ submissions that Mr Heiberg’s evidence would not have been supportive of the Authority’s case.  The Tribunal considers that Mr Heiberg had expressed concerns about Mr Urra’s description of the Eurex Letter and the explanations Mr Urra provided in response.  He had described his relationship and interactions with Mr Urra and how Mr Urra was very communicative, and then said it was striking that Mr Urra had not discussed his new strategy with him.  Given the timing of this interview, which was before the Authority had interviewed Mr Lopez (although we infer that the Authority would have had copies of the notes of Mr Lopez’s interviews with MHI), and the way in which the strategy is described, we infer that Mr Heiberg’s comments on understanding the approach and the concerns he expressed are about the Information Discovery Strategy.  Mr Jaffey submitted that Mr Heiberg had endorsed the Trading Strategy used by Mr Lopez, ie the Anticipatory Hedging Strategy, submitting that he made significant points about the legitimacy of this as a market making strategy and the sizing of the orders.  Whilst we agree that Mr Heiberg made significant points about Mr Lopez’s intended role at MHI, we are not persuaded that Mr Heiberg’s observations on anticipatory trading were an endorsement of Mr Lopez’s strategy - our reading of Mr Heiberg’s explanation is that he had not identified at that time that Mr Lopez was giving a different explanation to that of Mr Urra, and his references to anticipatory trading were in the context of anticipating the cash bonds that clients may seek to buy and building an inventory of such bonds.

145. The Compliance department’s conclusions on the activity were set out in the MHI Compliance Report, based on the review which it had conducted at that time.  That report refers to “sufficient doubt” as to the credibility of the explanations and expressed its conclusion as “indeterminate”.  The note of the Authority’s subsequent meeting with Mr Joshi (and the CEO of MHI) on 2 November 2016 included in the summary that MHI found the traders’ activity unacceptable but could not find unequivocal evidence that the Traders’ intention was to manipulate the market. 

146. In his interview with the Authority on 3 October 2017, Mr Joshi explained how MHI monitored traders, the various levels of surveillance and alerts which were in place, and the steps they took following receipt of the Eurex Letter.  The transcript of the interview records Mr Joshi as having said “we weren’t completely sure as to why these orders had been placed”; and price discovery is a good explanation (which was “logical, makes sense”) but they weren’t sure whether that was the “real reason”.  Mr Joshi explained that Compliance had received the data for the sampling exercise before MHI responded to MSUSA and the Exchange and had identified other examples of large orders on the other side of small orders and they thought this was “a little bit unusual”.  Mr Joshi said that MHI gave the Exchange a factual response based on their investigation and what MHI had found out about the trading from the Traders.  Mr Joshi emphasised that the Exchange was not asking MHI to do an investigation, or whether this had occurred on other days, or whether MHI considered this could be market abuse - he made the point that the Exchange had their own surveillance department.  

147. The Tribunal considers that Mr Joshi was distancing Compliance from the explanations given by the Traders (which would have been just Mr Urra and Mr Sheth at that time, as Mr Lopez had not been interviewed before MHI responded to Eurex) and was very careful not to say that he had accepted their explanations.  He referred to Mr Urra and Mr Sheth as having given the “primary, clear explanation” that this was for price discovery purposes.  It was also not apparent that Mr Joshi had identified at any time that Mr Lopez’s explanation for his trading (which was given after MHI replied to the Exchange but before Compliance finalised the MHI Compliance Report) was different to that of Mr Urra and Mr Sheth.  In any event, the Tribunal does not consider that the evidence supports the Traders’ submission that Mr Joshi was not called as he would not have been supportive of the Authority’s case.

148. Mr Hill had been interviewed on 8 June 2018 (ie after the Authority had interviewed the Traders, albeit only Mr Lopez had provided substantive answers to the Authority’s questions) and was asked about the Information Discovery Strategy, and did refer to somebody else being “desperate to trade” and taking the opportunity to access the liquidity.  However, Mr Hill seems to be talking about price discovery, ahead of a client trade, referring to being able to show a client a better price where there is liquidity at that price point, and emphasised that the question of how many lots to offer is very subjective and will depend on the market at the time (volatility, width of spread, time of day) and what you are trying to achieve.  His opinion on plausibility is significantly constrained by these differences, and the Tribunal, on reading the whole of the transcript of this interview, is not persuaded that the Authority would have concluded that his evidence would be unhelpful to them.

149. Whilst we do not draw adverse inferences from the Authority’s decision no to call these persons as witnesses, the transcripts of the interviews are nevertheless hearsay evidence before the Tribunal and we have taken them into account alongside all of the evidence before us, in particular in relation to Mr Sheth’s status on the Desk, the recruitment of Mr Lopez, Mr Heiberg’s view of the hit ratio and the way in which MHI conducted monitoring of the activities of the trading floor.

Approach of the Tribunal

150. The Tribunal is required to make findings of fact in relation to trading activity which took place more than 8.5 years ago.  In carrying out our fact-finding, we remind ourselves that the burden is on the Authority to satisfy us as to what was more likely than not to have happened in relation to any particular event. 

151. The Tribunal takes account of the following: 

(1)        The allegations of market abuse are very serious allegations for financial services professionals, with serious consequences for their future career and substantial penalties.  The seriousness of the allegations here may mean that it is less likely that the event occurred such that the stronger should be the evidence before we conclude that the allegations are established on the balance of probabilities. 

(2)        The Traders referred to Carnwath LJ’s speech in Mohammad Jafari-Fini v Skillglass Ltd & others [2007] EWCA Civ 261 where he referred to the exhaustive review of the authorities on the standard of proof in R(N) v Mental Health Review Tribunal [2006] 2 WLR 850, and said as follows:

“40. …Thus in civil proceedings, the “presumption of innocence” is not so much a legal rule, as a common sense guide to the assessment of evidence. It is relevant not only where the cause of action requires proof of dishonesty, but, wherever the court is faced with a choice between two rival explanations of any particular incident, one innocent and the other not. Unless it is dealing with known fraudsters, the court should start from a strong presumption that the innocent explanation is more likely to be correct.”

(3)        However, it has been repeatedly emphasised that there is no heightened standard of proof where serious allegations are made.  In Bank St Petersburg PJSC v Arkhangelsky [2020] EWCA Civ 408 Males LJ said at [117], before citing the judgment of the House of Lords in In re B (Children) [2008] UKHL 35:

“117.  In general it is legitimate and conventional, and a fair starting point, that fraud and dishonesty are inherently improbable, such that cogent evidence is required for their proof. But that is because, other things being equal, people do not usually act dishonestly, and it can be no more than a starting point. Ultimately, the only question is whether it has been proved that the occurrence of the fact in issue, in this case dishonesty in the realisation of the assets, was more probable than not.”

(4)        The Tribunal recognises that caution is required where a party builds their case on indirect evidence and inferences.  Nevertheless, as identified in Lakatamia Shipping Co Ltd v Nobu Su [2023] EWHC 1874 (Comm) at [102] “Conspiracy claims are invariably based on circumstantial evidence, and inferences to be drawn from the material that does exist….”.

(5)        The Tribunal has been presented with competing explanations of the trading activity and needs to assess those explanations in the light of all the evidence that is available, and taking account of the evidence that is not available, which includes information that would have been available to the Traders during the Relevant Period. 

(6)        Whilst the Tribunal accepts that the Traders’ inability to recall the specific details of this activity is understandable and that they cannot be criticised for this, the Tribunal does not accept the Traders’ submission that the Authority is not entitled to rely on an absence of evidence as supporting any assumption it seeks to make about the Traders’ intentions or rationale during the Relevant Period.  Such an approach would come close to deciding that all factual issues in relation to which there is no supporting documentary evidence should be resolved in the Traders’ favour.  The burden of proof is on the Authority to make its case, and when assessing whether the Authority has met this burden the Tribunal takes account of the evidence that is available (including, eg, the Electronic RFQs received, and cash bond transactions executed) and how the Traders explain their activity by reference to this evidence (eg as to client demand, or the identification of clients against which the Information Discovery Strategy was being tested).

Findings of fact

152. We make the following findings of fact.

EGBs, market making, BTPs and BTP Futures

153. The SOABF sets out the agreed background.

154. During the Relevant Period, the Italian government was a large issuer of debt and the largest component of this debt was BTPs.  BTPs were traded in MTS and Brokertec (electronic bond exchanges) and through the voice broker market. 

155. SOABF [7] refers to the obligations and costs associated with being a Primary Dealer in BTPs.  However, Primary Dealers have access to superior market information compared to all other market participants.  Market information comes from their obligations as Primary Dealers to purchase newly issued bonds, provide quotes in the secondary market and their high level of client and wider market engagement.

156. SOABF [9] and [10] record that market making in EGBs is very competitive, and that providing competitive prices involves traders continuously assessing what prices they are prepared to offer by reference to their own book and market conditions.  In addition, we find that market making is extremely dynamic and volatile.  To remain competitive, market makers continuously need to monitor supply and demand and adjust prices to reflect changing market conditions, and manage any acquired positions to ensure they are in a position to offer the most favourable pricing in both directions.

157. To state the obvious, to win an RFQ a market maker must offer a more competitive price than rival market makers.  However, without sufficient information or adequate pre-positioning, a market maker may find that securing an RFQ results in losses (if they are not able to hedge the order profitably).  For smaller market makers, this challenge is particularly acute.     

158. SOABF [19] records that cash bonds are balance sheet and capital intensive, whereas Futures require limited capital until the futures contract expires and the cash bond(s) are delivered.  In addition, we find that there were typically very different spreads in the cash market and the futures market.  The spread in the cash market was generally about four ticks, whereas the spread in the Futures market was generally one tick.  This meant that where a trader needed to hedge a trade it had executed in the cash market, it would be more expensive to cross the spread in cash than in Futures. 

159. Most market makers, having traded in the cash bond market and who needed to hedge their risk, would hedge with Futures and would do so swiftly in order to manage their exposure to market movements.  A market maker may not need to hedge if their portfolio had not been flat prior to the cash trade.

160. Futures may be traded electronically on Eurex.  The exchange is anonymous, ie as a market participant it is not possible to tell who (whether institution or individual) has placed the orders. 

161. Orders for Futures can be placed on Eurex at the current best price available in the market, ie the touch, or a different price.  Orders placed at the touch are more likely to execute promptly, but orders do not need to be placed at the touch in order to trade.  The quantity of lots available at any price point will vary. 

162. Addressing orders which are not at the touch:

(1)        Trading behind the touch - a trader may choose to place an order at a level behind the touch, ie if you are buying, it is at a level below the Best Bid, and if you are selling, it is a level above the Best Offer.  This is called a passive order, in that you are waiting for others to trade with you at the chosen price.  

(2)        Aggressing the market - a trader seeking immediate execution can “aggress” the market (also described as “crossing the spread”) by entering an order to buy at a price at, or better than, the Best Offer (or, in the case of selling, an order to sell at a price at, or lower than, the Best Bid) and immediately trade.  This, evidently, comes at a cost and there can be a multitude of reasons why a market participant would decide to do this.

163. The ability to iceberg an order in Futures is described in SOABF [24].  As it is a feature of the trader’s systems, it may be affected by the latency of a trader’s computer systems.  The process of the Exchange sending a message to the trader’s systems that the visible slice has been executed and the trader’s system placing an order for the next slice may give rise to a delay between the execution of one slice and the placing of the next.  However, on the basis of the specific timings of the trading activity in the Instances, which included the trading of Small Orders which had been iceberged by the Traders, the Tribunal finds that such latency or delay can be measured in milliseconds. 

164. As the use of iceberging is widespread and well-known, market participants are aware that actual liquidity in terms of available bids and offers is likely to be higher than what is visible on the stack, as portions of any given order may remain hidden as iceberged orders.  As (can possibly be seen) in Appendix 2, the presentation of the stack in the Replay graphs in the Instances did show the individual orders at each price point (albeit only the visible non-iceberged portion of each order) and the horizontal scale meant that our assessment of order size was estimated rather than precise.  It was not the Authority’s case that this is how the stack would have appeared to market participants at the time, and we were taken to screenshots of Bloomberg Escalator in the MHI Compliance Report which had been provided to Compliance by the Desk shortly after the Relevant Period.  That presentation showed the total number of visible lots available at each price point away from the touch, and we also find that the total number of orders live on the stack was stated.   

165.   The use of iceberging means that the stack shows only the visible liquidity at any point of time - the actual liquidity is not known.  In addition, counterparties could also agree block trades of 250 lots or more off market; these would be reported to the exchange after execution.

The Traders - roles at MHI and experience

166. Until the matters which are the subject of the references before us, none of the Traders had been the subject of any regulatory investigation or disciplinary sanction.    

Mr Urra

167. During the Relevant Period Mr Urra was Head of European Rates Trading and Head of the EGB Desk at MHI.  These roles involved oversight and management of MHI’s market-making activities across a variety of securities, and he was responsible for the supervision of the other traders on the Desk, ie Mr Lopez and Mr Sheth as well as Mr Hill and Mr Barouti.

168. By the Relevant Period, Mr Urra was an experienced trader, having worked in the financial services industry for 18 years and with 15 years’ experience in trading EGBs, and was well-regarded in the industry.  He began his career at JPMorgan in London, and later worked at Lehman Brothers, Standard Chartered Bank and Credit Suisse.  He had gained extensive expertise as a market maker, with a particular focus on Government bonds. 

169. Mr Urra joined MHI in October 2013 as Head of Government Bond and Swaps Trading and was promoted to Head of European Rates Trading in 2015.  In this role Mr Urra reported to the Head of Fixed Income Trading, Mr Heiberg, and his primary responsibilities as set out in his job description included: 

“To manage and develop the European Rates Platform at MHI.  To work in conjunction with the Head of Fixed Income trading to build a Rates platform capable of delivering recurrent client-flow orientated revenues and contributing positively to the Fixed Income bottom-line.

…

• To evaluate and manage the existing team of EGB traders with a focus on co-ordinating risk taking from a top down perspective

• To be responsible for the collective risk of the EGB desk in respect of Front Office and Risk Committee limits - to work to enhance the sophistication of the relevant limit infrastructure and controls

• To ensure pro-active risk management of positions, right sizing the risk appetite and balance sheet utilisation according to liquidity conditions and client flows

• To market-make and trade the 0-5 yr part of the curve for all countries with a view to enhancing client flows and generating revenues”

170.  Then:

“Managing Performance. As a manager it is a requirement of your role to ensure that each member of your team is competent to perform their role...

Overall Statement of Responsibility. All employees of the Company at whatever level of seniority have responsibility on a day to day basis for ensuring that there are clear and appropriate reporting lines between them and others for whom they are responsible and that any delegation of duties is to suitable persons, subject to appropriate supervision and monitoring.”

171. Mr Urra had a range of responsibilities, and they included managing all EGB trading across both the Core and Peripheral markets, supervising repo finance trading, overseeing all EUR- and US-denominated interest rate swap trading by MHI as well as leading interest rates strategy research.  He would therefore spend only part of his time working on his responsibilities as Head of the EGB Desk, and this would include both managing the other traders on that Desk and trading on his own book.  However, Mr Urra only had one physical desk at MHI, which was with the other traders on the EGB Desk.

172. Mr Urra was, and was recognised as being, a successful trader at MHI and Mr Heiberg was satisfied with Mr Urra’s personal performance as a trader.    

173. Mr Urra left MHI in September 2019.

Mr Lopez

174. Mr Lopez was a Director and Senior Rates Trader at MHI during the Relevant Period.

175. Mr Lopez was an experienced trader.  He had worked at Credit Suisse and BNP Paribas in non-trading roles early in his career.  He returned to Credit Suisse in July 2007, with a sales role in fixed income derivatives and structured products.  In September 2012 he transferred to work as a junior EGB trader on the fixed income desk; this was his first job as a trader.  He reported in that role to Mr Urra (who had joined Credit Suisse in August 2011) until September 2013, when Mr Urra moved to MHI.  Mr Lopez obtained his own book in June/July 2013 and was promoted in January 2015 but left Credit Suisse in December 2015, as Credit Suisse informed him that they would be closing the desk.

176. Mr Lopez joined MHI in February 2016 as Director and Senior Rates Trader in the Fixed Income Sales and Trading Department on the EGB Desk.  Mr Urra had become aware that Mr Lopez was looking for a new role and informed the relevant individuals at MHI - Mr Lopez was interviewed by a number of individuals at MHI before being offered the role.  Mr Lopez had been in advanced discussions in relation to a potential role at a hedge fund - that role would have involved more money, but he accepted the role at MHI as it was within his expertise and he wanted to prioritise spending time with his family.

177. Mr Lopez was still in his six-month probationary period at MHI during the Relevant Period.  Nevertheless, he had his own book and was afforded a lot of independence in his role.  Mr Heiberg’s opinion of him was that he was a “highly responsible individual with a strong moral compass”. 

178. Mr Lopez left MHI in February 2020.

Mr Sheth

179. Mr Sheth was an Associate EGB Trader at MHI during the Relevant Period.

180. Before joining MHI, Mr Sheth had studied Electrical Engineering and Communication and Signal Processing (including writing a thesis on the Black Scholes option pricing model), worked for almost three years as an interest rates analyst at a proprietary trading firm, Futures First, in India and then studied for a Masters in Finance at the London School of Economics. 

181. Mr Sheth joined MHI in June 2014, initially as an intern.  In September 2014 he was promoted to Analyst on the Desk, where his job description was that he was an EGB Trader.  Mr Sheth’s role involved both analysing Mr Urra’s book and making recommendations and trading on Mr Urra’s book.  He also worked on analysis and financial models during this time.  Mr Sheth was promoted to Associate in April 2016 but his job description remained that of EGB Trader.  He was given his own book to manage at that time, the ETA book - small sizes of EGBs (less than €5m) were traded in this book, which used an auto-quote algorithm and was an active book designed to develop relationships with a select group of clients with whom MHI wished to do larger trades.  Mr Sheth managed the risk in this ETA book by hedging cash trades with Futures and monitoring the balance sheet.

182. Mr Sheth was the junior trader on the EGB Desk; he was the least experienced of all of the traders on that desk, was described as junior by both Mr Hill and Mr Barouti, and this was reflected in his remuneration (his salary was £70,000, and his bonus for the year ended March 2016 was £20,000).  As with all of the other traders on the Desk, his supervisor was Mr Urra.

183. Mr Shivji nevertheless submitted that Mr Sheth had significant financial markets experience by the time he joined MHI in June 2014.  The Tribunal agrees that Mr Sheth was well-educated in financial markets by this time and understood risk, and had some practical experience (from his time at Futures First) in managing his own book, but in a very stable market where he had very low risk limits.  The Tribunal does not, however, agree that Mr Sheth was, during the Relevant Period, an experienced industry professional.  He was the junior member of a team, and still learning and gaining experience.

184. Mr Sheth left MHI in October 2020.

MHI and the EGB Trading Desk

Volcker Rule and the EGB Desk Mandate

185. The SOABF [11] records that US legislation known as the “Volcker Rule” prohibits banks from engaging in proprietary trading, ie short-term trading for their own profit and refers to one of the statutory exemptions from this general prohibition for “market making-related activities”.

186. Under the Volcker Rule, MHI was required to have an internal compliance program designed to ensure compliance with the requirements, including by written policies and procedures and internal controls.  This includes policies identifying the instruments in which the trading desk is a market-maker, the actions it will take to “demonstrably reduce or otherwise significantly mitigate promptly the risks of its financial exposure”, including “the products, instruments and exposures each trading desk may use for risk management purposes” and the “techniques and strategies each trading desk may use to manage the risks of its market making-related activities and inventory”. MHI was also required to identify (and enforce compliance with) limits for each trading desk, “based on the nature and amount of the trading desk’s market making-related activities”.

187. The Desk’s Mandate was prepared in compliance with the requirements of the Volcker Rule.  There were two different versions of this mandate in force during the Relevant Period but the differences were very minor.  The findings below are based on the Mandate with an effective date of 1 July 2016, and we find that the earlier version of the mandate, which applied during June 2016, had equivalent provisions.

188. The Mandate identifies the Desk Head as Mr Urra and the Division Trading Head as Mr Heiberg and applied to the EGB Desk (ie both Core and Peripheral EGBs).  The Mandate provides that the Desk Head retains ultimate responsibility for the Desk’s adherence to the Mandate, records that the Desk is relying on the market making exemption from the Volcker Rule, and was signed by Mr Urra.  The key relevant provisions were:

(1)        “Desk Mission & Strategy” - “To provide a client facilitation trading service in European Government Bonds (EGBs) and associated products.  Client facilitation trading includes market-making and active trading to support client business…”

(2)        “Authorised Instruments” - This lists various products and has columns headed “Market Making” and “Hedging”, and detailed footnotes.  “Government Bonds & Bills” are listed with a tick in the Market Making column, and “Government Bond futures” are listed with a tick in the Hedging column.  This section separately lists Exchange Traded Interest Rate Futures and ticks both Market Making and Hedging, but the footnote sets out that market-making approval for these instruments relates only to block trades in response to client demand. 

(3)        “Risk management” - “The Desk’s market making activities give rise to the following principal market risks: interest rate risk, curve risk and basis risk.  The Desk may purchase or sell hedging instruments, as set out within this mandate, in order to mitigate these risks.”

(4)        Annex A1 - Authorised Instruments - Overall trading and hedging strategies must be evaluated and approved by the Desk Head and Division Trading Head (after consultation with the Risk Management and Compliance departments).  The Desk Head and Division Trading Head must demonstrate that proposed new or amended strategies will not present a high risk to MHI, can be managed within limits, will be appropriately controlled and will comply with the Volcker Rule’s requirements.

(5)        Annex A3 then expands on Risk Management - The Desk must conduct its trading activities within the limits set by the Fixed Income Division Head and by MHI’s Risk Management Department, and they must be monitored on a daily basis.  The limits distinguish between market maker inventory and hedging positions, and the MHI limit for hedging positions is that hedges are not to increase market risk.  The Desk must manage exposures within these limits either by selling down long or covering short market-making inventory positions, or selling or purchasing authorised hedging instruments.  This section then provides:

“Transactions in hedging instruments must not be conducted, where such transactions increase the Desk’s market risk exposures (other than as described below), and hedge positions must be closed out once underlying inventory positions have been unwound.

The Desk may execute transactions in hedging instruments in anticipation of a highly likely near term exposure to risk, where there is a sound risk management rationale for such anticipatory hedging.”

(6)        A4. Market making policy - The Desk “must routinely stand ready” to purchase and sell one or more types of instruments authorised under this Mandate and must be willing and available to quote or transact in these instruments for its own account, in commercially reasonable amounts and throughout market cycles.  The size of the Desk’s market making inventory must be designed not to exceed, on an ongoing basis, the reasonably expected near-term demand of market making customers.  This assessment must be based on market conditions and a demonstrable analysis of historic customer demand and other relevant factors.

Risk Management and Limits

189. The Mandate refers to the limits set for the Desk, and they were set out in the MHI Compliance Report.  They included a Value at Risk (“VaR”) and a “PV01 Hedge” limit.  The PV01 Hedge limit monitors the sensitivity of the portfolio to a movement of one basis point in interest rates. 

190. These limits were assessed across the whole Desk (ie both Core and Peripheral markets) and applied at the end of the day:

(1)        the VaR limit was £2.5m; and

(2)        the PV01 Hedge limit was $60,000. 

191. There were no intraday limits (although, as set out in the context of Monitoring of activity at [215] to [220] below, Mr Heiberg did receive alerts of exposures during the course of the day).

192. The MHI Compliance Report set out the Desk’s Futures positions and risk limit utilisation at the end of 29 June 2016 (ie the date of the trading activity which had been questioned by Eurex).  The Desk was holding 2,287 lots of Futures, with a market value of £266m.  Its risk limit utilisation at the end of that day was a VaR exposure of £1.01m and flat PV01.  The report stated that the Desk “was operating well within its market risk limits at this time”. 

MHI’s EGB Business

193. As recorded at SOABF [27], MHI was a “third-tier” market maker in EGBs, and the principal purpose of the Desk was to service client demand.  Given its position in the market, MHI had less visibility over client flows in the market than top tier market makers.  It would receive fewer RFQs and received RFQs of smaller sizes.  The Desk saw around 20-25% of the overall market activity, and the Tribunal finds that it had an information disadvantage in a very competitive market.

194. During the Relevant Period, the average size of Electronic RFQs for cash BTPs received by MHI was €5.1m.  Only 3% of Electronic RFQs in cash BTPs received by the Desk were for €20m of bonds or more and had a maturity of five years and over remaining, for which 200 lots or more of Futures would be used as a hedge or partial hedge.  The Desk had a low hit ratio when it came to winning cash trades valued between €15m to €25m with a maturity over five years - Mr Lopez only won three such RFQs in the Relevant Period.

195. The Desk generated its profits by capturing the Bid-Offer Spread and earning revenue on the inventory of securities it held from time to time.  The Tribunal accepts the evidence of Mr Urra and Mr Lopez, which was supported by Mr Heiberg, that the Desk could not just be reactive: to be ready and competitive, they needed to take the risk of holding an inventory of cash bonds that they anticipated might be interesting or attractive, in anticipation of client demand and market movements.    

196. The process for submission of and responding to RFQs is set out at SOABF [12] to [15].  RFQs are directed by clients at selected market makers chosen by the client, not the entire market, and RFQs are often sent to multiple firms simultaneously.

197. RFQs could be submitted by clients in different ways, namely Electronic RFQs, Voice RFQs and Bloomberg/Chat RFQs.  The combined time that the original RFQ and the market maker’s price is live would typically be no more than five to ten seconds.  The client will select and deal with the best price.

198. When the Desk carried out a client cash BTP trade of any size, it would generally hedge its position swiftly using Futures in order to manage the directional risk associated with the cash trade.  MHI did not have an auto-hedging system, but did have software which assisted in calculating the size of the hedge.  Such hedging would not be necessary if the order book was itself not flat at the relevant time and could therefore accommodate a particular risk.  However, the Tribunal finds that the Traders generally tried to keep their books flat.  This enabled them to stand ready to quote competitively on both sides of the market.

199.  The Traders transacted all transactions in Futures manually, whether entering, amending or cancelling orders. 

200. All of the orders in Futures placed by the Traders on the Exchange were executable. 

Financial Targets

201. The “Desk Mission” in the Mandate was to provide a client facilitation service.  It was still an objective of MHI for the Desk to generate a profit and to develop the client business flow.

202. Around the beginning of MHI’s financial year 2016/17:

(1)        the Desk was set a target of £8.5m for 2016/2017; and

(2)        Mr Heiberg, who wanted to expand MHI’s business and for it to win more RFQs, told the Desk in May 2016 that he wanted to increase the Desk’s overall “hit ratio” to 15%.

203. This 15% hit ratio was not clearly defined at that time, either in writing or orally.  Mr Heiberg said (to Compliance on 6 October 2016) that he had not put the Desk under particular performance pressure and that they were generally in line with budget and (to the Authority in July 2017) that he expected to see performance in terms of both notional value of the trades and the pure number of trades.  Mr Heiberg considered that Mr Urra was already meeting this 15% hit ratio. 

204. The Tribunal finds that this hit ratio represented an increase on what the Desk as a whole had previously been achieving, Mr Urra had been meeting this ratio individually, it was an informal target, and was not introduced as one of the defined performance objectives for any of the Traders.

205. In relation to the target of £8.5m:

(1)        Whilst this target was generally referred to by the parties as a “P&L” target, it was not consistently explained in the evidence to which the Tribunal was referred as to whether it was a target for profitability or revenue.  The MHI Compliance Report referred to this as a revenue target, setting out P&L numbers separately (apparently showing the Desk as loss-making), and Mr Heiberg confirmed to the Authority that this was “pure revenue”.  The difference is clearly commercially significant; however, the Tribunal proceeds on the basis that the measure is less significant than whether it was met and indeed accepted as being met by Mr Heiberg and others within MHI.

(2)        The Desk had not achieved its target in the two preceding financial years.  The target had been £12m for 2014/2015 but in Mr Urra’s appraisal for that year it was recorded that at the end of February 2015 the Desk had achieved £6m.  The target for 2015/2016 was reduced to £8.5m, but this was not achieved.

(3)        The Desk’s target then remained unchanged at £8.5m for 2016/17.  It was described by Mr Heiberg as “modest”, and he drew attention to the changes in personnel on the Desk (including the recruitment of Mr Lopez and Mr Barouti).

(4)        Mr Urra had a personal target of £5m.  For 2014/15 his target had been £3m, which he had easily exceeded, with performance of £5.7m by the end of February that year.  For 2015/16 his target was £5m, and whilst he was said to be on track to achieve this in his performance review (having generated £4.27m to date), he fell short by the end of the year.

(5)        Mr Lopez was on probation during the Relevant Period.  His target was £1m for the financial year.

(6)        Mr Sheth also had a target of £1m.  Mr Sheth’s evidence was that he had been unaware of the target during the Relevant Period.  On the basis of his appraisal form for that year (both the stated objectives and his own comments thereon) the Tribunal finds that he was aware of this target by the end of the financial year, ie in February or March 2017.  The Tribunal accepts Mr Sheth’s evidence that he had not known during the Relevant Period that he had this specific target; we do, however, find that Mr Sheth did know that he, like all traders, was expected to generate an overall profit from his trading.

206. The Traders’ books were marked to market at the end of each trading day.  The Desk would generally make less than a tick on cash trades.  The Traders would be required to explain any loss of €50,000 or more.

Remuneration

207. The Traders each had a base salary alongside a discretionary bonus structure. 

208. Bonuses were discretionary and based on performance, including by reference to performance objectives which were measured on a Balanced Scorecard which took into account several factors:

(1)        40% of the appraisal weighting for Mr Urra’s performance objectives related to P&L performance of the Desk (including the Desk as a whole and his own P&L generation) and was the biggest single determinative factor in his appraisal. 

(2)        40% of the appraisal weighting for Mr Lopez’s performance related to his own personal P&L performance.  This was the biggest single determinative factor in his appraisal.

(3)        21% of the appraisal weighting for Mr Sheth’s performance related to revenue generation, for which he had a target of £1 million.  30% of his appraisal related to management of the ETA book, with a specific target hit ratio.  These objectives were together the biggest determinative factors in his appraisal.

209. Staying within MHI’s risk limits and staying on top of risk management also formed part of all of the Traders’ appraisals.  

210. For 2016/2017, their salaries were £400,000 (Mr Urra), £135,000 (Mr Lopez) and £70,000 (Mr Sheth).  We make the additional findings in relation to bonuses:

(1)        For 2014/2015 Mr Urra’s total bonus award was £430,000 in both cash and shares (with £86,000 cash and £86,000 shares payable upfront and the remainder being deferred).  For 2015/2016, Mr Urra’s total bonus award was £325,000 (with £97,500 cash and £97,500 shares payable upfront and the remainder being deferred).  His proposed bonus for 2016/2017 was £400,000 but all of that was withheld.

(2)        Mr Lopez had only recently joined MHI.  His proposed bonus for 2016/2017 was £50,000 but that was withheld.

(3)        For 2014/2015 Mr Sheth’s total bonus award was cash of £4,750 and for 2015/2016 it was cash of £20,000.  His proposed bonus for 2016/2017 was £100,000 but that was withheld.

Training

211. Each of the Traders had received training at MHI on market manipulation and market abuse.

212. Each of the Traders had completed MHI’s online “Market Conduct 2015” training, which required those being trained to complete exercises and answer questions as they progressed through the module.  That training included an example headed “Adding some liquidity” as follows:

“A trader is looking to unload 3,000 shares of QuestQuo plc, a relatively illiquid stock with an average order size of 1,500 shares on either side of the order book.  As the price looks a little low, the trader placed an order to buy 100,000 shares just under the lowest bid on the market.  As a result of the large order, there is an uptick in the market price, upon which the trader places an order to sell his 3,000 shares into the market, which is fully executed.  The trader then immediately deletes his buy order.”

213. The answer on the following slide is then:

“Making trades purely to manipulate market prices is prohibited.  The traders’ behaviour is known as spoofing/layering and is very likely to be seen by the regulators as market manipulation.”

214. The Tribunal finds that the Traders understood the principles underpinning the rules against market abuse, understood what spoofing was, and that spoofing (and indeed placing any orders without an intention to trade) constituted market abuse.

Monitoring of activity

215. There were various levels of monitoring which applied during the Relevant Period:

(1)        MHI monitored (and recorded) electronic messages and telephone communications of all traders on the trading floor, including the Traders. 

(2)        MHI monitored and maintained records of trades executed by the Traders on Eurex. 

(3)        Mr Heiberg had access to the Desk’s “live positions”, ie executed trades, throughout the day using a “Bento” app.  He would check these positions throughout the day, and he would receive alerts to exposures.  Such exposures would include the situation where a cash trade was executed and the risk was not hedged immediately.    

(4)        The Desk’s overnight positions were monitored by Mr Heiberg and by Compliance to check that the Desk’s book was within its limits.

216. MHI did not monitor orders in Futures placed on Eurex; the monitoring tool used at that time, TransWatch, did not handle transactions in Futures.  Orders (and cancellations) were monitored by MSUSA, as described in the MHI Compliance Report:

“The Compliance function conducts transaction and communications surveillance, in accordance with a risk based approach, as the second line of defence function responsible for monitoring market conduct.

An enhanced market abuse monitoring system (Trading Hub) was selected in 2015 and is in the process of being fully implemented;

At the time at which the order placing and cancellation practices described above were taking place, neither the Front Office nor Compliance were able to monitor such activity, as orders were routed through MSUSA and MHI maintained records of trades executed, but not orders. Orders were being monitored by MSUSA. Arrangements are currently underway to ensure that Eurex orders are captured within Trading Hub.”

217. No alerts or questions were sent by MSUSA to MHI in respect of the Traders’ activity in the Relevant Period.

218. Each of the Traders expected and understood that their communications and online activities would be monitored and recorded by MHI.  The Traders expected and understood that their trading on Eurex would be monitored by the Exchange, and that other market participants may flag to the Exchange any behaviour which they considered to be of concern.

219.  The Tribunal finds that the Traders did not know the specifics of the monitoring that was conducted by MHI, save that they were aware that Mr Heiberg had access to their live positions throughout the day and that, as limits applied at the end of the day, Compliance would have access to their positions at the end of the day.  That the Traders would not have known the specifics of the monitoring was clear not only from the evidence of the Traders, but is reflected in Mr Heiberg’s explanation of the monitoring to the Authority, where he explained the process of introducing “Trading Hub” as a replacement tool for TransWatch, as he himself had little visibility into the exact configuration of the systems, and no details of exactly what they could monitor, making the point (also reflected by Mr Urra) that it probably makes sense that the people being monitored don’t have a “perfect understanding” of how the monitoring operates. 

220. The Tribunal finds that the Traders did not, at the beginning of the Relevant Period, know that orders in Futures were not being monitored by MHI. 

Traders’ roles on the EGB Desk and interactions

221. As Head of the Desk, Mr Urra was responsible for supervising Mr Hill, Mr Lopez, Mr Barouti and Mr Sheth.  Mr Hill and Mr Barouti focused on Core and semi-Core EGBs, and did not generally deal in BTPs.

222. The Desk was located in an open-plan working environment throughout the Relevant Period.  The Traders sat next to each other, with Mr Urra in the middle.  MHI moved premises during the Relevant Period, but this seating layout remained the same.  In the first premises, the desk on which Mr Hill and Mr Barouti sat was diagonally behind that of the Traders; in the second location, Mr Hill and Mr Barouti sat in front and facing them (divided by the Traders’ screens and their own).  In both locations, the Desk was near to desks of other MHI employees.

223. Whilst Mr Urra had a number of roles and responsibilities within MHI, his only desk was at the EGB Desk, and he spent most of his time there.

224. In terms of the way in which the Traders operated:

(1)        The Traders did not have real-time visibility of the orders or executed trades of the other Traders - the Desk did not have a “common blotter”.  Whilst Traders could see the orders which had been placed by all market participants on Eurex, they could not see the identity of the trader that had placed the order and would not therefore know, unless specifically told by the relevant trader, that it had been placed by one of the other Traders.

(2)        The Traders each had their own “book” (with Mr Urra having two books), ie the record of the securities that they had bought and sold at any time, and the associated profit or loss.  The Traders were responsible for managing their own risk on their books. 

(3)        The Traders would assist each other with hedging, particularly when they were busy.  Mr Lopez and Mr Sheth both agreed they would hedge Mr Urra’s trades, and this could be seen in some of the Instances.

(4)        If a Trader wanted another member of the Desk to know about a particular trade or hedge, ask them to hedge a trade for them, or give them instructions as to placing of an order (whether as to pricing or size) they would need to communicate that information orally. 

(5)        There was no formal division of responsibility for dealing with RFQs of particular sizes or from particular clients.  When RFQs were received by the Desk, in whatever form, the Trader proposing to provide a price in response to that RFQ would need to inform the others that they were responding.  Mr Heiberg was trying to create a structure where Mr Urra and Mr Hill would focus only on the big tickets, with Mr Lopez and Mr Barouti then focusing on the medium-sized tickets (with €25m being used as an example of a medium-sized ticket for this purpose).  This clear division of roles and responsibilities had not yet been introduced by the time of the Relevant Period; all three Traders were responding to RFQs of this size.  The Tribunal does accept, based on the explanations given by Mr Heiberg in his interview with the Authority, that Mr Lopez had been recruited with a view to focusing on the medium-sized tickets.

(6)        The Traders could each execute cash trades which could be allocated to another Trader’s book.    

225. The Traders were talking to each other frequently throughout the trading day.  The dispute of fact is the subject-matter of that discussion.  Whilst we need to revert to this in the Evaluation, the Tribunal make the following preliminary findings:

(1)        The physical layout of the Desk meant that conversations between the Traders could be overheard by Mr Hill and Mr Barouti.

(2)        The Traders discussed market conditions (including market news and events) and trading strategies in general terms.  Mr Hill and Mr Barouti would sometimes join in these discussions. 

(3)        Mr Hill and Mr Barouti saw and heard nothing which led them to suspect that the Traders were engaged in anything untoward.  

Eurex Letter

226. The Eurex Letter was sent to MSUSA as the member firm on the Exchange on 26 July 2016 and that letter was forwarded by MSUSA to MHI on 28 July 2016.

227. The Eurex Letter asked about two instances which occurred on 29 June 2016:

“On 29th June 2016 starting from 12.58.50 pm (CET) four bid orders of 200 and 500 contracts were entered in the order book of the FBTP SEP 16.  Some seconds later starting from 12.58.57 pm (CET) ask orders with a volume of 25 contracts in total were executed.  Immediately after that, the four bid orders of 200 and 500 contracts had been cancelled without execution.

Almost the same scenario could be observed on the same day at 15.30 pm (CET).

The two described trading scenarios could constitute a breach of §17 of the Exchange Rules of Eurex Deutschland and Eurex Zürich and §20a of the German Securities Trading Act (WpHG).

…

3.  Please explain the underlying strategy of the orders and transactions as detailed as possible.  Please describe in particular why you entered bid orders with a high volume and traded on the ask side.  Please also explain why you deleted the bid orders without execution.”

228. The two instances identified are now referred to as F174 and F176, ie they are within the Instance Pool.  These instances involved orders placed by Mr Lopez and Mr Sheth; no orders were placed by Mr Urra in these instances, although the Small Order in F176 was a hedge for a bond in Mr Urra’s book.  More particularly:

(1)        F174 involved Mr Sheth placing three buy orders of 500 lots each and Mr Lopez placing a buy order of 200 lots at the same time as Mr Lopez had a 25 lot sell order on the market. The buy orders were cancelled after the sell order traded. 

(2)        F176 involved Mr Sheth placing a series of seven buy orders of 500 lots each (not all of which overlapped) whilst having an 80 lot sell order on the market, followed by a 10 lot sell order.  Some of the buy orders were cancelled before the sell orders had traded fully, and once the sell orders had traded the remaining buy orders were cancelled.

229. Mr Urra was informed of this letter on 28 or 29 July 2016.  He attended an interview with Compliance at 8.30 am on 29 July 2016, returned with a note of his initial thoughts at midday, produced a revised note on 2 August 2016 and was interviewed again on 5 August 2016.  Mr Sheth was interviewed by Compliance on 3 and 10 August 2016.  (Mr Lopez was on holiday for the first two weeks in August 2016 and he was not interviewed by Compliance until 29 September 2016, ie after MHI had responded to the Eurex Letter.)

230.  The pattern of trading activity ceased immediately.  On 28 July 2016 there were nine instances (F225 to F233) which between them involved each of the Traders, with F233 taking place at around 4pm on 28 July 2016.  The Tribunal finds that Compliance told Mr Urra that the Desk must no longer place large orders in Futures away from the touch and that he informed Mr Lopez and Mr Sheth of this instruction.

231. MHI responded to MSUSA on 22 August 2016.  That letter, which was sent by Mr Joshi, had been reviewed by Mr Heiberg and others at MHI, as well as MHI’s legal advisors.  MSUSA then provided this response to the Exchange. 

232. MHI’s response to MSUSA included:

(1)        Futures are used for hedging of positions (or potential positions) arising from cash bond transactions executed with clients.

(2)        The desk also seeks to build “basis/curve” positions using cash bond positions and futures positions at levels which would be attractive for clients.

(3)        The “ask” orders were entered and executed to hedge transactions in cash bonds, the “bid” orders were “to establish a “basis/curve” position of some size within the firm’s market making book at an attractive level for clients.  These orders also help traders to understand the direction and depth of the market”.  The “bid” orders were placed at about the same time as the ask orders as the “desk was of the view that there may have been interest in an order of this size due to potential wider selling activity in the government bond market at that time”.  The bid orders were deleted without execution as the desk determined that the market interest at this size and price level had not materialised.

(4)        MHI’s traders have emphasised that the orders put to the market are real executable orders; if the order is filled, this is fully honoured and the transaction is settled.  MHI has in an economic sense its capital at risk.

(5)        The large multiple orders were a result of a trader who intended to amend the bid price for an existing order but instead entering new orders and failing to cancel the previous orders and as being a mistake. 

(6)        MHI’s management has instructed its traders “not to place these types of order on the Exchange until further notice”.

233. The Exchange acknowledged receipt of this letter and did not ask any further questions.

Interviews with Compliance

234. Members of MHI’s Compliance team interviewed the Traders and Mr Heiberg.  The notes of those interviews were prepared by Compliance, and were signed (unamended) by the relevant Trader.  Giving evidence, the Traders challenged the accuracy of these notes.

235. Mr Urra was interviewed by Mr Joshi, Mr Algeo and Mr Lloyd (all from Compliance) at 8.30am on 29 July 2016.  The note of that meeting records that Mr Urra had noted that the orders looked strange.  It records: 

“Mr Urra noted that the desk had concerns that futures orders placed by MHI were being front run, possibly by algorithms … MHI had raised this with Bloomberg and MSUSA.  Mr Urra stated that MHI undertook “price discovery” by putting capital at risk and that adding size to an order neutralised this front running effect.  Mr Urra was asked whether this was the reason for these trades that Eurex had highlighted and he stated that this would not have been the reason for placing these orders.” 

236. When he returned with a written response at midday, the note of the meeting records Mr Urra “stated that the large bid orders related to liquidity provision to get some size into the books into a curve position, for example if other dealers were hit by cash bond sellers at once.  Mr Urra added that the multiple large orders being placed on the book was an error - the correct procedure should have been to amend the price of the original order”.

237. The Tribunal did not have a copy of Mr Urra’s written response which he provided at that time.  Mr Urra produced a second draft on 2 August 2016.  That note, which was included in the MHI Compliance Report, says “… the ask side trades refer to hedging at market level of cash transactions and the bid side liquidity provision below the market refers to an interest to get some size into the books into a curve position in case other market participants decide to take advantage of the size of the bid, which could have happened should the cash bond sellers or futures block hedging of the time were hitting several dealers at once”.

238. Mr Urra was then interviewed by Mr Joshi, Mr Algeo and Mr Tippetts on 5 August 2016.  The note has a sub-heading “Price Discovery rationale” and refers to MHI not having the same level of information as primary dealers in the market as they do not see the flow, that the desk may contribute liquidity to the market and that they only decide to use price discovery if they feel “blind” to what is going on and need to gather more information to give them direction.  The Desk began trading in this way around March 2016.

239. Both of these notes were signed by Mr Urra on 12 October 2016, and when giving evidence Mr Urra said he had not scrutinised them at that time.

240. Mr Sheth was interviewed by Mr Algeo and Mr Lloyd on 3 August 2016.  The note of the interview says that Mr Sheth explained that the large buy orders placed beneath the current market execution price provided information on the market.  The futures market was not very liquid, and price transparency was not great as much use was made of hidden orders.  Mr Sheth said that the Desk suspected that their orders were being “front run” by other firms’ algorithms, and “Mr Sheth noted that the placing of large opposite side orders away from the current traded market level often had the effect of enabling the desk to get their smaller orders executed at the current traded market level”.  Mr Sheth explained that the large orders were used for price discovery purposes on other occasions, eg to price up a security for a client, rather than as part of filling a client order.  The entry of multiple large orders was said to be a mistake. 

241. Mr Sheth was interviewed by Mr Joshi and Mr Tippetts on 9 August 2016.  The background section of the note of that interview refers to Mr Joshi saying that he would like to understand the rationale for the transactions and to understand more clearly how the trading strategy aids price discovery.  The note records that Mr Sheth explained that the Desk did not always put on orders on the opposite side of a small order.  When a client asks for a bid/offer, the Desk does not have the complete picture of the market since they only see 10-20% of what is happening.  Large orders were often priced two or three ticks away from the market, but in BTPs that is not substantially away.  The multiple orders were a mistake - a likely reason for this might have been that he had intended to cancel or amend the trade.

242. The notes of these interviews were signed by Mr Sheth but not dated.

243. MHI responded to the Eurex Letter, and Compliance then conducted interviews with Mr Lopez and Mr Heiberg.

244. Mr Lopez was interviewed by Mr Joshi, Mr Algeo and Mr Lloyd on 29 September 2016.  The note of the interview records that Mr Lopez had stated that in April the Desk had been asked to get more business and increase flow.  It needed to offer tighter pricing in order to achieve this and, as a result, better market information was needed.  Mr Lopez said his order of 200 lots (in F174) had been for price discovery purposes, as some clients hit multiple dealers with orders at the same time.  He also noted that if the orders were filled then they would be of benefit to the positioning side of the Desk’s activities.  Mr Lopez expanded on the rationale for “price discovery”, explaining that the BTP market was less liquid than core markets and prices were more erratic.  The purpose is to understand how liquid the market was in order for the Desk to be able to offer the tight pricing required by clients.  This activity of placing orders, in part for price discovery purposes, had been common at his previous employer and it was a widespread practice in the market.  Mr Lloyd had queried the sequence for price discovery - Mr Lopez agreed that a price discovery trade would be placed first, then based on the results a firm quote would be provided to the cash client and then a hedging futures order would be placed.  The note includes a section on the placing of multiple orders, with Mr Lopez recorded as saying that this was “in part due to the configuration of the order book on his PC”. 

245. Mr Lopez signed the note of that meeting on 11 October 2016.

246. Mr Heiberg was interviewed by Mr Joshi and Mr Lloyd on 6 October 2016.  The note of that interview referred to them having previously had a number of conversations with him.  In his view on the explanations it is said he had not been aware that the Desk was engaging in “this pattern of trading prior to the Eurex letter”, that Mr Urra had indicated that “this” was a one-off and had not disclosed that there had been a repeated pattern of similar trading.  Mr Heiberg said that capturing a potential basis position was a possible explanation for these orders, but that the frequency, their size and timing meant this was “not a fully convincing explanation”.

Investigation by MHI Compliance

247. Compliance conducted a wider review of the Desk, the outcome of which was set out in the MHI Compliance Report of 13 October 2016.

248. As described in the report, in the review Compliance:

(1)        reviewed in detail the letter from the Eurex Exchange to MSUSA and the specific transactions referred to therein;

(2)        analysed in detail all futures orders placed by the Desk and transactions for five other days, selected randomly, in June and July 2016;

(3)        reviewed orders and transactions conducted between January and May 2016 - one day per month, selected randomly;

(4)        interviewed Mr Heiberg, Mr Urra, Mr Lopez and Mr Sheth;

(5)        reviewed the Desk’s organisation, strategy, performance, limits and exposure;

(6)        reviewed electronic communications: e-mails for the period January to July 2016 and Bloomberg chats for the period April to July 2016; and

(7)        reviewed telephone calls between Mr Urra, Mr Sheth and Mr Lopez for the period April to July 2016;

249. The interviews with the Traders were those which have been referred to above.  Compliance had requested details of the futures orders placed by the Desk shortly after receiving the Eurex Letter, and received this information in early August 2016.

250. The summary of its analysis and findings is then as follows:

(1)        The Desk “appears to have operated within its mandate and end of day limits”.

(2)        The placing of multiple large orders when a smaller order on the other side was being executed “appears to have been a regular practice and was not confined to the two instances highlighted by Eurex”. The Desk has stated that in many cases it had not intended to place multiple large orders, rather the trader(s) were seeking to update the price of an existing order and in doing so did not cancel previous orders (which they felt were unlikely to be filled because they were away from the prevailing price / not attracting interest). This resulted in large cumulative outstanding orders.  “If this is the case, it would appear to us to be poor practice in terms of order entry and management.”

(3)        The Head of the Desk had informed them that the practice of placing larger orders had been used since around April 2016 when the Desk was asked to serve a wider client base.

(4)        The futures orders placed by the Desk were fully transparent and available to the market for execution. The orders are visible for a sufficient period of time to enable market participants to decide whether or not to act on them.

(5)        “Our review also indicates that large orders are not solely placed at times when the traders are seeking to execute smaller orders on the other side. On the days we sampled, 33% of large orders were not placed whilst a smaller order was in the process of being executed on the opposite side. This appears to indicate that the sole purpose of placing the large orders is not necessarily to seek to influence the execution of the smaller lot on the other side.”

(6)        “We have noted that, when a large order has been placed in the Italian Government Bond futures market and a smaller order is being executed on the other side of the Eurex order book, the large order is usually cancelled a short time after the smaller order has been executed - literally within seconds.  It could be said that this indicates that there was never any intention to execute such orders and that their sole purpose was to enable the Desk to execute the smaller orders at a better price.” 

(7)        Compliance had undertaken surveillance of the Traders’ Bloomberg messages and chat (from January to July 2016) and e-mails (from April to July 2016).  This was in addition to the ongoing electronic monitoring of such traffic.  This did not reveal any evidence of manipulative or abusive behaviour.

(8)        The Traders had explained that the purpose of the large orders is to seek to build “basis/curve” positions at relatively cheap levels if possible.  Where such orders are not filled they provide the Desk with useful information as to the depth and direction of the market (the Desk has referred to this as “price discovery”).  The Desk thinks clients could be dealing through multiple market makers.

(9)        At the time at which these order placing and cancellation practices were taking place, neither the Front Office nor Compliance were able to monitor such activity, as orders were routed through MSUSA and MHI maintained records of trades executed, but not orders. Orders were being monitored by MSUSA.  Arrangements are currently underway to ensure that Eurex orders are captured within Trading Hub.

251. It summarised its conclusions as follows:

(1)        The orders could be regarded as “genuine” or “legitimate” in the sense they would have been honoured if they had been filled. MHI had put its capital at risk.

(2)        The practice of placing orders which resulted in multiple outstanding orders when this was not the intention indicates a lack of sufficient due care and poor practice in terms of order management.  Such conduct is not in accordance with the standards accepted by MHI.

(3)        The pattern of activity could give the perception that the purpose of the large orders was to facilitate the execution of the smaller order on the other side of the order book. We have received explanations from the Traders in relation to the Desk’s activities from which it could be concluded that these were legitimate.  However, there is sufficient doubt as to the credibility of these explanations to leave open the possibility that the trading patterns could be regarded as abusive.  We have not found conclusive evidence in this regard.

(4)        The trading practice described above does not appear to have resulted in any abnormal impact on the market.  It may be that the Desk’s dealing pattern did not result in “creating a false or misleading impression” as to the supply or demand for the relevant futures contracts.  There does not appear to have been any significant distortion in the way in which the market was operating.

(5)        Even if the conduct does not result in a clear breach of the market abuse regulations, such conduct was not in accordance with the standards expected by MHI and put the Desk and the firm at risk of being accused of not complying with high standards of market conduct.

(6)        There is a case for disciplinary proceedings to be taken against Mr Urra, Mr Lopez and Mr Sheth for not acting with sufficient due skill, care and diligence, and possibly in relation to adherence to high standards of market conduct, and they referred these issues to the Head of HR in order that they may institute a disciplinary investigation.

252. The report had considered the hypothetical impact of outright purchases of 500 lots and 2,000 lots by reference to the Desk’s risk limits:  

(1)        An outright transaction of 500 lots of Futures equated to approximately £1.3m of VaR; 2,000 lots equated to approximately £3.3m. 

(2)         2,000 lots would equate to $186,000 of PV01 Hedge risk.  The table in the report sets out that the exposure on 500 lots would be $2,000.  On the basis of all of the evidence, the Tribunal determines that this is a typographical error, and should be in the region of $40,000-50,000 (but in any event within the Desk limit).

253. The report states that the Desk’s “intended trade” of 500 lots, even if it had not been hedged by the end of day, would not have breached the Desk’s limits.  However, had the actual order of 2,000 lots been filled, then the Desk would have breached its limits, if it had not been hedged by the end of the day.  Risk limits are formally monitored based on end of day exposures.  This then led to the conclusion that the large Futures orders placed on 29 June 2016 “to the extent that these were placed with a view to establishing a hedge against expected near term customer trades or to establish a basis position for future sale to clients, would fall into the category of “anticipatory hedging” and are permitted by the Desk’s mandate.”

254. The report sets out the result of its communications review (having first observed that as the Traders sit on the same desk “there would thus appear to be little need for these traders to communicate electronically or by telephone in respect of the placing of these orders”).  The results of its review were:

(1)        Review of Bloomberg Messages and Chats - Searches were performed for all Bloomberg messages and chats exchanged between Mr Urra, Mr Lopez and Mr Sheth between 1 January and 31 July 2016.  No Bloomberg messages were sent between Mr Urra, Mr Lopez and Mr Sheth during this period.  All chats between Mr Urra, Mr Lopez and Mr Sheth during this period were reviewed and no issues were identified.

(2)        Review of Email Communications - Searches were performed for emails sent by Mr Urra, Mr Lopez and Mr Sheth where one or more of the others were recipients between April and July 2016.  No issues were identified.

(3)        Review of Telephone Records - Searches were performed for all telephone calls made between Mr Urra, Mr Lopez and Mr Sheth in all combinations between 1 April and 31 July 2016.  Only one call was found, a 50 second conversation between Mr Urra and Mr Lopez on 21 June at 11:26.  The call was in Spanish and did not appear to be related to the trading of Eurex contracts.

255. The report then sets out its overall assessment on compliance with market abuse regulatory requirements and includes:

“Given that the orders were in keeping with the desk’s market making remit and risk limits, that the orders were available to the market for long enough to be executed and that there is no direct evidence to the contrary, we conclude that there was an intention to execute the orders as placed.

…

We have obtained no direct evidence that the Desk entered large orders in order to create abnormal or artificial prices on the Exchange.

Whilst the Desk may have intended to execute the large orders if filled…the possibility remains that the primary motivation in placing such orders was to facilitate the execution of the Desk’s concurrent smaller orders at particular prices…

Our overall conclusion is thus indeterminate as to whether the Desk was attempting to create artificial prices…”

256. MHI sent a copy of this report to each of the Traders and they were each told that a further investigation was to be conducted to consider disciplinary action.  The report includes a copy of the Eurex Letter, MHI’s response to MSUSA and the notes of the interviews with each of the Traders.

MHI disciplinary process

257. MHI then took disciplinary action against each of the Traders which involved further interviews with each of the Traders.  These interviews were conducted by HR, not Compliance, and the Tribunal had notes of these meetings.   

258. Mr Sheth was interviewed on 14 October 2016.  He explained that the large order was a price discovery process, to test the market.

259. Mr Urra was interviewed on 18 October 2016.  Mr Urra is recorded as saying that the large orders (which he said were not actually large) were placed to give the Desk a better understanding of the market, and that clients could be selling bonds because they had seen news that the Desk had not or the client could be splitting larger flow into smaller clips. 

260. Mr Lopez was interviewed on 18 October 2016.  Mr Lopez referred to the MHI Compliance Report, saying that in his specific case his trading is limited to sizes not bigger than 200 lots and overall he did not see any remote link to anything in the report.  Where there are two orders, the one of 200 lots is a delta positioning trade, whereas the other is linked to a cash trade.  The purpose of some of the large orders look to have been price discovery, others could have been delta or basis positioning.    

261. Mr Lopez sent an email to HR the following day.  That email contained a summary of his background and trading style, with more details on F174, describing the 200 lots as corresponding to duration trading, with the idea that the market was going back to a particular level and he could take on some delta risk, well within his limits.  On F174, he said he “can provide some details to the best of what I can remember”, said he had had a look at market levels, Futures and bond activity and said “the thought process for my orders originates from a previous enquiry where a client sold 14mm of btp 2029 bonds”, which triggered a sell off towards a technical support level in the ten-year Futures.

262. The outcome of the disciplinary process was recorded in letters from MHI dated 10 November 2016 to each of the Traders: 

(1)        Mr Urra was given a final written warning on the basis that he had not acted with sufficient due skill care and diligence in his role as manager by failing to take reasonable steps to ensure that the Desk was controlled effectively.  This included a finding that there was inadequate monitoring of transactions and business practices.  This related to the failure to stop Mr Sheth engaging in the “poor practice” of entering new trades rather than amending orders which “attracted the attention of Eurex and an enquiry from them”.

(2)        MHI did not find that any of the allegations against Mr Lopez were well-founded.  No disciplinary action was taken against him.

(3)        Mr Sheth was given a first written warning on the basis that he had acted without sufficient due skill, care and diligence by engaging in what he now recognised to be a poor practice in terms of failing to amend orders but instead entering new trades and cancelling existing trades.

Interviews by the Authority

263. The Authority met with Mr Joshi and Michiel de Jong, CEO of MHI, on 2 November 2016.  The Authority’s note of that meeting includes the summary that MHI found the Traders’ activity unacceptable; and MHI “could not find unequivocal evidence that the traders’ intention was to manipulate the market” and the Traders themselves apparently remain convinced that they were pursuing a reasonable strategy.

264. As part of its investigation, the Authority interviewed Mr Heiberg on 17 July 2017 and Mr Joshi on 3 October 2017.

265. The Authority then interviewed the Traders under caution in 2018.  Mr Lopez was interviewed on 18 and 19 January 2018, Mr Sheth on 15 March 2018 and Mr Urra on 16 April 2018.  Mr Sheth and Mr Urra each exercised their right to silence or to give “no comment” interviews, which is why their interviews were each able to be conducted over a single day.

266. The Authority interviewed Mr Hill and Mr Barouti in June 2018.

Traders’ explanations of rationale for the Large Orders

267. There is no contemporaneous documentary evidence of the design or operation of the Trading Strategies that are now relied upon by the Traders as being their explanations for placing the Large Orders (and the other large orders during the Relevant Period). 

268. We have summarised above the notes of the meetings with Compliance and HR (and revert to those in the Evaluation).  The descriptions in this section of the decision of the evidence given by the Traders are not findings by the Tribunal that these strategies were being pursued by the Traders. 

269. Mr Urra’s evidence was that he was pursuing what we have referred to as an Information Discovery Strategy.  Mr Sheth’s evidence was that he was also pursuing this strategy, which he said had been shown to him by Mr Urra and which he had been given permission to try.  In closing submissions, Mr George and Mr Bailin both referred to these strategies as being “similar” to each other.  Mr Lopez’s evidence was that he was pursuing the Anticipatory Hedging Strategy.

Information Discovery Strategy - Mr Urra

270. Mr Urra described his trading style as being multi-dimensional, being informed by a range of information and strategies.  Mr Urra frequently used relative value trading strategies, seeking to exploit pricing discrepancies or misalignments between related financial instruments.  His perspective was that a key advantage of relative value trading was that such trades could be structured to offset directional market risk.  Instead of speculating on the overall direction of the market, the primary risk exposure associated with relative value trading is the relative price change between instruments. 

271. This Information Discovery Strategy was devised by Mr Urra and was based on his belief that the Desk was vulnerable to an information disadvantage and that clients may be splitting orders with MHI and other market makers.  Mr Urra’s hypothesis was that another market maker, referred to by Mr George as a “Posited Trader”, who might be dealing with a possibly much larger part of the same client order, might now be exposed to significant risk and be looking to hedge their position swiftly and might be prepared to pay a premium to do so.  A large order, away from the touch, might be attractive to such a Posited Trader and be profitable for the Desk.

272. Mr Urra’s evidence was that this strategy was consistent with the objectives of the Desk, low risk and in line with the Mandate - the Desk was allowed and encouraged to take positions as long as the overall activity was within the Desk limits.

273. The benefits for the Desk were:

(1)        If Mr Urra was right in the assessment which led him to place a Large Order, he would have correctly identified the existence of a Posited Trader who was looking to transact promptly and at volume, and who was prepared to pay a premium to do so.  A successful fill could then be used by Mr Urra in ways which would be profitable and advantageous to the Desk, eg by holding the position and using it to hedge future client business more cheaply, allowing more aggressive pricing and thereby winning more client bids for the Desk, building a relative value position at a favourable level, or just simply offloading the duration at a profit.  Mr Urra considered it likely that the market would revert to its previous market levels once the order had traded, creating opportunities for profitable trades. 

(2)        Whether or not the Large Order did (as primarily intended) result in a successful trade, there was a “price discovery” or informational benefit.  If the Large Order was filled, it suggested that Mr Urra’s theory about a client splitting a larger flow was correct. This new information could guide future pricing by the Desk and alert Mr Urra to the need to exercise greater caution when dealing with that client in future.  If (contrary to Mr Urra’s primary intention) the Large Order was not filled, it revealed that the client was likely disclosing their full order.  With a better understanding of the market and MHI’s clients, Mr Urra could adapt his trading strategies and increase the ability of the Desk to win more client orders by offering favourable pricing.

274. The Large Orders were not large for the market, they were fully transparent and available for a sufficient period of time to enable market participants to decide whether or not to transact.  If the orders had been filled, they would have been fully honoured and the transactions settled.

275. A consequence of the Information Discovery Strategy was that large orders would be likely to frequently (but not always) overlap with small orders trading in the opposite direction; both orders could have been triggered by the same client activity or market event, despite serving entirely different purposes.  A cash enquiry from a client that traded away (ie MHI was asked for a price but the client traded elsewhere) may still have prompted him to use this strategy and place a large order, but in that situation the Desk would not be hedging a client trade (ie it would not be placing a small order on the opposite side).

276. Mr Urra’s evidence was that he had had some successful fills in the past which made him think the strategy was worth implementing.  He did recall that in late May or early June he had been filled all at once in what he described as a medium-sized order behind the touch “which, in addition to representing a good deal with a potential for profit, had allowed [him] to understand better market positioning or some client activity at the time”.  (The more specific evidence in relation to past success was put forward by Mr Sheth - Mr Sheth had a specific recollection that Mr Urra had a successful trade in May 2016 and had told him about this.  Mr Sheth was subsequently able to identify (from the records of trading activity) that this had been on 12 May 2016 and involved an order of 450 lots.)  Mr Urra explained that he became less enthusiastic about this strategy, and decreased the number of times he used it as the Relevant Period progressed.

277. Mr Urra could no longer recall the specific trigger for placing the Large Orders which are in the Instance Pool, and referred to the information that was no longer available, including Voice RFQs, Bloomberg/Chat RFQs and market information, which would have informed his trading decisions at the time.  Mr Urra had not, in his witness statement or giving evidence, attempted to reconstruct the rationale for his trading (although he had given some explanation in relation to F47 and F150).   

Information Discovery Strategy - Mr Sheth

278. Mr Sheth’s evidence was that he was also pursuing the Information Discovery Strategy, explaining that this had been shown to him by Mr Urra in May 2016, he understood it had been successful and Mr Urra had given him permission to use this himself.  Any Futures orders in round numbers in excess of 200 lots would have been placed in pursuit of this strategy. 

279. Mr Sheth explained his understanding that the purpose of the strategy was to obtain information as to whether there was hidden liquidity in the cash market and create profitable positions as market participants looking to hedge cash trades would be willing to pay a premium to access liquidity:

(1)        Informational purpose - This was to conduct “price discovery”, and he understood this to mean that it would help the Desk discover the actual liquidity of the market and to enable the Desk to better price its trades, saying that it was to address the information imbalance suffered by MHI.  The Large Orders were placed to test the theory that some counterparties would split up large cash orders between multiple market makers and therefore withhold their full position from MHI.

(2)        Creation of profitable positions - If clients were splitting their orders, the increase in cash market activity would in turn increase the activity in the Futures market and potentially affect Futures prices.  It was assumed that market participants looking to hedge the cash trades would be willing to accept a less favourable price to access liquidity quickly.  Given that the resulting movement in the market was generated in response to activity in the cash market, the theory was that the Futures market, all things being equal, was likely to bounce back, allowing the Desk to unwind the Futures position at a gain.  In cross-examination, Mr Sheth said that he would have already lined up orders in a cash bond which the Desk considered was expensive in relative value terms.  He would sell those, create a basis position where he is short an expensive bond versus Futures and could then market that position to relative value hedge funds.  He would not have held the position for any longer than necessary.

280. Mr Sheth became frustrated that this strategy was not working (none of his Large Orders traded) and, around the time of the Brexit referendum, he discussed this with Mr Urra who suggested to Mr Sheth that he move his large orders closer to the touch.  Mr Sheth did try this, and it was his amending the price which led to the Multiple Large Orders, but was not successful and used it less.

281. As to whether the Information Discovery Strategy complied with the Mandate, Mr Sheth’s evidence was that he didn’t believe he had seen the Mandate; for him, his manager had shown him this strategy and so it must be within the Mandate.

Anticipatory Hedging Strategy - Mr Lopez

282. Mr Lopez’s evidence was that he placed orders of 200 lots or more as anticipatory hedges, seeking to look ahead to likely client demand and position his trading book and inventory of bonds to service that demand.  He had a particular focus on seeking to improve MHI’s success rate for RFQs for cash bonds in the range of €20-30m, and placed large orders at times when he considered it was highly likely that client interest would materialise and when there was a reasonable chance of such orders executing at a price that he considered represented good value.  Pre-positioning in this way was, he considered, the way to enable him to offer competitive prices to clients. 

283. Mr Lopez’s evidence was that he was specifically hired by Mr Heiberg to focus on RFQs of €20-30m, although he accepted that Mr Urra would also respond to RFQs of this size.  Mr Lopez was seeking to identify and predict patterns of client flow - in particular, he said he identified that, most days in the Relevant Period, Banca d’Italia approached MHI with an RFQ for MHI to sell €25m nominal of bonds within a defined maturity range (often June 2026), and another client said to be predictable was Banco Popular.  Mr Lopez produced a list of the 93 Electronic RFQs received by the Desk during the Relevant Period which would have required hedging with 200 lots or more of Futures (the “93 RFQs”). 

284. A significant amount of his time was spent watching market movements and analysing where he considered resistance levels (ie upper price levels where he expected the market to turn down) and support levels (ie lower price levels where he expected the market to turn up) to be.  He would keep in mind resistance and support levels from previous days when considering where he expected the market to go on any one day, as well as recent and upcoming news events, relevant RFQs and relevant auction dates.  This would feed into his view, which he would constantly update, as to where he considered to be a good level to buy and sell. 

285. Between 1 June and 10 June 2016, his VaR was very high, showing he had a higher risk appetite in this period, and during this time he placed anticipatory hedge orders of 300 and 400 lots (whereas, later in June, as the Brexit referendum (held on 23 June 2016) came nearer, and in July, 200 lot orders were more common as his risk appetite was lower).  He considered orders of 200 lots to be below standard traded sizes in the BTP market at the time, which he estimated to be between €25-50m of ten-year BTPs (ie 250 to 500 lots).    

286. Mr Lopez did not use the iceberg function for these orders.  This was principally because he wanted to trade the full number of lots, and showing the full number of lots gave rise to the possibility that someone would hit the order and take it all.  If he placed the order as an iceberg, every time a small and partial fill was completed, he would go to the back of the queue.  This would make it less likely that the whole of the order would be filled.  In contrast, he would iceberg his small hedging orders as he was happy for them to execute slowly (providing an opportunity for a client cash trade to come the other way), and this also allowed him the opportunity to amend the price. 

287. He typically placed his anticipatory hedge orders several ticks away from the touch; he wished to pre-position at an attractive price, which he expected would enable him to meet client orders on an attractive and profitable basis.

288. On Mr Lopez’s strategy, the Large Orders and the Small Orders within an Instance were unconnected.  He had placed a total of 1,139 small orders during the Relevant Period. 

Trading Activity of the Traders in the Relevant Period

289. The Authority was provided with data from MHI and Eurex recording all orders in Futures placed by market participants during the Relevant Period; the Authority identified the Instance Pool from this data.

290. The Criteria applied by the Authority are set out in its Amended Statements of Case and involve: 

(1)        At least one order was placed by the Traders of at least 250 lots for Mr Urra and Mr Sheth, and at least 200 lots for Mr Lopez.  Where a Multi Trader Instance involved Mr Urra or Mr Sheth, then their orders of more than 200 lots were included.

(2)        This order was not placed as an iceberg.

(3)        At least one small order was placed by the Traders of under 200 lots on the opposite side of the order book.  Two of the Small Orders in the Instance Pool are 200 lots or more.  They were iceberged, and the Authority relies on them appearing as 22 lots and 25 lots to the market.

(4)        The period in which the large order and small order were live on the order book overlapped.

(5)        The large orders were cancelled shortly after the small orders fully or mostly traded.

291. The Authority identified 233 Instances involving a total of 341 Large Orders.  Of those Large Orders, 338 were cancelled without having traded at all, and three had traded in part (all three of which had been placed by Mr Urra).

292. The Authority identified a total of 40 Specified Instances from the Instance Pool.  Mr Creaturo and Mr Kasapis both considered that the Specified Instances were broadly representative of the Instance Pool as a whole.  Mr Urra drew attention to the difference in the average cancellation times of the Large Orders after the Small Order filled.  The average cancellation time of the Large Orders for the 18 Specified Instances which involved his trading was 2.5 seconds after the Small Order filled, whereas the average cancellation time for the Instance Pool as a whole was 5.3 seconds.  Mr Kasapis’s opinion was that this is a “significant difference within the context of the fast-moving Eurex market”.

293. The Tribunal accepts the expert evidence that the Specified Instances are broadly representative of the Instance Pool as a whole, but also acknowledges that they do not purport to be a statistical sample of such pool.  The Tribunal accepts that there is a difference in the average cancellation times as identified by Mr Urra.

294. We summarise here some of the features of this trading activity, not only how the Criteria relied upon by the Authority appear but also outline those aspects of the trading relied upon by the Traders.

Illustration of application of Criteria to Trading Activity in Instances

295. The trading activity relied upon by the Authority, ie the placing, amendment, filling and/or cancellation of orders, is agreed.  The Authority’s case is made by reference to inferences it submits that the Tribunal should draw from the trading activity in the Instances and from the trading activity of other market participants. 

296. We illustrate the Authority’s case by reference to a Specified Instance for each Trader.

Mr Urra - F7 at 15.31.06.983 on 7 June 2016

297. The trading activity for this instance is set out in Appendix 2 hereto.  The length of this Instance is 13 seconds.

298. As can be seen from the table of trading activity, Mr Urra had placed a Small Order at 15.31.06.983, selling 40 lots iceberged to nine, which was filling.  After 18 lots had filled, Mr Urra placed the Large Order, a buy of 444 lots which was live for 1.899 seconds in total, and was cancelled 1.339 seconds after the Small Order filled.

299. Mr Urra’s evidence was that he had placed the Large Order in pursuance of the Information Discovery Strategy.

300. This was one of the Instances for which Mr Kasapis had identified a cash bond trade in proximity to the Instance, which could have prompted Mr Urra to speculate that a Large Order might be tempting to a Posited Trader.  The cash bond trade was the purchase by MHI of a €10m bond from Industrial & Commercial Bank of China Financial Services LLC (“ICBC”), for which the Small Order was a (partial) hedge.  The cash bond trade is timed at 15.36.05, ie after the Instance, but it was a Voice RFQ and we accept that the time for which it was entered was likely to have been after the price had been agreed by the Desk.  Mr Urra agreed that this was a plausible match and that the Large Order could have been triggered by the trade with ICBC. 

301. In terms of the activity itself, the Authority relied on:

(1)        Mr Urra was trying to sell his Small Order at or very close to the daily high: the market had been in a rally but had just fallen back at the time of the Instance.

(2)        18 lots of the Small Order had filled within 0.5 seconds of the Small Order being placed.  However, trading then paused for ten seconds.  It was only after the Large Order was placed that the Small Order continued to fill - trading resumed within 0.2 seconds of the Large Order being placed.

(3)        The Large Order was cancelled 1.339 seconds after the Small Order filled.  Whilst Mr Urra’s evidence was that he cancelled the Large Order as he was discouraged by the price action and his best guess was that he had concluded that the order was not going to transact, the Authority relied on the Large Order being live for less than two seconds, submitting this was not enough time to think and reach this conclusion.

Mr Lopez - F56 at 17.02.08.899 on 15 June 2016

302. The trading activity was as follows:

Trader

Side

Quantity

Eurex Time

Transaction Quantity

Price

Status

Iceberg Display

Lopez

S

5

17.02.08.899

5

140.62

Sent-ack

3

Lopez

B

200

17.02.21.324

200

140.59

Sent-ack

0

Lopez

S

5

17.02.21.474

3

140.62

PtFil

3

Lopez

S

5

17.02.21.522

2

140.62

Sent-ack

3

Lopez

S

5

17.02.21.675

2

140.62

Filled

3

Lopez

B

200

17.02.24.591

200

140.59

Cancel

0

 

303. Mr Lopez’s evidence was that the Small Order to sell five lots was placed for curve management purposes and the Large Order was placed as an anticipatory hedge.  In his witness statement he referred to movements earlier that day in the Bund futures market, and that the Bund/BTP Futures spread was widening which would have indicated to him market risk aversion in BTPs, and he would have expected client selling interest.  But from 15.20 the Futures market started to rally, which, together with the RFQs received, would have prompted him to think it could be a good time to position himself long in anticipation of client buying.  Mr Lopez referred to two RFQs received by the Desk (at 15.42 and 15.46) for client buys of €25m June 2021 BTPs, both from Banca Popolare di Milano (“Banca Popolare”).  Mr Lopez’s evidence was that he cancelled the Large Order as he likely thought the pull back to 140.59 was not going to happen; and he may also have seen the Bund futures rallying and anticipated that there may not be further client buying interest.

304. In terms of the trading activity, the Authority relied on the length of time for which the Large Order was live, namely 3.267 seconds, and submitted that if Mr Lopez had wanted to take a directional position by buying for 140.59 he could have done so by keeping the Large Order on for five minutes.

Mr Sheth - F55 at 16.55.33.255 on 15 June 2016

305. The Instance lasts ten seconds and the trading activity is as follows: 

Trader

Side

Quantity

Eurex Time

Transaction Quantity

Price

Status

Iceberg Display

Sheth

B

5

16.55.33.255

5

140.63

Sent-ack

0

Sheth

B

10

16.55.38.515

10

140.59

Cancel

0

Sheth

S

500

16.55.41.886

500

140.66

Sent-ack

0

Sheth

B

5

16.55.42.150

5

140.63

Filled

0

Sheth

S

500

16.55.43.532

500

140.66

Cancel

0

 

306. Mr Sheth’s evidence was that he could not identify anything which would have prompted a sell-side Large Order.  He did identify an RFQ from Banca Popolare at 16.54 for a €2.5m buy-side order on a bond (which had been priced by Mr Lopez but traded away).  If that had been the prompt, Mr Sheth said he should then have placed a large order to buy 500 lots to pursue the Information Discovery Strategy, and suggested that the placing of an order to sell may have been an error on his part and would explain why he took the decision to cancel the Large Order so quickly. 

307. In cross-examination, Mr Shivji challenged Mr Sheth’s evidence that he did not discuss with Mr Lopez (who had priced the RFQ) his decision to place a Large Order against the direction of that client flow.  Mr Shivji also challenged Mr Sheth’s decision to deploy the Information Discovery Strategy where the RFQ had traded away.  In terms of the trading activity, the Authority relied on the Small Order of five lots having been live for eight seconds without filling, the Large Order was placed, the Small Order traded and the Large Order was cancelled 1.382 seconds after the Small Order filled. 

Dates of Instances

308. The Tribunal has set out at [17] the number of Instances relied upon against each Trader, which total 128 for Mr Urra, 125 for Mr Lopez and 104 for Mr Sheth.  Whilst the Relevant Period is 1 June to 29 July 2016, the Instances relied upon by the Authority did not start at the beginning of the period and they were not evenly spread throughout this period. 

309. The first Instance, F1, was on 7 June 2016.  There were seven Instances that day, all of which involved Mr Urra, and some of which involved Mr Lopez (who placed a Small Order in three of these Instances and a Large Order of 300 lots in one, F6, an Instance in which Mr Urra had also placed a concurrent Large Order of 444 lots). 

310. The Instances on 8 June 2016 (F8 to F15) only involved Mr Urra (where his Large Orders ranged in size from 333 to 475 lots), but thereafter there was no discernible pattern as to the involvement of each of the Traders. 

311. The second Instance involving concurrent Large Orders was F31, a Specified Instance that is a Multi Trader Instance, and involved a Large Order placed by Mr Lopez of 400 lots and a Large Order placed by Mr Urra of 499 lots.

312. The first Instance involving Mr Sheth’s trading activity was F36 on 13 June 2016, in which he placed both the Small Order and the Large Order.  After this date, his trading activity appeared in both Single Trader Instances and Multi Trader Instances.

313. The last Instances occurred on 28 July 2016, F225 to F233.  All three Traders were involved in various of these Instances.

314. 179 of the Instances occurred in June 2016 and the remaining 54 in July 2016. 

Number and size of Large Orders placed by the Traders in the Instance Pool

315. The Authority has characterised whether an order is a Large Order differently in respect of the three Traders.  For Mr Lopez, the Authority has identified his orders of 200 lots or more, rather than 250 lots or more as used for Mr Urra and Mr Sheth (save where they were involved in a Multi Trader Instance, in which case the lower threshold of 200 lots is applied to them as well).

316. Adopting this approach, the Authority relied on the following numbers of Large Orders being placed by each Trader:

(1)        Mr Urra - 135;

(2)        Mr Lopez - 55; and

(3)        Mr Sheth - 151.

317. Mr Urra’s Large Orders were usually in the range of 400 to 499 lots, typically 400, 444 or 499 lots.  He did occasionally place a Large Order of a smaller size, namely 250 or 333 lots.  The Tribunal only identified one Instance where Mr Urra had placed a Large Order of 500 lots and that was in F30, although he did occasionally place concurrent Large Orders in an Instance where the total size was then more than 500 lots, F82 where he placed two Large Orders, each of 490 lots (and which were concurrent with a Large Order placed by Mr Sheth of 500 lots).

318. Most of Mr Lopez’s Large Orders were of 200 lots.  Where he placed a Large Order in a larger size it was generally of 300 or 400 lots.  Whilst some of Mr Lopez’s Large Orders were live concurrently with Large Orders which were placed by Mr Urra or Mr Sheth, the Tribunal has not identified any Instance where Mr Lopez had placed concurrent Large Orders.

319. Mr Sheth’s Large Orders were usually of 500 lots, although he did occasionally place Large Orders of 250, 300 or 400 lots.  The 151 Large Orders relied upon against Mr Sheth include all of his Large Orders within the Instance Pool, ie including those Multiple Large Orders which he submits were a mistake.

Small Order already trading

320. In 22% of the Instances, and 20% of the Specified Instances, the Small Order had already started to fill before the Large Order was placed. 

321. This could be seen in, eg, the following Specified Instances: F7 (Mr Urra), F64 (Multi Trader Instance), F84 (Multi Trader Instance), F150 (Mr Urra), F194 (Mr Urra) and F201 (Mr Sheth).

322. The Small Order had not started to trade in any of the Single Trader Instances which were Specified Instances for Mr Lopez.  However, Mr Lopez’s Small Order had started to trade in some of the Multi Trader Instances, including in F84 which was a Specified Instance.

Amendment of price of Large Order after the Small Order filled

323. In some of the Instances, the price of the Large Order was amended after the Small Order filled.  Mr Kasapis’s report included what was said to be a non-exhaustive list of these.  He identified six Instances in which this occurred: F32 (a Specified Instance for Mr Urra), F86, F90, F161, F170 and F171. 

324. In his submissions Mr Jaffey referred to there being 12 Instances where the price of the Large Order was amended after the Small Order filled.  However, the Tribunal has not been taken to any evidence of these additional six Instances and we have not been able to identify them.  We have proceeded on the basis that there were the six Instances identified by Mr Kasapis. 

Small orders which overlapped with (and on same side as) Large Orders

325. There are some Instances where there is another small order, placed by one of the Traders, on the same side of the order book as the Large Order (the “Overlapping Small Orders”). 

326. A non-exhaustive list of Instances which involved an Overlapping Small Order had been provided to Mr Kasapis, namely: F27, F31, F32, F40, F48, F53, F54, F83, F106, F109, F114, F166, F168, F169, F170, F180, F181, F203, F204 and F222. 

327. Mr Kasapis’s report included the order numbers for these Overlapping Small Orders, and he confirmed that he had checked that the order number appeared in the full list of trading data and that it was live whilst the Small Order identified by the Authority was live on the opposite side of the book (identifying whether the Overlapping Small Order was placed or cancelled within the time of the Instance).  Mr Kasapis did not include any additional information, eg the Trader who placed the order, time at which it had been placed, time for which it was live, whether it overlapped (in time) with the Large Order, size of the order, the distance from the touch or whether the order ultimately filled.

Trading Activity of the Traders outside the Instance Pool

328. The Traders’ position was that the trading activity in the Instance Pool was not representative of their trading activity throughout the Relevant Period.  They relied in particular on the large orders they placed which are outside the Instance Pool and the number of small orders that they placed during the Relevant Period.

Non-Instance large orders and Lone Large Orders

329. During the Relevant Period the Traders placed uniceberged orders of 200 lots or more that are not within the Instance Pool (the “Non-Instance large orders”).  There are a total of 124 Non-Instance large orders and they comprise two different categories:

(1)        90 had a small order placed by one of the Traders on the opposite side of the book whilst they were live (but did not meet all of the Criteria which is why they are not in the Instance Pool); and

(2)        34 did not have a small order placed by one of the Traders on the opposite side of the book whilst they were live - these are the “Lone Large Orders”.

330. On the basis of Mr Creaturo’s evidence, the Tribunal finds that the numbers of such orders are as follows:

 

Number of Large Orders in Instance Pool

Number of Non-Instance large orders

Number of Lone Large Orders

Mr Urra

135

40

15

Mr Lopez

55

33

5

Mr Sheth

151

51

14

Totals

341

124

34

 

331. Three of these Non-Instance large orders traded - two in part and one in full, and all three of these had been placed by Mr Urra.  The remaining 121 were cancelled without having traded.

332. The Lone Large Orders were not iceberged, were placed in round numbers (with Mr Urra’s generally being between 400 to 499 lots, most of Mr Lopez’s being for 200 lots and all but one of Mr Sheth’s being for 500 lots), and were live for various lengths of time (including 2, 3, 17, 24, 34 and 43 seconds). 

333. The Tribunal finds that there is nothing to distinguish the Lone Large Orders from the Large Orders in the Instance Pool, save for the fact that there was no small order which had been placed by one of the Traders on the opposite side of the order book. 

Number of small orders placed

334. The Traders each placed a large number of small orders during the Relevant Period which did not overlap with a large order, whether a Large Order within the Instance Pool or a Non-Instance large order.   

335. Based on the evidence of Mr Kasapis, there were 1,458 small orders which did not overlap with any of the Large Orders (ie those within the Instance Pool), and 1,375 which did not overlap with either a Large Order or a Non-Instance large order. 

Trading Activity of other participants in the market

336. The SMARTS data was relied upon by the Authority in its Amended Statements of Case as supporting the inference that the Traders’ conduct was abusive.  The Tribunal makes the following findings of fact on the basis of the SMARTS data.

337. During the Relevant Period, the average daily number of active participants in the Eurex market, being participants who entered orders (including MSUSA, the market code for which was MIZNY), was 110.

338. The comparison of MIZNY’s activity with that of all other market participants during the Relevant Period included:

(1)        Number of orders - 2,872 large orders (defined as orders of 200 lots or more) were placed, accounting for 0.02% of the total number of Futures orders.  MIZNY entered the most large orders, 466, being 16.23% of all large orders by number and 23.24% of the total volume of large orders placed.  There were 2,406 large orders placed by other market participants.

(2)        Execution rates:

(a)        MIZNY executed fewer of the large orders it placed than other market participants, both by number of orders and volume.  MIZNY traded 0.44% of the contracts it entered in large orders.  The other market participant who submitted the most large orders after MIZNY executed 4.56% of its large orders.  Viewed as a whole, the other market participants traded 20.7% of the contracts they entered in large orders.  MIZNY entered 466 large orders for a total of 187,996 contracts, 1.5% of which partially or fully traded.  Other market participants entered a total of 2,406 large orders for a total of 620,977 contracts, 72.28% of which partially or fully traded and 27.72% of which did not trade at all. 

(b)        MIZNY executed more small orders than large orders, having executed 66.98% of its small orders (by volume) compared with 0.44% of its large orders.  MIZNY executed more small orders than other market participants, who executed 16.58% of the 40,799,356 contracts entered.

(3)        Pricing - MIZNY entered proportionately more large orders further away from the touch than other market participants.  1.93% of MIZNY’s large orders were placed at the Best Bid or Best Offer price, or at improved prices.  The remaining 98.07% were placed between two and nine ticks away from the best price.  80.34% of large orders placed by other market participants were at the best or improved prices.

339. The data by individual market participant included the following:

 

Market Participant

Number of Large Orders

Volume of Large Orders Entered

Volume of Large Orders Traded

Execution Rate by Volume

1

MIZNY

466

187,996

828

0.44%

2

CAIPA

580

159,453

7,267

4.56%

3

JPMFR

377

106,168

11,853

11.16%

4

NEDAM

352

70,400

0

0

5

UBSCX

241

66,136

10,532

15.92%

6

CSILO

153

36,000

4,050

11.25%

 

340. The code CAIPA refers to Credit Agricole, and JPMFR refers to JPMorgan.  NEDAM (Flow Traders) is a specialist in algorithmic and high frequency trading and placed 352 large orders for a total of 70,400 lots, none of which traded.

Market abuse

341. The Tribunal has set out the relevant law on market abuse at [36] to [51] above and set out the allegations made by the Authority in taking the referred actions. 

342. The issue for this Tribunal is whether the Traders undertook an abusive trading strategy in Futures on the Exchange, alone and in collaboration with each other, by placing the Large Orders to facilitate the execution of the Small Orders on the opposite side of the order book and which they did not have an intention to trade, and whether such conduct was dishonest and lacked integrity.  The critical question is whether the Traders each had a (subjective) intention to trade the Large Orders, as explained by Andrew Baker J in Burford - the impression or signal given or likely to be given is only false or misleading if the Traders did not have an intention that the orders they placed would trade.

343.   It is not necessary for the Authority to establish that the conduct had an actual market impact.  It is about the impression or signal that the orders gave or were likely to give.  In Hannam v The Financial Conduct Authority [2014] UKUT 233 (TCC) the Tribunal concluded at [118] that the word “likely” in s118C(2)(c) FSMA 2000 is properly to be construed as meaning that there is a “real (in contrast with fanciful) prospect of that information having an effect on the price of qualifying instruments”, rejecting the submission that it meant “more probable than not”.

344. In making its findings of fact above, the Tribunal has not considered it necessary (or helpful) to set out the trading activity in all of the 233 Instances relied upon by the Authority (or even the 40 Specified Instances).  The facts of that trading activity are agreed between the parties.  We have taken account of all of the activity to which we were referred by the parties (whether in the course of oral or written submissions, witness statements, expert reports or cross-examination of witnesses) and refer to some of the Instances during the course of the Evaluation.

345. The Tribunal has found it helpful to address first the expert evidence in relation to market impact.

Evaluation - Whether Large Orders are likely to impact the market

346. One of the issues which Mr Creaturo and Mr Kasapis had been instructed to address in their expert reports was: “What impact, if any, would the trading instances identified in the Instance Pool have on the market (considered by reference to the impression and/or signal as to the supply or demand for, or price, of BTP futures that the trading instances would give or be likely to give to the market, if any)?”

347. The Authority’s position was that the Large Orders placed by the Traders would have been capable of giving an impression and/or signal of significant supply and demand to other market participants.  They relied on Mr Creaturo’s evidence, which included that the Futures market was particularly susceptible to changes in supply and demand due to the fact it had relatively low liquidity and high volatility.  The Traders relied on evidence from Mr Kasapis that the Large Orders would not impact the market based on various analyses including in relation to market liquidity and market price, as well as the size of the orders, which led him to conclude they would have had zero impact.  They also relied on the SMARTS data in relation to the large orders placed by other market participants and the execution rates (and cancellation rates) of such orders. 

348. We address here the evidence from the experts in the light of the challenges put in relation thereto which addressed both the relevant expertise and credibility of the two experts and their analysis.    

Tribunal’s assessment of the Experts

Mr Creaturo

349. Mr Creaturo had extensive experience having worked in fixed income trading for various global banks from the late 1980s to 2019, working for Lehman Brothers International, followed by Barclays Capital from 1997 and BNP Paribas in 2004 where he oversaw their core EGB trading, and then Nomura International in 2007 (until 2019).  He became more involved in trading BTPs during the Eurozone debt crisis as Italy experienced a significant spike in volatility during that time.  As Head of the EGB Desk at Nomura, he had to manage the overall credit risk with a strong focus on Italian government debt. 

350. Whilst all three Traders submitted we should prefer the evidence of Mr Kasapis, Mr Jaffey put forward the most direct challenges to Mr Creaturo’s expertise and credibility.

351. Mr Jaffey submitted that the majority of Mr Creaturo’s experience of trading EGBs derived from roles that he held with institutions that were principally primary dealers (in particular at Nomura, BNP Paribas, Barclays and Lehman Brothers); and that the vast majority of his trading was in Core EGBs not Peripheral EGBs.  Whilst the Tribunal recognises that the majority of Mr Creaturo’s experience in EGBs was in Core EGBs, Barclays Capital and Nomura (where he had worked during the Relevant Period) had not been primary dealers in EGBs when he joined them and we are satisfied that Mr Creaturo was well-placed to give opinions on trader behaviour across the EGB market, including in relation to the trading of BTPs and Futures. 

352. Mr Jaffey’s further challenges mainly related to Mr Creaturo’s opinions in respect of the trading activity of the Traders and his opinions on the explanations they put forward of their Trading Strategies.  Mr Jaffey criticised Mr Creaturo for failing to provide an explanation for the Lone Large Orders (which Mr Creaturo said were outliers) or the Traders amending the Large Orders after the Small Order filled, and submitted that his analysis did not address the significance of the Overlapping Small Orders (where Mr Creaturo accepted that it was a possibility that they were relevant) or the small orders where there was no large order on the opposite side of the book.  Mr Creaturo had explained his opinions on the Trading Strategies (as we address separately in the context of our assessment of the plausibility of them); the Tribunal does not consider that he was required to seek to explain every aspect of the Traders’ activity.

353. The broader challenge made was that Mr Creaturo had asserted opinions without reasoning, eg denying that mean reversion could be a meaningful concept in the context of the Desk’s activities in relation to the Information Discovery Strategy.  We do not consider that this is a fair challenge - Mr Creaturo’s opinions were clearly expressed and sufficiently reasoned, and were expressly based on his own relevant experience.  It is correct that he had adopted a different approach to setting out his opinions on matters relating to market liquidity and market impact than Mr Kasapis.  We address the latter’s approach below.  However, this was not a weakness in Mr Creaturo’s evidence.

354. Overall, the Tribunal concluded that Mr Creaturo’s evidence was well-informed and persuasive.

Mr Kasapis

355. Mr Kasapis is an experienced expert witness and is a Senior Director at Kroll Advisory (“Kroll”).  Prior to joining Kroll in 2017 he had worked at Rabobank, KPMG and National Australia Bank, and had experience in a range of financial products.  He has significant expertise in analysing trading data. 

356. However, as Mr Kasapis acknowledged, he had no relevant trading experience himself, with no experience of market making in EGBs (or any bond market).  Where he expressed his opinion by reference to his experience, the Tribunal was not satisfied that he had the requisite relevant experience, and certainly not such as would substantiate the opinions he expressed.  By way of example:

(1)        In the Summary of his expert report for the Tribunal Mr Kasapis said:

“With respect to the placement of these orders away from the touch, in my view a larger order offset to the touch would have meant that the order would be given much less consideration by other market participants compared to if those larger orders were placed on the touch (with zero offset to the touch).  In my experience, traders focus much more on orders placed on the touch rather than offset two or three cents behind the touch (in a highly liquid market, as this was, the bid ask spread will be narrow as the market is highly competitive, therefore bids and offers would be aggressive and narrow the spread).  Orders placed offset to the touch will in my experience be ignored by market participants, as would have been the case with the larger orders within the Specified Instances.”

(2)        Addressing the placing of multiple orders by Mr Sheth:

“In my experience of managing desks with junior traders, this explanation is plausible, and it is not an unusual error.”

357. Most of Mr Kasapis’s evidence was based on his analysis of the trading data.  Mr Shivji criticised various aspects of Mr Kasapis’s approach, in particular the assumptions made in the context of his “Other Participant Trade Analysis” which Mr Shivji submitted were unrealistic and overstated the “immediate liquidity” at the time of the Large Orders, and the averaging approach used in the “Market Trend Analysis” which smoothed out price fluctuations. 

358. The Tribunal considers that there were flaws in the approach adopted to analysing data (including rounding of time intervals where actual timings of traded volumes were available, and averaging of price movements) but nevertheless have found some of Mr Kasapis’s presentation of the data helpful.  This included his identification (or confirmation of data which had been presented to him) of data relating to the trading activity in the Instances, in particular the Overlapping Small Orders and the amendments to Large Orders after the Small Order traded.

359. There were areas where Mr Shivji criticised the scope of the work which had been done by Mr Kasapis, submitting that he had gone beyond the role of an expert in producing a list of cash bond orders relating to Specified Instances for Mr Urra and Mr Sheth which had not been identified or relied upon by the Traders (and some of which were then denied or doubted by them).  Whilst there was a clear error in this exercise, separate from whether the Traders agreed the cash transactions were relevant, in that Mr Kasapis had, for F67 (a Specified Instance for Mr Sheth) identified a cash bond trade with Banca Aletti without recognising that this was a split trade and thus failing to refer to the other leg of the trade in the other direction, the scope of the exercise undertaken was clear and we consider that Mr Kasapis was permitted to undertake this exercise; his report explained that this was his own identification of cash trades (in contrast to other sections where he set out that material had been provided to him and he had checked it and then put forward his opinions). 

360. Mr Shivji was particularly critical of Mr Kasapis’s report where he produced graphs to show that two very large trades (of 7,540 lots and 3,242 lots) in 2017 had traded without there being an impact on the market to move the price towards the opposite side, submitting that he had sought to pass off material that had been given to him by his client (Mr Urra) as his own independent analysis and had not sought to verify the information given to him. 

361. The Tribunal accepts that Mr Kasapis had not questioned or verified this information and did not in fact have the underlying data; and the Tribunal considered that the subsequent correspondence with the Authority following the Authority asking for the underlying data is unclear.  However, the Tribunal does not accept that the presentation of this data in Mr Kasapis’s two reports to the Tribunal was misleading - there is no reference to the source of the data in those reports (whether Eurex or otherwise) and the footnote simply says “the data for which I was provided and assume to be accurate”; he does state that these examples had been produced in his first report.  That is a reference to his report dated 12 November 2021 before the RDC, where (in section 4.3.5) he stated “I have been provided with…” without further explanation.  This criticism as to the presentation of these trades is a separate issue from whether Mr Kasapis should then have addressed the possibility that these trades were block trades (reported after execution) or roll trades.  

Summary of evidence of Mr Creaturo

362. Mr Creaturo’s opinion was that the Large Orders were likely to have an impact on the market:

(1)        The Futures market was particularly susceptible to changes in supply and demand due to the fact it had relatively (compared to more developed markets such as the Bund and US Treasuries) low liquidity and high volatility.  He described liquidity as referring to the ease of buying and selling at each price point. 

(2)        The Futures market used the iceberg functionality more than any other market.  It was an important feature of the market because the market is less liquid and participants were generally not encouraged to show their size - the market responded to outsize moves in client trades or larger trades.

(3)        There were typically small volumes shown at the touch.  Market participants use the volume in the stack to gauge whether the market is in balance or skewed towards the bid or offer price.  More volume on one side of the market than the other is called a “volume skew” and effectively suggests to other market participants that demand is high (when skewed towards the bid side) or supply is high (when skewed towards the offer side).  This implies that the next move in the price is likely to be upwards or downwards respectively.

(4)        This may cause a reaction in any market participant, but high frequency traders (and any algorithmic traders) are particularly sensitive to these calculations because they operate through algorithms, and are seeking to make profit by conducting high volumes of trades in both directions one or two ticks apart.  Even a small skew towards the bid or offer side can trigger algorithms to sell or buy so as to avoid a loss caused by a one-tick move in the touch.

(5)        The Large Orders were not iceberged and they were sufficiently large to give the impression or signal to market participants of a noticeable change to supply or demand in this market at the point they were placed.  His opinion was that the Traders’ challenge to whether the Large Orders were in fact “large” in size “misses the point” - it is the visible size of an order relative to the rest of the market at the time placed that is relevant when considering impact.  In this context, Mr Creaturo’s opinion was that whilst 200 lots of itself was not unusually large, the orders were large compared to other visible orders on the market at the time they were placed. 

(6)        Placing the Large Orders behind the touch was not consistent with the placing of large orders by other market participants.  When looking at the SMARTS data, he saw a number of institutions that were engaging with the market and wanting to trade; the outlier was MHI not trading.

(7)        That the Large Orders caused a volume skew at the time they were placed can be seen visually from the Replay graphs provided for each Instance.  They were often close to the touch (within three or four ticks), to which market participants are particularly sensitive. 

(8)        The visible size of the Large Orders (whether alone or combined), the way they were priced, and the liquidity and volatility of this market, all made it likely that there would be an impact on the market, and that the most likely market reaction was to move away from the Large Orders, in the opposite direction.  Therefore, if a Large Order was placed on the bid side it would be likely that the price of Futures would move up in line with higher demand, so that sell orders close to the touch were more likely to trade, and vice versa. 

(9)        The Small Orders in the Instances were on the opposite side of the Large Orders, usually at or very close to the touch.  In his opinion, a trader placing the Large Order would have taken the view that the most likely effect of that order would be to assist the execution of the Small Order on the opposite side of the book.  

(10)     It was striking that MHI had such a high success rate (ie execution rate) on their small orders and such a low success rate on their large orders; he would expect it to be consistent with the ratio of other banks, but in smaller sizes.

363. We assess Mr Kasapis’s challenge to this approach and the conclusions reached by Mr Creaturo below.  Mr Kasapis’s opinion was that the Futures market had “high general liquidity” in the Relevant Period and that there was “high immediate liquidity” surrounding the Specified Instances, such that the Large Orders could be “easily absorbed by the deeply liquid market” and the Large Orders would have “zero influence in such market conditions on other market participants”. 

Market liquidity

364. Mr Kasapis’s analysis was based on three limbs:

(1)        analysis of the liquidity of the cash market;

(2)        volumes traded in Futures, including the liquidity ranking given by the Exchange and a comparison of volumes of Futures in the Relevant Period with other times; and

(3)        an “Other Participant Trade” analysis which looked at traded volumes around the time of each Specified Instance.

Liquidity of the cash market

365. Mr Kasapis referred to there being instantaneous liquidity of approximately €200m in the cash market on any given bond (with a Futures equivalent of 2,000 lots).  This assessment was made on the basis that there were 20 primary dealers who were each obliged to make a two-way price of €10m each.

366. The Tribunal does not accept that this is a realistic assessment of the actual liquidity in the market at a particular price point.  The €200m is the maximum volume available, and is not all going to be available at the touch.  As Mr Creaturo explained, it would not be practical to access all 20 primary dealers at the same time, and trading with some of them in the same direction would be likely to cause a shift away in pricing of both the cash market and the Futures.  Mr Kasapis did accept this in cross-examination, and recognised that this was “not necessarily liquidity that someone can grab immediately”, noting that “everything comes at a price”.  

Comparison of traded volumes of BTP Futures in the Relevant Period with other times and markets

367. BTP Futures were the fifth and sixth most traded interest rate derivatives (out of 32) on the Eurex Exchange in June and July 2016 respectively.  The Tribunal notes that there is a huge range, and volumes plummet within the “top 10” in both months.  The most traded derivatives were Euro-Bund Futures, at 19,253,830 contracts in June 2016, with BTP Futures then at 3,399,472, and the shorter term BTP futures (in 10th) at 932,524.  The shape of the “top 10” is similar for July 2016, with Euro-Bund Futures at 11,203,674, BTP Futures at 1,567,151 and shorter term BTP futures at 380,380.

368. This data shows the difference in traded volumes in Futures between June and July 2016 - June was the month with the highest traded volumes for 2016, and July was second lowest (behind August).  The picture had been similar in 2015, with traded volumes of 3,191,548 in June and 1,633,723 in July.  The Tribunal finds that this significant difference between traded volumes in June and July 2016 was broadly predictable. 

369. Mr Kasapis conducted an exercise of comparing the volatility of Futures traded in the Relevant Period with two other timeframes - the Eurozone debt crisis (2010 to 2012) and the Italian Sovereign Bond crisis (May 2018).  Volatility in Futures prices in the Relevant Period was said to be low compared to these two periods, which was said to be consistent with the principle that a market’s liquidity has a significant, inverse, impact on price volatility - lower liquidity usually results in a more volatile market and causes prices to gap, higher liquidity usually creates a steadier less volatile market in which prices fluctuate in smaller amounts.   

370. Mr Kasapis set out his opinion that “Consequently, the ideal conditions to manipulate a market such as the BTP futures market would be a period of high volatility and intermittent low liquidity (an environment of relatively few buy and sell orders).”  The Tribunal accepts this proposition.  However, we do not consider that this assists with considering whether the Large Orders were likely to have a market impact in the Relevant Period.

371. Mr Kasapis had criticised Mr Creaturo for comparing the liquidity of Futures to US Treasuries and Bund futures saying that this was a “false comparison” but had then engaged in a similar comparison himself by reference to other futures traded on Eurex.  The short point is that these analyses of relative traded volumes between different futures contracts or between different time frames are of little assistance when considering the actual liquidity of the Futures market at the times of the Instances.  As Mr Kasapis accepted, what matters is the “actual liquidity that you can capture”.   

Other Participant Trade Analysis

372. Mr Kasapis concluded that the Large Orders of 200 to 500 lots would be easily absorbed by the high immediate liquidity of contracts being traded at that time.

373. Mr Kasapis looked at traded volumes around the time of the Specified Instances - in the five minutes before the start of the Instance (excluding any MHI trades), the traded volumes whilst the first Large Order was live (excluding any MHI trades) and then aggregating volumes and also looking at mean percentages.  He concluded that on average the market was highly liquid both before and during the Specified Instances and could easily absorb the Large Orders.

374. Mr Shivji criticised this approach, submitting it was flawed in several respects including:

(1)        Five minutes before the Instance is a very long time in a fast-moving market, and there is no evidence that market participants would have been looking at this.

(2)        Mr Kasapis had not looked at the prices at which these volumes had traded.

(3)        When calculating the liquidity whilst the first Large Order was live, Mr Kasapis had rounded to include traded volumes in the full minute in which the Large Order was placed, and the full minute in which that order was cancelled.  This was unnecessary where the actual traded volumes were available, and had the effect of overstating traded volumes whilst the Large Order was live.  For example, in F232, Mr Kasapis states that the traded volume in the five minutes before the Large Order is placed was 247 lots, and during T (ie the Large Order being live less the Small Order) the traded volume was 204 lots.  However, the Large Order was only live for 2.51 seconds, and Mr Kasapis’s approach meant that he used the traded volume for two minutes - the result is that 194 of the 204 lots identified by Mr Kasapis were traded at a time when the Large Order was not live.

(4)        He used only the first Large Order in the Instance, yet subsequent orders were sometimes closer to the touch, and the concurrent Large Orders included not just Multiple Large Orders placed by Mr Sheth but concurrent Large Orders across the Traders.  So in F31, Mr Kasapis focused only on Mr Urra’s Large Order even though Mr Lopez’s Large Order was closer to the touch.  

375. The Tribunal agrees with Mr Shivji’s criticisms of the approach taken.  However, the inclusion of the five minutes before the Instances, whilst not, we accept, relevant to the views of market participants at the time (with Mr Creaturo not being persuaded that any market maker would look at liquidity in this way), was helpful to us in understanding levels of traded volumes in the market around the times of the Instances (as otherwise we only had the traded volumes during the Instance itself).   

376. An extract from Mr Kasapis’s table, with the Authority’s addition of the concurrent Large Order size in each Instance, shows the following:

Trader

Specified Instance

Cumulative contracts traded

Concurrent Large Order Size

 

 

5 mins before T

During T (rounding minutes)

During T (less Small Orders traded)

 

Mr Urra

F7

807

327

287

444

 

F12

797

258

185

475

 

F32

17

45

36

450

 

F47

566

195

186

490

 

F60

656

268

256

444

 

F150

1,581

268

178

444

 

F152

748

281

248

444

 

F194

509

500

475

444

 

F203

346

72

64

250

 

F232

247

229

204

450

 

 

 

 

 

 

Mr Lopez

F42

257

5

-

200

 

F43

682

225

205

200

 

F56

696

24

19

200

 

F62

1,333

299

264

200

 

F63

2,603

153

148

200

 

F74

417

116

96

200

 

F132

800

99

96

200

 

F177

416

190

100

200

 

F190

261

89

44

200

 

F215

390

84

72

200

 

 

 

 

 

 

Mr Sheth

F55

818

369

364

500

 

F67

1,010

577

552

500

 

F103

1,509

282

272

1,000

 

F104

159

337

331

1,000

 

F125

1,551

325

310

400

 

F153

488

118

113

1,000

 

F158

794

376

342

1,000

 

F175

695

376

366

1,000

 

F176

571

405

315

2,000

 

F201

376

267

257

600

 

 

 

 

 

 

Multi Trader

F30

310

963

949

500

 

F31

531

279

240

899

 

F64

632

339

179

960

 

F82

454

615

578

1,400

 

F84

1,236

469

214

490

 

F94

1,029

128

113

500

 

F121

2,862

564

542

1,250

 

F174

644

219

194

1,700

 

F205

423

168

138

650

 

F209

1,298

447

408

700

 

377. The Tribunal identifies that there is a wide variation in the traded volumes across the Specified Instances, both objectively and by reference to the size of the concurrent Large Orders.  We do not agree with Mr Kasapis’s conclusion that this analysis shows high immediate liquidity during all of the Specified Instances. 

Whether Large Orders may influence other market participants

378. Mr Kasapis concluded that the Large Orders would have “zero influence in such market conditions on other market participants”.  This is based on:

(1)        “Market Trend Analysis”;

(2)        bid-offer spread analysis;

(3)        volume skew; and

(4)        two very large trades which went through the market in 2017.

Market Trend Analysis

379. Mr Kasapis said that the purpose of the market trend analysis was to identify whether the market was moving towards the Small Order before the Small Order was placed, as this would indicate that the Small Order would trade without the alleged inducement from an opposing large order.  Mr Kasapis did this exercise for the Specified Instances that were Single Trader Instances, not the Multi Trader Instances. 

380. Mr Kasapis calculated the three-minute moving average of Futures prices (using the price at each minute interval) and assessed whether that average was moving towards or away from the Small Order at the point the Small Order was placed.  (Two of the Specified Instances for Mr Sheth (F158 and F176) involved two Small Orders.) 

381. On this basis, Mr Kasapis set out the result as follows:

Market Trend (%)

Mr Urra

Mr Lopez

Mr Sheth

Towards the Small Order

70

40

67

Away from the Small Order

20

60

33

Neutral

10

0

0

 

382. Mr Shivji submitted that this analysis is flawed, with his reasons including:

(1)        Mr Creaturo’s evidence was that he didn’t know of any market maker using such an approach, and that this misses the point.

(2)        A three-minute moving average smooths out fluctuations in the market.  It ignores the usual price movements which take place within each minute of the day - and the Eurex market moves fast.

(3)        There is no logical rationale for selecting three minutes - the average length of the instances is one minute 36 seconds; and there is no suggestion that market makers would be using a three-minute moving average.  Mr Kasapis explained he had chosen three minutes based on the length of the Instances, not because of how the Futures market moved.

(4)        Mr Kasapis had just looked at the Small Order, without looking at the possible impact of the Large Order.

383. The Tribunal agrees with the Authority that the rounding and absence of marking of the time at which the Large Orders were placed are significant weaknesses in the approach taken by Mr Kasapis.  In addition, in some of the Instances the daily price graph did not appear to support Mr Kasapis’s conclusions - in F104, F153 and F176, the daily price graphs indicate that the market was trending away from the Small Order until the Large Order was placed.  These are all Specified Instances for Mr Sheth. 

Bid-Offer Spread Analysis

384. The purpose of this was said to be to assess the supposed induced movement compared to the daily movement of the Futures market, on the basis that a small supposed induced movement indicates that it was not likely to have been influenced by opposing large orders.

385. The average daily price movement in Futures is around 75 ticks, and the Bid-Offer spread is usually steady at one tick.  Mr Kasapis’s opinion was that because a one tick move is a very small and usual move, and because the Large Orders would be easily absorbed by the market (on the basis of his market liquidity analysis), the Large Orders would have zero influence in such market conditions on other market participants.

386. The Tribunal is not persuaded that price movements of one tick means that prices are incapable of being impacted by a Large Order - the average daily price range shows the level of volatility in this market.

Volume skew

387. Mr Kasapis disagreed with Mr Creaturo’s opinion that market participants would pay regard to any volume skew on the stack.  Mr Kasapis’s evidence was that whilst other market participants would know that the visible orders did not represent all the liquidity at any point in time (because of iceberging of orders), they would nevertheless be able to identify that the Large Orders were not part of an even larger iceberged order (as someone willing to show size of, say, 250 lots, would be considered to be unlikely to be hiding more). 

388. When challenged as to how market participants would know that the visible 250 lots were part of a single order (as, in contrast to the way the Replay graph was presented by the Authority for the purposes of these references, the Bloomberg Escalator showed the number of lots available at each price point without separating that information into orders), Mr Kasapis said that market participants would see the increase in volumes when the Large Order was placed and that the number of orders had increased only by one.  Whilst the Tribunal finds that the total number of orders live on the stack at any particular point of time was shown on the Escalator, we do not accept that human market participants would be able to unpick this in real time, and we had no evidence that, in a fast-moving market, algorithmic traders were set up to unpick the activity by millisecond to identify that a particular number of lots was part of a single order.

389. The Tribunal finds that other market participants would expect that there was more liquidity than was shown on the stack, but whilst they may have made informed guesses based on their own experience as to the amount of invisible liquidity, they would generally assume that the shape of the stack reflected the level of supply and demand. 

Two very large trades in 2017

390. Mr Kasapis’s reports had identified two very large trades which he put forward as examples of larger orders having no impact on the market.  They were trades for 7,540 Futures on 6 June 2017 and 3,242 Futures on 5 December 2017.

391. As referred to at [360] to [361] above, Mr Kasapis did not have the underlying data from Eurex and had not verified the information (which had been given to him by Mr Urra). 

392. The Tribunal accepts Mr Creaturo’s evidence that these transactions are very likely to be Futures positions being rolled by market participants - Futures contracts have an expiry date and it is common practice for market participants to roll positions forward to the next contract.  Further, these may well have been put through as block trades which means that the time when they were traded was different to the time when they were reported.  In all of the circumstances, they were likely not to have been outright transactions in Futures on the Exchange.   We have not taken these transactions into account.

Conclusions on market impact

393. On the basis of all the evidence before us, including not only the evidence from the experts but also the trading data from the Instances and the SMARTS data, the Tribunal reaches the following conclusions:

(1)        The Futures market was more liquid than most of the other interest rate derivatives markets on the Eurex Exchange in the Relevant Period.  However, it did not have high levels of immediate liquidity at all times of day - there was significantly less liquidity after the cash bond market closed at 17.30 CET, eg in F32 (Mr Urra) where only 17 lots had traded in the five minutes before the Instance, and there was often not a high level of immediate liquidity around the time of the Specified Instances.  This is evident from the Replay graphs produced by the Authority (as at the time each Large Order was placed in the Specified Instances) and by the Other Participant Analysis conducted by Mr Kasapis.  By way of illustration:

(a)        There was minimal liquidity at the touch in the following Instances (by reference to the actual size of the Small Order):

(i)         F12 (Mr Urra) - the Small Order to buy 73 lots is at the touch and there is visible liquidity of about 145 lots on the buy-side (including this 73 which was not iceberged and was at the back of the queue), but only 12 to 15 lots was visible at the touch on the sell-side (and about 80 to 90 lots on each side one tick away from the touch).  A total of 258 lots traded during the time the Large Order was live (using Mr Kasapis’s approach of including the traded volumes for the full minute in which the Large Order was placed and that in which it was cancelled) and the Large Order was for 475 lots;

(ii)        F125 (Mr Sheth) - whilst 1,551 lots had traded in the five minutes before the Instance, in F125 itself the Small Order to sell 15 lots (which was not iceberged) is at the touch and there are about 65 lots visible on the sell-side (more than 20 of which were ahead of the Small Order in the queue), whereas there were only about 17 to 20 lots visible on the buy-side at the touch and about 26 lots visible one tick further away on the buy-side; and

(iii)      F150 (Mr Urra) - the Small Order to sell 90 lots is at the touch, not iceberged and at the front of the queue.  It had started to fill and there were 61 lots still to be done when the Large Order was placed.  There were just ten to 15 lots visible on the buy-side at the touch and about 30 to 35 lots visible one tick away on the buy-side.  The Large Order was just two ticks from the touch on the buy-side.  Visible volumes were low on both sides of the stack until about 11 ticks away on the sell-side, where there were about 180 lots visible; and

(iv)      F177 (Mr Lopez) - the Small Order to sell 90 lots is at the touch and iceberged to six.  There were about 20 lots visible at the front of the queue, and about five to six lots behind it, which would be ahead of subsequent slices of the Small Order.  There were about 18 lots visible at the touch on the buy-side, with a further 43 lots visible one tick away from the touch.  Just 416 lots had traded in the five minutes before the Instance.

(b)        There were some Instances where, although liquidity throughout the stack was not high, there was a good level of visible orders at the touch and one tick away, comparative to the size of the Small Orders, eg F42 (Mr Lopez), F63 (Mr Lopez) and F64 (Multi Trader Instance).

(c)        There were Instances where the visible liquidity throughout the stack, not only at the touch but also up to four to five ticks away was multiples of the actual size of the Small Order.  This could be seen in, eg, F7 (Mr Urra).

(2)        Only 0.02% of Futures orders placed on the Exchange in the Relevant Period were of 200 lots or more, and whilst the number of such orders (2,872, of which 2,406 were placed by other market participants) means that they were showing on screens multiple times per day, this is a very small proportion of the Futures activity that was seen on the market.  This leads us to conclude that the Large Orders were large for the Eurex market.  (They were also large by reference to the size of MHI’s market making business and related hedging, but we do not consider that to be relevant when assessing market impact.)

(3)        That the Large Orders were large in size for the Eurex market was illustrated by the Replay graphs in most of the Instances to which we were taken throughout the cross-examination of the Traders at the hearing.  By way of illustration:

(a)        The Large Orders were a significant distortion to the shape of the stack at the time they were placed in most of the Instances, eg F12 (Mr Urra), F32 (Mr Urra), F64 (Multi Trader Instance), F125 (Mr Sheth) (where the Large Order was 400 lots and the visible lots were less than 100 at every price point on the entire stack), F150 (Mr Urra), F158 (Mr Sheth), F177 (Mr Lopez) (where the Large Order was 200 lots, although there were 130 lots visible on the sell-side seven ticks away from the touch) and F194 (Mr Urra).

(b)        There were some Instances where although the Large Order is distortive there were also distortions at other levels, eg F8 (Mr Urra) (where the Large Order of 333 lots is one tick from the touch and 14 and 15 ticks away from the touch are visible orders of 250 lots, on the same side as Mr Urra’s Large Order), F60 (Mr Urra) (where the Large Order of 444 lots stands out two ticks away from the touch but there were 225 lots visible four ticks away on the opposite side) and F62 (Mr Lopez).

(4)        Whilst the Futures market used iceberg functionality more than any other market, such that other market participants knew there was additional liquidity of unknown amounts throughout the stack, the visible lots of the Large Orders would not have been identifiable as single orders.  A consequence of this is that other market participants could not have reasoned that orders of this size were unlikely to be iceberged.

(5)        None of the Large Orders to which we were taken were placed at the touch (although there were Specified Instances for Mr Urra where he amended the price to move the Large Order to the touch, but then moved the order away again).  The Traders’ evidence was that market participants primarily focus on what is happening at the touch.  Whilst we accept that this would have been important, we accept Mr Creaturo’s evidence that the Large Orders were close enough (generally two to four ticks away) that they would not only have been visible but also would have formed part of other market participants’ views on supply and demand at the time at which they were placed.  This is particularly the case given the relatively low levels of visible liquidity at the touch in most of the Instances.

(6)        The result of this is that the Large Orders were each likely to give the impression or signal of significantly increased supply or demand, and we accept Mr Creaturo’s evidence that the most likely market reaction would be for the market to move in the opposite direction to the Large Orders, ie towards the Small Orders.  We do not accept the Traders’ submission that this conclusion is inconsistent with the SMARTS data in relation to the large orders placed by other market participants (which were also not iceberged).  Mr Jaffey’s submission, in particular, was that the Authority’s position was akin to arguing that all of these large orders would have been moving the market yet, he submitted, there was no evidence of this.  That is not our conclusion; we rely on the size, time at which they were placed and price of the Large Orders in the Specified Instances.  We did not have specific evidence in relation to the market, timing or the price of the 2,406 large orders placed by other market participants but 72.28% of these partially or fully traded and 80.34% of these were at the best or improved prices.  These were thus mainly at the touch, or crossing the spread, ie, as Mr Creaturo put it, engaging with the market.  (This can also be said of the “larger” Small Orders which were not iceberged, eg in F150 Mr Urra’s Small Order of 90 lots was not iceberged and had been placed at the touch.)  We have taken these other large orders into account, but the fact that they were placed, and we infer that those that did not trade (or trade fully) were cancelled (as a matter of good order management), does not undermine our conclusion in relation to the Large Orders placed by the Traders.

(7)        We accept Mr Creaturo’s evidence that a Trader placing the Large Order in these circumstances would have known of this likelihood. 

Evaluation - Whether traders committed market Abuse

394. The Authority’s pleaded case on market abuse is that the Large Orders were not placed for legitimate reasons, and in placing the Large Orders the Traders gave or were likely to give a false or misleading impression and/or signal as to the supply of, or demand for, the Futures to which the Large Order related; and that the Traders did not wish to trade the Large Orders and the purpose of placing them was to facilitate the execution of the Small Orders at a more advantageous price, or on a more timely basis, than would otherwise have been achieved.  The Authority relied on inferences which it submitted can be drawn from both the Criteria being met and from comparing the behaviour of the Traders with that of other market participants during the Relevant Period. 

395. According to the Authority’s case, the Traders were spoofing, a practice which is agreed to be abusive, and both the Information Discovery Strategy and the Anticipatory Hedging Strategy were said to be fabrications or inventions created after-the-event to fit the facts that we do know about.  Mr Shivji submitted that each Trader says they developed their new Trading Strategy at roughly the same time, following Mr Heiberg’s setting of a hit ratio.  Mr Urra and Mr Lopez developed different strategies in isolation from each other, with Mr Lopez saying that he had not heard of the Information Discovery Strategy before the Eurex Letter.  The strategies said to be being pursued are not the same, with even Mr Urra and Mr Sheth now describing their strategies in closing as “similar”.  The Mandate required that they be approved, yet none were.  All were outside the scope of market making, and were not used by other market makers.  They were deployed independently.  They did not result in a successful trade.  Mr Lopez says that the overlap with Small Orders is a coincidence.  All three Traders say that they did not collaborate.  Yet the pursuit of these strategies results in an identical pattern of behaviour; this is not a tiny sample, but it is most of their activity in placing large orders during the Relevant Period.   

396. Mr George, Mr Jaffey and Mr Bailin submitted (with different emphases based on the different positions of each of the Traders):

(1)        The Traders were each pursuing a legitimate Trading Strategy and had an intention to trade the Large Orders.  The Trading Strategies were not ultimately successful, but the Traders were each willing (indeed keen) for the Large Orders to trade whilst they were on the market.

(2)        The delays by the Authority and the lack of information that would have been available to the Traders during the Relevant Period has hindered the Traders’ ability to explain their specific trading activity.  They have necessarily had to try to reconstruct based on the information that is still available. 

(3)        The Criteria have been selected by the Authority as a means of identifying the Instance Pool from all of the trading activity.  It is therefore inevitable that the trading activity on which the Authority relies meets these Criteria and no sound inference of market abuse can be drawn from this.  The question is not whether there is a pattern of activity but whether there is an intention to trade.

(4)        In any event, the Criteria are compatible with, and explained by, the operation of both the Information Discovery Strategy and the Anticipatory Hedging Strategy, such that they do not justify an inference that the trading activity was abusive.

(5)        The Authority’s approach fails to take account of the full trading activity of the Traders during the Relevant Period, and offers no explanation for, eg, the placing of Lone Large Orders or why Large Orders were amended after the Small Order had traded.  This approach is unfair; Mr Bailin submitted it is an example of the “Texas sharpshooter” fallacy.  The placing of the Lone Large Orders is particularly significant - they are meaningful in number and they can only be explained by the Traders pursuing their Trading Strategies, as there was no small order on the opposite side of the order book.  The Authority has not put forward any explanation for these orders.

(6)        The Multi Trader Instances are not that at all; in each Instance, the Traders were not acting in concert with the others.

(7)        The Authority’s case is itself implausible.  The Large Orders were not sufficiently large in size as to create an impression or signal to the market as to supply of or demand for Futures.

(8)        The SMARTS data in relation to the behaviour of other market participants shows that a high number of orders of more than 200 lots were placed on Eurex and cancelled during the Relevant Period.

(9)        The Authority has no pleaded case on motive - it has not explained why any of the Traders would risk their reputations and careers to commit market abuse, which would have been exposed by the monitoring of activity on Eurex, to achieve small price advantages on the trading of the Small Orders; the scheme to commit market abuse would not have benefitted the Traders.

(10)     There is nothing abusive about placing orders with an intent to trade, or subsequently cancelling those orders in response to a perceived lack of demand.

(11)     The Authority’s case is that the Traders conspired to commit market abuse but never committed any part of their plan to any form of electronic communication, coordinated in an open plan office where they could be overheard by others, yet the other traders nearby did not hear anything suspicious, and the Traders have put forward two different explanations for their trading activity; the Traders submit this is not credible.

397. They referred to the differences between the Traders’ positions, submitting that these differences were supportive of the absence of a conspiracy or collaboration:

(1)        The three Traders were said to have been engaged in a single conspiracy to commit market abuse, yet had explained their placing of Large Orders by two different strategies;

(2)        Only nine of Mr Lopez’s orders were of 250 lots or more, with the majority of his Large Orders being for 200 lots.  If 200 lots were sufficient to give a false impression or signal to the market, why did Mr Urra and Mr Sheth feel the need to place orders of generally higher numbers, usually 450 lots or more, and why did the Traders sometimes place Large Orders concurrently, if an order of 200 lots was thought to be sufficient to facilitate the execution of the Small Order; and

(3)        Mr Sheth had placed Multiple Large Orders in Multi Trader Instances, involving both Mr Urra and Mr Lopez.  This served to draw attention to the behaviour; and in circumstances where Mr Sheth is alleged to be involved in an abusive scheme in collaboration with Mr Urra and Mr Lopez, their existence casts doubt on the Authority’s case as it would be expected that Mr Urra or Mr Lopez would have noticed Mr Sheth’s mistake and corrected him.

398. In addition, Mr Bailin submitted that for Mr Sheth it is not simply about the plausibility of the Information Discovery Strategy; he was a junior trader, and was doing what he had been shown by his manager, and this was a strategy which Mr Heiberg had considered plausible.  Mr Bailin submitted that if the Tribunal were to find as fact that Mr Urra showed Mr Sheth the Information Discovery Strategy, then Mr Sheth has not committed the deliberate and dishonest market abuse alleged.

399. The actual trading activity within the Instances is known and agreed; we need to assess the range of inferences that may be drawn from that activity.  The Tribunal takes the approach of assessing the various themes which arose from the parties’ submissions and the evidence, and then taking into account all of our findings of fact and this analysis and reaching conclusions in relation to the conduct and intentions of each Trader. 

Criteria used to identify the Instance Pool

400. The Authority submitted that the fact of the Instances meeting the Criteria gives rise to an inference that the Large Orders gave or were likely to give a false or misleading impression and/or signal as to the supply of, or demand for, Futures.

401. The Authority relied on Mr Creaturo’s evidence, which included that the trading activity in the Instance Pool has particular unusual features that are inconsistent with market making, in particular:

(1)        There is a notably high rate of cancellation in the Large Orders.  338 out of 341 Large Orders did not trade at all; three orders (all placed by Mr Urra) partly traded but were then moved away from the touch and/or cancelled so that they could not trade in full.

(2)        Despite the Large Orders being large for this particular market, none of the Large Orders were iceberged.  This is the opposite to what he would expect if there was a genuine intention to trade the Large Orders. 

(3)        The pricing of the Large Orders, having regard to their size, was unusual in this market, as they were nearly all placed away from the touch and nearly all kept away from the touch, which made them unlikely to trade.  The Traders placed 1.93% of their Large Orders at or better than the touch; other market participants placed 80.34% of their Large Orders at or better than the touch. 

(4)        The size of the Large Orders was inconsistent with MHI’s status as a lower tier market maker that did not engage in proprietary trading and considering the usual size and number of RFQs it received.  

(5)        The overlap between the Large Orders and the Traders’ Small Orders on the opposite side of the book was unusual.  Where traders need to deal with two orders, he would generally expect the priority order to be dealt with first, or for the orders to be netted. 

(6)        The repeated pattern of the last Large Order being cancelled once the Small Order has traded in full (or nearly traded in full) cannot, in his view, have been a coincidence. The frequency of the overlap and the timing of cancellation suggests to him a deliberate design.  His evidence was that it is the proximity in timing of the Small Orders trading and the Large Orders being cancelled that in his opinion is “particularly powerful” in showing that the Large Orders were a function of the Small Orders.

402. Mr Creaturo’s opinion was that “Taken collectively, [these] factors…are inconsistent with market making as I have known it throughout my career in trading.  In my opinion, the only plausible explanation for the strategy used is for the purposes of assisting the execution of the Small Orders and is designed to avoid trading the [Large Orders]”.  He had formed the opinion that the trading activity identified in the Instance Pool is “not explained by or consistent with the Traders’ explanations for their trading identified in their Replies and their witness statements”.

403. The Traders submitted by reference to the decision in Burford that the question is not whether there is a repeated pattern of trading activity but whether the Traders had an intention to trade.  We agree with that submission, although note that here the Authority is relying on the pattern (including, eg, moving the price away from the touch or cancelling the Large Order when the market gets close, the time for which the Large Order was live, and the timing of cancellation after the Small Order trades) as giving rise to an inference that the Traders did not intend to trade the Large Orders.

404. The Criteria and the resulting pattern relied upon by the Authority can be readily seen throughout the Specified Instances.  In the Single Trader Instances referred to below by way of illustration, the Large Order (or, where Mr Sheth had placed Multiple Large Orders, the Large Order closest to the touch) was cancelled within four seconds of the Small Order being filled:

(1)        F7 (Mr Urra) - The Small Order had been filling, but this paused for more than ten seconds.  The Large Order was placed, and the Small Order started to fill within 0.2 seconds.

(2)        F31 (Multi Trader Instance) - Mr Urra’s Small Order was live for more than 11 seconds.  Mr Urra placed a Large Order, and made several price amendments towards the touch and away again.  Mr Lopez then placed a Large Order and the price of the Small Order was amended closer to the touch.  Mr Urra then made further price amendments to his Large Order, towards the touch and away again, then back towards the touch, and within this time Mr Lopez also amended the price of his Large Order closer to the touch.    Mr Urra’s Small Order started to fill, and during that 2.5 seconds in which it was part filling in slices there were no price amendments to the Large Orders.  Trading in the Small Order paused for more than 3.5 seconds, the price of the Small Order was amended, there were further price amendments to Mr Urra’s Large Order towards the touch and the Small Order started to trade again.

(3)        F42 (Mr Lopez) - The Small Order had been live for more than 28 seconds, more than half of which at its amended price, and did not fill.  The Large Order was placed and the Small Order started to fill within 0.7 seconds.

(4)        F56 (Mr Lopez) - The Small Order was live for more than 12 seconds without filling.  It started to fill within 0.2 seconds of the Large Order being placed.

(5)        F60 (Mr Urra) - The Small Order was live for more than nine seconds without trading.  The Large Order was placed and the Small Order started to fill within 0.3 seconds.

(6)        F103 (Mr Sheth) - The Small Order was live for more than one minute 26 seconds without filling, more than half of which at an amended price.  The Large Order was placed and then cancelled after more than nine seconds.  The price of the Small Order was amended.  A second Large Order was placed and then, after four seconds, a third Large Order was placed, closer to the touch.  The Small Order filled within 0.2 seconds of that third Large Order being placed.

(7)        F150 (Mr Urra) - The Small Order had started to fill, but this paused for more than seven seconds.  The Large Order was placed and the Small Order started to fill within 0.1 seconds.

405. The Tribunal agrees with the Authority that the pattern of the trading activity (including the time of cancellation of the Large Orders) in the Instances could potentially support an inference that the Large Orders were placed to facilitate the execution of the Small Orders and that the Traders did not intend to trade the Large Orders.

406. The Traders submitted that the Criteria relied upon by the Authority are also consistent with them pursuing their Trading Strategies, and that therefore the mere fact of them being met does not support the inference drawn by the Authority.  We heard from Ms Mayer for Mr Urra and Mr Bailin for Mr Sheth in relation to the Information Discovery Strategy (whose submissions we have considered together) and from Mr Jaffey for Mr Lopez in relation to the Anticipatory Hedging Strategy.

407. Ms Mayer and Mr Bailin submitted that the Criteria are consistent with the Information Discovery Strategy being pursued by Mr Urra and Mr Sheth:

(1)        The size of the Large Orders was part of the Information Discovery Strategy - showing volume was part of what was intended to make the orders attractive.

(2)        This required that the Large Orders were not iceberged - the process of each slice being placed at the same price but at the back of the queue may slow execution; and the strategy requires that you show the full volume available to tempt a Posited Trader into paying a premium to offload their risk in one go.

(3)        They were seeking to trade at an advantageous price, in the belief that the Posited Trader would be willing to pay a premium to transact at volume, and Mr Creaturo had accepted in principle that there is nothing wrong with placing an order away from the touch.  Further, orders at the touch would be attractive and accessible to all, and would have reduced the informational value gleaned if the order were filled; placing away from the touch offered the trader the opportunity to create a profitable position.

(4)        Cancellations were very common in the market, and it was inherent in the strategy that the Posited Trader might not exist in which case it was highly unlikely that the Large Order would be filled at the price or volume at which it was placed.  Three of Mr Urra’s Large Orders traded in part, and one is a Specified Instanced, F30.  The cancellation is consistent with the strategy - Mr Urra wanted a Posited Trader to aggress the market; it was counterproductive to trade slowly on the back of a more general market movement.

(5)        The placing of the Small Orders and the Large Orders may have been prompted by the same client trade, which would explain their relative proximity or overlap.  The correlation in timing of the Small Order filling and the Large Order being cancelled does not necessarily imply causation - an underlying event, such as the market ticking up, may have caused the Small Order to fill and prompted the Trader to cancel the Large Order.  In addition, both Mr Urra and Mr Sheth had placed Lone Large Orders.

(6)        The Large Orders were not hedges for an executed cash trade.  Where no RFQ has been identified, the Large Order may have been prompted by other market events or information that is no longer available.

(7)        Mr Urra and Mr Sheth didn’t know how many lots a Posited Trader would be looking to trade (with Mr Sheth referring to this as a “strategy of discovery”); they simply needed to ensure the order was sufficiently large.  Round numbers fit with this.

(8)        The orders served distinct strategic purposes - the Small Orders were typically a hedge for a client trade, and the purpose of the Large Order was to provide an opportunity for MHI to trade at favourable prices.  It would have been illogical and commercially counterproductive to net the orders against each other.

(9)        The Traders agreed they deliberately deployed the strategy repeatedly.  Mr Sheth had also adjusted his approach after speaking to Mr Urra about his frustration with the strategy not working, by amending the price of the Large Orders and moving them closer to the touch.  Both Mr Urra and Mr Sheth decreased their use of the strategy when it did not seem to be working. 

408. Mr Jaffey submitted that the Criteria are similarly consistent with the Anticipatory Hedging Strategy:

(1)        The average size of Mr Lopez’s Large Orders was noticeably smaller than those placed by Mr Urra and Mr Sheth.   The size of 200 lots was chosen by reference to the size of RFQs he expected to and did receive and was within his risk limits, so he was able to keep the hedge, including overnight.

(2)        Mr Lopez would iceberg an order to allow him to execute slowly and adjust the price during execution.  For the Large Orders, he wanted them to execute, and to execute in full, at a competitive price that may only be available during a very brief pull-back in pricing. 

(3)        Mr Lopez was seeking to place anticipatory hedges at very precise levels where he thought there might be interest, and placed them at a price where he thought he could be aggressed and the orders could trade at a competitive price.

(4)        The pricing of the orders meant that the orders were less likely to trade; this also reflected the information disadvantage faced by the Traders and Mr Lopez’s cautious approach.

(5)        Mr Lopez was often doing a number of things at the same time; he placed 1,139 small orders in the Relevant Period and it is not surprising that some of them overlapped.  The Small Orders and the Large Orders were not connected.

(6)        The Large Orders were not hedges for executed cash trades.  They were placed in anticipation of future RFQs which Mr Lopez predicted would be received by the Desk.

(7)        The number of lots, usually 200, was based on the predicted size of the RFQs Mr Lopez expected to receive, and was a hedge for medium-sized RFQs.

(8)        Netting would make no sense.  The Small Orders were a hedge for his book; the Large Orders were trying to capture a price advantage.

(9)        Mr Lopez deliberately and repeatedly deployed his strategy.  It is accepted that none of his Large Orders traded.

409. The Tribunal has some reservations about the Traders’ submissions above, in particular in relation to:

(1)         Mr Urra’s and Mr Sheth’s position that they needed to show the size of their Large Order to the market (which goes to both the number of lots used and to the decision not to iceberg their orders) - in a market in which all participants knew that there would be invisible liquidity as a result of iceberging, and therefore that there was likely to be more available at the touch than was visible, they gave no explanation as to why a participant who wanted to buy in size would not place an order in size at the price they were prepared to pay and see how much of that traded through the stack.  If such other participant did indeed need to hedge, or were interested in trading at that level, they could see how much they could get done at that level.  The orders placed by Mr Urra and Mr Sheth did not need to be visible for the Posited Trader to do this;

(2)        Mr Lopez’s decision not to iceberg his orders; and

(3)        the coincidence in timing between the Small Orders and the Large Orders (which primarily goes to the proximity between the Small Order being filled and the cancellation of the Large Order for all Traders, and the fact of the overlap between the Small Orders and Large Orders for Mr Lopez). 

410. We consider that the first of these issues goes to the question of the plausibility of the Information Discovery Strategy and whether it was being pursued.  The fact of the overlap between the Small Orders and the Large Orders placed by Mr Lopez and the absence of iceberging is not necessarily inconsistent with him pursuing the Anticipatory Hedging Strategy when placing Large Orders.  However, the coincidence in timing between the Small Order being filled and the cancellation of the Large Orders (including in Multi Trader Instances where Large Orders had been placed by more than one Trader) is striking, and we revert to this below, but we accept that this coincidence in timing is not itself necessarily inconsistent with the Trading Strategies as described by the Traders.  

411. The Tribunal thus accepts that the Criteria relied upon by the Authority may potentially be consistent with the pursuit of the Trading Strategies as described by the Traders.    

The Trading Strategies - contemporaneous explanations

412. We have considered the way in which the explanations of the Trading Strategies have emerged as explanations for the trading activity.  This is not just with a view to assessing contemporaneous evidence of the existence of the Trading Strategies and consistency (which may not of itself merit any weight if the explanations now have been fabricated and designed to fit the available evidence), but also to consider the information that was available to MHI at various times and also to the Authority when it interviewed Mr Heiberg and Mr Joshi.

During the Relevant Period

413. There is no contemporaneous evidence of any of the Traders designing or seeking to implement a new trading strategy.

414. We accept that the only reason to produce a written explanation of the proposed Trading Strategies at the time would be for the purpose of discussing it with, and seeking approval from, Mr Heiberg and Compliance in accordance with the Mandate.  The process required by the Mandate for evaluation and approval of strategies was not followed in relation to either the Information Discovery Strategy or the Anticipatory Hedging Strategy - none of the Traders consulted with Mr Heiberg, Compliance or Risk Management in respect of their Trading Strategy, and Mr Lopez did not seek the approval of Mr Urra for the Anticipatory Hedging Strategy.

415. However, we would expect to see some records or monitoring of plans or outcomes during the Relevant Period, even if imperfect:

(1)        Mr Urra and Mr Sheth did not keep written records of the RFQs or clients where they decided to test the theory that clients were splitting orders, in circumstances where Mr Urra accepted that he would not have tested the strategy against, eg, Banca Popolare, and where Mr Sheth tested against some clients on multiple occasions but not for every RFQ, eg in relation to ICBC.  They did not keep records of how long they kept a Large Order on the market before deciding to cancel on each occasion.  They did not share with each other the outcomes from having tested the strategy in respect of particular clients.  Mr Sheth’s evidence was that he had expressed (to Mr Urra) his frustration that the strategy was not working, but this was not said to be any more detailed than saying that his Large Orders were not trading.  One purpose of the Information Discovery Strategy is said to be to gain information, and that there is some information to be gained even where a Large Order does not trade, yet there was no attempt to record this information. 

(2)        Whilst Mr Lopez has exhibited a list of the 93 RFQs the Desk had received during the Relevant Period where the cash trade would have required a hedge of 200 lots, there was no evidence that he had kept written records of patterns or predictions as to clients, timing, or direction at the time.  We would not, however, necessarily expect that Mr Lopez would have been keeping a record of his pricing analysis, or where he considered resistance and support levels were forming, throughout the day.

416. It was not said by any of the Traders that they had kept such records but that they were no longer available.

Reactions to the Eurex Letter

417. Each of the Traders has described the Eurex Letter as being about the placing of Multiple Large Orders by Mr Sheth: 

(1)        Mr Urra was shown the Eurex Letter by Compliance in the meeting on 29 July 2016.  His evidence was that the letter raised a concern over market manipulation and that the trading activity identified involved both Mr Sheth and Mr Lopez.  However, he said the issue raised by Eurex was the placing of multiple overlapping orders by Mr Sheth on these two occasions and, in cross-examination said he had immediately checked the market to look for market impact and then spoke to Mr Sheth but was reluctant to accept that he would also have spoken to Mr Lopez at that time.    

(2)        Mr Lopez’s evidence was that Mr Urra spoke to him about this letter at the end of July; that Mr Urra suggested that the issue involved Mr Sheth only; and that Mr Urra did not give him the impression that it was a very serious matter or that there was any sense of urgency. 

(3)        Mr Sheth’s evidence was that his understanding was that the Eurex Letter concerned his multiple orders.

418. We find that Mr Lopez and Mr Sheth were first told of the Eurex Letter by Mr Urra, and may not have seen it at the end of July 2016.  However, even if they had not previously seen or been given a copy, it was attached to the MHI Compliance Report (which they received in October 2016).

419. All three Traders maintained these explanations in their evidence, namely that the Eurex Letter was about multiple overlapping orders, rather than a large order opposite a small order where the large order is cancelled after the small order trades.  

420. The Tribunal finds that it is clear from the Eurex Letter itself that the questions being raised were not confined to multiple orders but about “bid orders with a high volume” whilst trading on the ask side, and the letter expressly referred to deleting the bid orders without execution.  Each of the Traders should have known this as soon as they saw a copy of the Eurex Letter; it is apparent from not only the description of the trading activity but also from question 3 asked by the Exchange.

421. We do not accept Mr George’s submission that Mr Urra’s understanding (or that of the other Traders) was reasonable, or that it was shared by others at the time.  MHI’s response to MSUSA on 22 August 2016 sets out an explanation of the bid orders as being used to establish a basis/curve position for some size and to understand the depth and direction of the market, and addresses why they were placed about the same time as the ask orders.  There is then a separate explanation of “why such large multiple orders were placed on 29 June 2016” by reference to a trader having intended to amend the price.  It is clear from this response, which was signed by Mr Joshi and had been seen in draft by Mr Heiberg, that MHI understood that it was being asked not just about the placing of multiple orders but about large orders opposite a smaller order where the large order was cancelled.  The Tribunal considers that Mr Joshi and Mr Heiberg did not take a different position when interviewed by the Authority in 2017.

422. The refusal of Mr Urra, Mr Lopez and Mr Sheth to acknowledge the scope of the enquiry raised by the Exchange, even in cross-examination at the hearing, is difficult to understand.

Interviews with Compliance

423. The notes of the Traders’ interviews with Compliance do not present a clear picture of the explanations that were given to Compliance by the Traders.  There are several references to the Traders being concerned that orders placed by the Desk were being front-run by algorithms, although that was then said not to be the reason for the trades highlighted by Eurex.  The notes all refer to price discovery (even those for Mr Lopez, who is said to have stated that price discovery was common at his previous employer), there are some references in the notes of Mr Urra’s interviews that he was trying to get size into the books in a curve position if other dealers were hit by cash bond sellers at once (and this was reflected in the draft note he provided on 2 August 2016), and the note for Mr Lopez refers to the multiple orders saying that this may have been a result of the settings on his PC.

424. The Authority sought to rely on the fact that these notes had been signed by each of the Traders without amendment.  They were signed by Mr Urra more than two months after the interviews, and by Mr Lopez two weeks after his interview; Mr Sheth’s signature is undated.  The Traders said they didn’t realise the importance of checking the accuracy of the notes at the time. 

425. The Tribunal is not satisfied that these notes are an accurate summary of the interviews - we do not know when they were prepared, it is clear that there are some substantive inaccuracies as can be seen from the reference to, and explanation for, Mr Lopez placing multiple orders (which he did not do, and this paragraph does not seem to have been copied across from one of the other notes as the drafting is different).  The dates on which they were signed by Mr Urra and Mr Lopez would suggest that these notes were presented to the Traders for signature just before MHI finalised the MHI Compliance Report.   

426. From reading these notes, even proceeding on the basis that they are somewhat inaccurate summaries, it is not apparent to the Tribunal that Compliance at that time identified from the explanations that were given that the Information Discovery Strategy was one that Mr Urra had devised and was not used by other market makers, that Mr Lopez was using a different strategy to that which had been described by Mr Urra and Mr Sheth, or that these strategies had been deployed on multiple occasions before and after the activity identified by the Exchange.

MHI Compliance Report

427. The Traders submitted that the Tribunal is now being asked by the Authority to reach a different conclusion to that which was reached by MHI following its own review which was conducted shortly after the Relevant Period.

428. Whilst we have only summarised and cited parts of the MHI Compliance Report, and were taken to selected pages, the Tribunal has read and considered the whole report.  The Tribunal is not persuaded that any weight should be attached to MHI not having concluded that the Traders committed market abuse:

(1)        The review conducted by MHI was based on a limited range of trading data (the two Instances identified by Eurex and five further (randomly selected) days in the Relevant Period), the interviews with the Traders described above were the only meetings between the Traders and Compliance before MHI produced its report and MHI did not have the benefit of expert evidence.

(2)        It is not clear to us that, at the time of this report, the Traders had explained and/or Compliance had identified that the Traders were providing two different explanations for their trading.  The report sets out, at section 4.1, what is said to be Mr Urra’s explanation - referring to large orders being to establish a basis position at an attractive price and price discovery (albeit with no reference to client splitting).  The first time anticipatory hedging is mentioned is when the report addresses the Mandate (as the Mandate contains its own description of anticipatory hedging) and says (at section 6.2) that “The large futures orders, to the extent that these were placed with a view to establishing a hedge against expected near term customer trades or to establish a basis position for future sale to clients, would fall into the category of “anticipatory hedging”, and are permitted by the Desk’s mandate.”  The report later refers (in section 7.2.3) to Mr Urra’s explanation, and includes that the Desk may have been used by clients to execute only a portion of their cash trades.  There is no specific reference to Mr Lopez or to a different explanation for the trading.

(3)        The conclusions reached by Compliance are in any event far from a vindication of the trading activity.  The report refers to “sufficient doubt” as to the Traders’ explanations, which left open the possibility that the trading could be regarded as abusive and expressed its overall conclusion as “indeterminate”.

429. The Tribunal does not place any weight on the conclusions reached by MHI in the MHI Compliance Report when considering whether the Traders’ activity constituted market abuse.  The Tribunal does consider separately the conclusions reached by MHI in relation to the Mandate, but again the reservations about the explanations, or understanding of those explanations, which had been provided at that time does significantly reduce the weight which is placed on those conclusions.

Disciplinary interviews

430. The disciplinary interviews were conducted by HR in October 2016.  The Tribunal’s reading of the notes of those interviews between the Traders and HR is that, whilst by no means being a transcript, they do seek to set out all the explanations provided by the Traders. 

431. The notes of the interviews with Mr Urra and Mr Sheth are broadly consistent with each other.  The note of the interview with Mr Lopez, which was then supplemented by his email sent the following day, sets out his explanation of the use of what he saw as the pricing levels in the market, although there is no reference there to seeking to predict the submission of medium-sized RFQs by clients.

Conclusions

432. There is a broad level of consistency to the explanations given by Mr Urra and Mr Sheth from their first interviews with Compliance to their disciplinary interviews.  The explanations are not entirely clear, they did not explain that this was a strategy which had been designed by Mr Urra or that these orders that were said to be to establish a basis position were being placed at the same time as a hedging order in the opposite direction. 

433. Whilst Mr Lopez’s Anticipatory Hedging Strategy does not emerge clearly from the note of his interview with Compliance, Mr Lopez’s explanations were consistent between the disciplinary interview with HR and the interview with the Authority, although he did not clearly explain any link with his focus on medium-sized RFQs and attempts to predict patterns in these RFQs which were received by the Desk. 

Mandate

434. The Traders each gave evidence that they believed their trading was permitted by the Mandate:

(1)        Mr Urra did not accept that the Desk was only permitted to carry out market making in cash EGBs and not in Futures.  He said he did not recall that the Mandate permitted trading in Futures only for the purpose of hedging.  He considered that speculative positions, even those taking directional risk, would be within the Mandate.  He did not think the Mandate was scrutinised, and interpreted it as a wide mandate, and took the position that as long as a trade was within the Desk’s limits, a trader was permitted (by the Mandate) to carry the risk.

(2)        Mr Lopez did not have a specific recollection of being shown the Mandate; but it is a document he would have known about, and the limits and concepts referenced therein were familiar to him and consistent with his understanding of what the Desk was allowed to do and did in practice.  Mr Lopez emphasised that the Mandate permitted “transactions in hedging instruments in anticipation of a highly likely near term exposure to risk, where there is a sound risk management rationale for such anticipatory hedging”.

(3)        Mr Sheth couldn’t recall having seen the Mandate at the time, but took the position that his manager, Mr Urra, had shown him the Information Discovery Strategy and therefore it must be within the Mandate.

435. Mr George, Mr Jaffey and Mr Bailin referred to the conclusions reached by MHI, submitting not only that the Trading Strategies complied with the Mandate but that what mattered was whether the Traders thought that they would be considered by Compliance to be in breach of the Mandate if they pursued the Trading Strategies.

436. The Tribunal finds it is objectively clear from the terms of the Mandate that:

(1)        The Mandate authorises market-making in EGBs, which would include cash BTPs, and hedging in EGB Futures (including BTP Futures) (which we find are Government bond futures and not exchange traded interest rate futures for this purpose). 

(2)        The Mandate does not authorise market making in Futures.   

(3)        The market making policy refers to the size of the Desk’s market making inventory and says it must be designed not to exceed, on an ongoing basis, the reasonably expected near-term demand of market making customers.  However, this reference to the inventory which may be held is clearly to the instruments in which the Desk may market make, ie EGBs/BTPs, but not the hedging instruments, the Futures.

(4)        The Mandate requires that overall trading and hedging strategies must be evaluated and approved by the Desk Head (Mr Urra) and Division Trading Head (Mr Heiberg) (after consultation with the Risk Management and Compliance departments).  It provides that Mr Urra and Mr Heiberg must demonstrate that proposed new or amended strategies will not present a high risk to MHI, can be managed within limits, will be appropriately controlled and will comply with the Volcker Rule’s requirements. 

437. We regard the Traders’ attitudes towards the Mandate as potentially significant. 

438. The Traders did not seek approval for their Trading Strategies in accordance with the requirements of the Mandate, a matter which was not addressed by the MHI Compliance Report.  This was in circumstances where Mr Urra was responsible for the Desk’s adherence to the Mandate and, on his own account, had devised a new strategy that was not being used by others in the market; whilst Mr Lopez described the practice of anticipatory hedging as “entirely normal practice”, this failed to reflect the fact that the size of the orders he was placing was targeted at a size of RFQ which represented only about 3% of those sent to MHI; and Mr Sheth put forward an explanation of not having seen the Mandate but assumed it must comply as he was being shown a strategy by his manager, even though the training he had undertaken at MHI must have made it clear to him that the Desk’s activities were constrained, and that he must act within limits.  

439. The MHI Compliance Report concluded that:

“The large futures orders, to the extent that these were placed with a view to establishing a hedge against expected near term customer trades or to establish a basis position for future sale to clients, would fall into the category of “anticipatory hedging”, and are permitted by the Desk’s mandate.”

440. We have already expressed our concerns in relation to the position as understood by Compliance at that time.  In addition, we consider that the framing of this conclusion, by reference “to the extent that…” the trades were placed for the stated purpose weakens the strength of this conclusion.

441. In his interview with the Authority Mr Joshi had described speculative pre-positioning in order to meet anticipated client demand as being within the scope of the Mandate.  Having read those pages, we agree that Mr Joshi was discussing the purchase of Futures rather than the cash bonds (as the Authority had referenced the orders being placed by Mr Sheth, asking how these fit into the facilitation of client business, expressly adding that this was effectively asking about the Desk’s Mandate).  Mr Joshi said it was legitimate to be positioning yourself in anticipation, not of the client coming back, but others in the market coming to execute an order.  The Authority questioned whether this was separate from making money from client business, and Mr Joshi said that they were not “outright prop trading” but nevertheless in order to service clients you need to maintain some inventory, and they need to manage that inventory.  However, Mr Joshi made it clear that they had asked themselves what the Traders were doing, as this was a client-based business, but had checked that they were within limits and had confirmed with the Chief Risk Officer that they were within their Mandate.

442. When Mr Heiberg was interviewed by the Authority, he told the Authority the Mandate permitted “anticipatory trading”, building a portfolio ready to sell to clients.  However, we consider that Mr Heiberg was discussing cash BTPs; the transcript does not support a conclusion that Mr Heiberg thought that the Mandate permitted building a portfolio of, and holding as inventory, Futures rather than cash BTPs.

443. Reading the MHI Compliance Report as a whole and the transcripts of the Authority’s interviews with Mr Heiberg and Mr Joshi, the Tribunal accepts that MHI had not at any stage concluded that the Traders were acting outside the Mandate.  We consider that MHI’s conclusion was at least in part based on the fact, which is agreed, that the Traders did not breach the Desk’s limits. 

Information Discovery Strategy - plausibility

444. We assess the plausibility of this strategy, which Mr Urra and Mr Sheth both say they were pursuing, albeit that Mr Urra was said to have devised it, Mr Sheth says he pursued it as he had been shown it by his manager (which Mr Bailin submitted was, for Mr Sheth, more significant than its plausibility given his junior status) and which, in closing submissions, Mr George and Mr Bailin submitted were “similar” to each other without providing any clarification as to how they differed.  We also refer to the Mandate here in the light of our conclusions as to the strategy, as the Authority’s submissions on the Mandate (essentially that the strategy was outwith its permitted scope) formed part of its submissions that the strategy was not being pursued. 

445. We have set out above Mr Urra’s explanation of the Information Discovery Strategy, including the problems it was seeking to address (MHI’s information disadvantage and clients splitting orders) and its intended benefits (information as to whether clients are splitting orders, which could be obtained both if a Large Order filled and if it did not, and a profitable position for MHI if the Large Order filled), as well as Mr Sheth’s explanation of his understanding of that strategy.

446. The Tribunal finds as fact that Mr Urra had placed an uniceberged order to buy 450 lots of Futures on 12 May 2016, that this order filled and that he told Mr Sheth about this at or around that time.  We make this finding on the basis of the data presented by Mr Sheth in his witness statement, which was not challenged by Mr Shivji.

447. The Tribunal also accepts that MHI’s status as a third-tier market maker meant it had an information disadvantage compared to larger, top-tier, market makers based on the agreed fact that it had less visibility over client flows in the market, receiving fewer RFQs and RFQs of smaller sizes than larger market makers.  We accept, as did Mr Creaturo, that the Desk would be seeking to improve its position in this regard.

Price discovery

448. Mr George submitted that Mr Creaturo and Mr Kasapis agreed that price discovery was a common and legitimate practice used by market makers - Mr Creaturo agreed that price discovery is a useful process for market makers operating in the BTP market as their job entails providing a price in bonds to clients. 

449. However, the Tribunal concludes that the Information Discovery Strategy as described by Mr Urra and Mr Sheth is not a “price discovery” strategy as that is understood and used in the market.  We accept Mr Creaturo’s evidence that price discovery is about validating the fair market value of an asset or security for a given size of transaction and occurs prior to a trade.  Mr Urra did agree that his strategy was not what other institutions would call price discovery.  We recognise that it was part of Mr Sheth’s case that the strategy he was pursuing was one of price discovery, but prefer the evidence of Mr Creaturo in this regard.  Furthermore, the Information Discovery Strategy was not used by others in the market - we accept Mr Creaturo’s evidence that he had not seen anybody else do trades like this, ever.    

450. That other market makers would place orders for price discovery, ie pursue a different strategy, does not therefore assist the Tribunal’s consideration of the Information Discovery Strategy.

Splitting of orders by clients

451. The Tribunal accepts that a client sending an RFQ to MHI may have been splitting its order with other market makers, and thus transacting at a larger size than was shown to MHI. 

452. Addressing the likelihood of this happening, we consider that this was not routine, it was simply a possibility.  In particular, based on the evidence of Mr Creaturo, we find that a client looking to do a large transaction would be very unlikely to split that order.  They would be best served by using a trusted counterparty, who would be able to transact at the desired size, to avoid leaving a footprint and running the risk that the market would start to move against it. 

453. Where a client did decide to split their order amongst market makers (rather than just seeking prices for a specified amount from multiple market makers), we find that they would typically divide their order between no more than three market makers, and would be likely to split in broadly even amounts.  Giving a simple example, a client showing €10m to MHI may be doing €10m or €20m with, respectively, one or two other market makers.  In reaching this conclusion we have preferred Mr Creaturo’s evidence based on his own experience; whilst Mr Kasapis was able to hypothesise as to the theory of how a client might approach multiple market makers, splitting €100m between, say, ten counterparties, he had no relevant experience of market making and we were not satisfied that his speculation as to this possibility would in fact ever be likely to happen in practice.    

454. The Large Orders placed by Mr Urra and Mr Sheth were mainly of 400 to 500 lots, which is approximately equivalent to €40m-50m of ten-year cash bonds.  The RFQs in respect of which they tested their strategy are much smaller than this.  The three trades identified by Mr Kasapis in respect of Mr Urra’s Specified Instances that Mr Urra agreed were plausible were each for cash bond trades of €10m.  Mr Sheth had identified cash bond trades of €0.5-10m in respect of his Specified Instances.  

455. Based on the size of the RFQs identified as having been sent to MHI in each of these Instances, reinforced in some cases by the identity of the client, the Tribunal considers that Mr Urra and Mr Sheth would have had no realistic basis on which to suppose that the client would, for these cash bonds, have done an additional €40-50m with another market maker.  We identify in this regard that their strategy was said to be seeking a single Posited Trader who had done a trade of that amount with the client, and not the position where a client had split a €40m trade across eight market makers, giving them €5m each.  

Likelihood of hedging by other market makers

456. The Tribunal accepts that another market maker who has traded cash is likely then to want to hedge swiftly with Futures.  However, this is likely, not certain, and will depend on other factors which cannot be known to MHI, including the positioning of the other market maker’s book, its risk appetite, the cost of hedging with a cash trade and its own view of anticipated short-term client flow.  In any event, such other market maker may not be prepared to pay a premium to hedge in Futures, and may not need to do so depending on the volumes available at or one tick from the touch.

457. Based on our conclusions about market impact, the Tribunal considers that placing a visibly large order away from the touch and expecting the market to come to that order was illogical, as the existence of that order would have a tendency to push the market the other way. 

458. In addition, in terms of the sizing of the order, not only would the market need to move towards the Large Order, but also the orders (visible and iceberged) closer to the touch would need to trade first.  Giving evidence, Mr Kasapis sought to suggest that orders at price points closer to the touch could fade, pull or cancel, so market participants don’t have to hit every single order in front of the Large Order.  We accept that some of the orders ahead of the Large Orders may be cancelled, or pulled back to a more attractive price.  But the point remains that another market participant seeking to trade in size would need to work through the stack to reach the Large Order, and there is invisible volume in that stack.

Whether placing Large Orders gave information benefit to MHI

459. There was said to be an information benefit to MHI resulting from placing the Large Orders, irrespective of whether or not the Large Order traded.  This was said by Mr Urra and Mr Sheth to be one of the two benefits of the Information Discovery Strategy (the second being the prospect of a profitable position if the Large Order traded).

460. Eurex is an anonymous Exchange.  If the Large Order traded, Mr Urra or Mr Sheth would not know who had transacted or why; it could be entirely unrelated to the RFQ which had been received by MHI.  Mr Sheth accepted that he could not be certain that a client was splitting orders, but said he would be entitled to a “high confidence guess” in this scenario that MHI’s client was splitting an order.  We consider that to be entirely unrealistic. One obvious alternative explanation would be that the filling of the Large Order would have been triggered by a market maker hedging a cash trade following a different, larger RFQ from another client that had not been seen by MHI (given that MHI only saw a small proportion of RFQs and RFQs of smaller sizes than other larger market makers). 

461. Furthermore, even if the trading of the Large Order had been a result of MHI’s client splitting its order (which would not be known), this would not inform Mr Urra or Mr Sheth as to the underlying size of the client trade or its likely approach in the future.  This is particularly the case where the client of MHI who had sent the RFQ was itself a clearer or broker, such as ICBC and Pershing Securities Ltd (“Sigma”), as they may be acting for a different client when they next approached MHI with an RFQ. 

462. Where a Large Order did not trade, this would not inform the Desk as to whether the client was splitting its orders.  The client may have shown MHI the full size.  They may have split their order, but the other market maker may have either not hedged in Futures, or hedged with the Futures that were available at or closer to the touch, eg in F47 (Mr Urra), there are about 250 lots that are visible ahead of the Large Order of 490 lots and the actual liquidity would not be known (to other market participants or to Mr Urra when placing the Large Order).

463. Mr Shivji submitted that there was in any event no need to place the Large Orders to test whether clients were splitting orders.  He submitted that MHI could have asked the client, or the broker, whether it had more to do; MHI could have observed by watching the volumes traded on Eurex at each price point straight after an RFQ was received; or MHI could have placed smaller sizes closer to the touch to see how quickly they traded (based on Mr Creaturo’s evidence). 

464. We accept that all market participants could see traded volumes on the Exchange; this might, if patterns were observed over time, provide some information, although we agree with Mr Sheth that the “signal” of others buying and selling could get lost in the noise of the market if they were hedging over time.  Similarly, we agree that MHI could have asked the brokers for information.  However, we agree with Mr Urra and Mr Sheth, as a matter of commercial sense, that a client seeking a price from MHI for their own desk would not, if asked, disclose if they were splitting their order.

465. However, whilst there are deficiencies in the other methods available to obtain information, the Tribunal is not in any event persuaded that MHI would obtain an information benefit from placing the Large Orders.

Prospect of a profitable position and risk

466. Mr Urra and Mr Sheth both gave evidence that if the Large Order traded, this would be a profitable position for the Desk, emphasising that the Large Orders were placed away from the touch.  Mr Sheth’s evidence in his witness statement was that he would have sought Mr Urra’s help to exit the position.

467. As a matter of fact, it is clear that the Large Orders in the Specified Instances were placed away from the touch, generally two to three ticks away, although some of Mr Urra’s Large Orders were at the touch during the course of the Instance.

468. At the risk of stating the obvious, any order, including the Large Order can only trade once the market price, the touch, has moved to the Large Order. 

469.  All three Traders submitted that the price of Futures would deviate from and then revert to a long-term mean - and more particularly that the Large Order would trade and then the market would immediately revert to where it was before the execution of the Large Order - and that this price movement could then be traded profitably.  This submission is particularly relevant to the Information Discovery Strategy, but Mr Jaffey also cross-examined Mr Creaturo on this in the context of the risks posed by the Anticipatory Hedging Strategy.   

470. Mr Kasapis’s evidence was that Mr Urra’s belief that the market would revert to its previous levels after transacting with a Posited Trader was well-founded.  In his view, in highly liquid markets, prices often exhibit “mean reversion”, ie a tendency to return to their average levels following temporary deviations or fluctuations.  In this context, when a trade is executed away from the equilibrium, the large number of active participants and high traded volumes act as a stabilising force, with participants taking advantage of any mispricing until prices are restored to their prevailing levels.

471. In cross-examination, Mr Creaturo did accept that prices fluctuate in the market, and can move back towards their previous levels after an initial movement, by way of reversion to mean.  He agreed that if a Posited Trader aggressively traded at a premium, the market could return to its previous level if there is no further demand or external factors driving a broader movement.  However, Mr Creaturo also considered it to be very speculative that the market would revert just because one trader had traded, and described any expectation of being able to profit in this situation as “wishful thinking”.

472. The Tribunal recognises that there was no data before us showing what actually happened where an order of 200 or 450 lots did trade.  However, based on the evidence of Mr Creaturo, the daily average high/low of the Futures market, the daily price graphs, the known directional risk, and MHI’s information disadvantage as a result of seeing only a minority of client flow which restricted its ability to ascertain any mean price to which the market was likely to revert (or the likelihood or timescale), or whether price variations in the market were an anomaly or client flow, we conclude that any trading based on the perception that there would be a mean reversion or correction immediately after a Large Order traded would have been speculative; Mr Urra or Mr Sheth may have been able to profit in this situation, but they could also have made a loss. 

473. During cross-examination, Mr Sheth gave the more detailed explanation of his “exit strategy” in the event that the Large Order traded, namely that he would execute an offsetting cash trade; and if he could not exit within a few seconds, he would have sought Mr Urra’s assistance.  Mr Sheth accepted that the cash market moves in line with the Futures, so if the Futures prices had moved to the price of the Large Order the cash would follow, but his evidence was that he observed that the cash market “moves with a lag” and that he would have a cash trade lined up as a hedge which he could execute before the cash market moved with the Futures (and that the disappearance of this cash hedge could explain the timing of some of his cancellations of the Large Orders).  However, there was no expert evidence supporting the existence of such a lag or that such a lag would be sufficient to enable a (human) trader, trading manually, to take advantage of this given that the cash market is automated.  This had not been addressed in any of Mr Kasapis’s four expert reports.  At the hearing, Mr Creaturo’s evidence was that it was not realistic to expect that Mr Sheth would be able to trade up to €50m in cash bonds at an attractive price before the cash market moved.  We accept that evidence, and recognise that in a relative value trade the trader would be expected to initiate the least liquid leg of the trade first, ie the cash trade, as the trader has more certainty on both liquidity and price in the more liquid instrument. 

474. The Tribunal concludes that if the Large Order had traded the relevant Trader would have acquired a speculative directional position, with the prospect of making a profit or realising a loss, and this would have represented a considerable risk on the book pending unwinding the position, the execution of an offsetting trade or winning a suitably sized client order in the right direction against which the position could be used as a hedge.

Mandate and the Desk’s aims

475. We have already recorded that Mr Urra and Mr Sheth did not seek evaluation and approval for the Information Discovery Strategy from Mr Heiberg in accordance with the requirements of the Mandate.

476. In addition, we conclude that the Information Discovery Strategy was not permitted by the Mandate:

(1)        It did not assist with the pursuit of the Desk’s aims, which were to routinely stand ready to purchase and sell BTPs and be willing and available to quote or transact in these instruments for its own account, in commercially reasonable amounts and throughout market cycles.  This is the essence of market making.  If the Large Order had traded, then the Desk would, until it closed out its position, have a directional position which meant that it would not be able to quote competitively in both directions.  This would have meant that if the client whose RFQ had prompted the use of the strategy came back to MHI for more, the Desk would be wrongly positioned against this client flow, which is contrary to the usual practice of market making (which is to trade in the same direction as client flow as opposed to against it).

(2)        The Mandate does not authorise market making in Futures.  It does allow hedging in Futures.  Transactions in hedging instruments must not be conducted where such transactions increase the Desk’s market risk exposures, other than “in anticipation of a highly likely near term exposure to risk, where there is a sound risk management rationale for such anticipatory hedging”.  If the Large Order had traded, this would have increased the Desk’s exposure to market risk, and this was not in anticipation of a highly likely near term exposure to risk.  The Information Discovery Strategy was not directed at an expected near term exposure to risk from its cash trades, let alone one that was “highly likely”.

477. The trading activity did not breach the Desk’s limits, and, if an order of 500 lots had traded, would not have breached those limits even if the Desk had not hedged that position.  The size of the individual Large Orders placed by Mr Urra were not of sufficient size alone to have resulted in breach to the Desk’s limits, although his trading within the Multi Trader Instances could have resulted in a breach if the concurrent Large Orders had traded, as could the concurrent Large Orders he placed in other Instances.  Mr Sheth’s Multiple Large Orders, if they had traded, could have breached the limits if the position were not unwound or hedged. 

Conclusions on plausibility

478. On the basis of all of the evidence, the Tribunal concludes that whilst MHI had an information disadvantage, and would want to improve its position, particularly given the introduction of the 15% hit ratio, the Information Discovery Strategy would not have been a plausible response to this:

(1)        Even if a client sending an RFQ to MHI was splitting its order with other market makers, there was no reason to suppose that a single market maker would have done cash trades of the magnitude envisaged with that client based on the size shown to MHI. 

(2)        Even if another market maker were, as was often described, “desperate to trade”, placing a visibly large order away from the touch and expecting the market to come to that order was illogical, as the existence of that order would have a tendency to push the market the other way. 

(3)        We do not accept that MHI would obtain an information benefit from placing the Large Orders, either if those Large Orders traded or if they were cancelled.

(4)        There was no guarantee or likelihood of obtaining a profitable position if the Large Order traded.  This would be a speculative position with directional risk, that was against the direction of client flow.

479. The Tribunal has also concluded that the Information Discovery Strategy would have been outside the scope of the activities permitted by the Mandate (although individually each of the Large Orders would, if they had traded and not been hedged, have been within the Desk’s limits.) 

Information Discovery Strategy - operation

480. We have considered whether the actions of Mr Urra and Mr Sheth are consistent with them pursuing the Information Discovery Strategy.  Mr Creaturo’s evidence was that it was not, and we have looked at various aspects of their trading activity. 

Clients in respect of whom the theory of splitting orders was tested

481. The premise is that Mr Urra and Mr Sheth were testing whether clients were splitting orders.  However, the approach they both adopted was not consistent in respect of each client and does not reflect the status of the particular clients (including their size, whether they were brokers or clearers and whether they would send multiple RFQs to MHI).

482. For the Specified Instances for both Mr Urra and Mr Sheth, Mr Kasapis had sought to identify whether there was an executed cash bond order or RFQ within close proximity to the Instance.  For five of the ten Specified Instances for Mr Urra, Mr Kasapis had identified cash trades which could have prompted Mr Urra to speculate that a Large Order might be tempting to a Posited Trader.  Mr Urra agreed that three of these appeared to be plausible matches, one was not and he had doubts about another:

(1)        Those Mr Urra agreed were plausible were in respect of F7 (the purchase by MHI of a €10m bond from ICBC), F12 (the sale by MHI of a €10m bond to CM Capital Markets Brokerage SA (“CM Capital Markets”)) and F150 (the purchase by MHI of a €10m bond from Sigma).

(2)        The one that was not was F47, where Mr Kasapis identified a cash trade for €25m with Banca Popolare, but that was a six month bond and Mr Urra said he would not be hedging this with nine lots of Futures, and he would not be testing his theory in relation to Banca Popolare.

(3)        He had doubts about F152 (the purchase by MHI of a €5m bond from Tullett Prebon (Europe) Ltd (“Tullett”)) as the cash bond had a maturity of 4.5 years and the Small Order was 33 lots, which Mr Urra considered would have been a lot of Futures to hedge that trade.

483. Mr Urra thus accepted, and we agree, that it would not necessarily make sense to test this theory against all clients.  Mr Urra agreed that he would not have been testing his theory against an RFQ from Banca Popolare in F47.  However, the decision to test the strategy in respect of ICBC, CM Capital Markets and Sigma, and Mr Sheth’s decision to test the strategy in relation to Method Investments and Advisors (“Method Investments”), raise significant questions about their approaches.

484. Mr Urra agreed that he may have been testing the strategy in respect of ICBC (in F7).  Mr Sheth said he had been using the strategy against ICBC in F67 on 16 June 2016, F125 on 23 June 2016 and F158 on 28 June 2016; and in one of his Lone Large Orders on 28 June 2016.  It is clear from the list of cash trades executed in the Relevant Period that ICBC was a regular client of MHI’s - there were 66 voice trades with them over the Relevant Period, and they would often come more than once in a day.  This could be seen on the following dates: 

(1)        On 10 June 2016 ICBC sent five RFQs to MHI, all in the same direction. 

(2)        On 13 June 2016 Mr Sheth did two trades of €5m each, both of which were MHI selling to ICBC. 

(3)        On 20 June 2016 Mr Sheth did three trades with ICBC between 11.02.12 and 11.09.37, all of which involved MHI buying €10m bonds. 

485. Mr Sheth had placed a Large Order in F82 at 10.43.48 on 20 June 2016, but did not place a large order during this interval where he was trading with ICBC - his next large order was placed at 11.15.33 (which was a Large Order during F83), and was placed shortly after he placed a Small Order of 50 lots.

486. The Tribunal agrees with the Authority’s submission that Mr Sheth’s large orders do not correlate with his trading with ICBC; but do bear a greater correlation with placing a small order in the opposite direction (even taking account of the number of Lone Large Orders placed by Mr Sheth). 

487. Furthermore, Mr Urra and Mr Sheth were both purportedly using the strategy against someone who regularly gave them repeat business, sometimes during a single day.  This means that if the Large Order had traded, they would be positioned against that client’s flow.  In addition, ICBC was a clearer for smaller financial institutions; so the Desk would not have known the identity of the underlying client on any occasion, and ICBC could have been acting for a different underlying client when it next approached MHI.

488. Mr Urra and Mr Sheth both said they may have been testing the strategy against other brokers or clearers - CM Capital Markets (Mr Urra in F12), Sigma (Mr Urra in F150, Mr Sheth in F176), Tullett (Mr Urra in F152).  Yet, the Desk would not have known the identity of the underlying client, which may be different each time the clearer sent an RFQ.

489. Mr Sheth’s evidence was that Method Investments triggered the use of the strategy in F103 (an RFQ that traded away), and two of the Lone Large Orders (on 21 and 28 June 2016).  Method Investments sent regular RFQs to MHI, but the average size was less than €600,000, generally with shorter maturities.  The RFQ Mr Sheth identified for the Lone Large Order on 28 June 2016 was a €1.3m cash bond of 4.5 years, which would have required about two lots to hedge.  The Lone Large Order was 500 lots.  For bonds of this maturity, Method Investments would have needed to place more than €300m elsewhere with another single market maker.  This seems incredibly unlikely in principle and Mr Sheth did not explain how he possibly thought this could be the case.

490. At the other end of the scale, for the Large Order in F153 Mr Sheth identified a cash trade where MHI sold €500,000 of a 15-year bond to Societe Generale (“SocGen”), and considered this had prompted him to deploy the strategy.  Yet SocGen was a primary dealer and would be likely to be dealing elsewhere; the Large Order placed was sized for 100 times the size shown to MHI.

491. The Tribunal recognises that the absence of Voice RFQs and Bloomberg/Chat RFQs has contributed to the difficulty for Mr Urra and Mr Sheth of identifying now the triggers for the use of the Information Discovery Strategy (although in the context of Mr Urra’s Specified Instances, this exercise had been conducted by Mr Kasapis and not Mr Urra).  They did have the data for all of the executed trades where MHI won an RFQ (in whatever form it had been submitted).  However, we have set out our concerns above in relation to Mr Urra’s and Mr Sheth’s apparently illogical and inconsistent decisions in relation to ICBC, CM Capital Markets, Sigma and Method Investments.  These were identified as the potential triggers in several Instances; and even if this reconstruction is not accurate for particular Instances, they gave clear evidence that they would have been testing their theory against these clients.

RFQ Traded Away

492. Mr Shivji submitted that it made no sense for Mr Urra or Mr Sheth to deploy the strategy and place Large Orders where there was an RFQ in which the client traded away, as in that situation they knew there was a trade out there being done with another market maker and its size, eg in F103 (Mr Sheth) where Method Investments traded away, and even came back several minutes later with another RFQ. 

493. We do not agree with this submission - the RFQ had shown MHI a particular size, and whilst that size was done elsewhere, it was no more or less likely in this situation that the client was splitting its order and doing more (which hadn’t been in the RFQ) elsewhere. 

Times of day

494. Mr Urra and Mr Sheth placed Large Orders at times of day where the possibility of a client having sent large RFQs to multiple market makers of the size envisaged was very unlikely (and where it would be difficult to close out the position if the Large Order had traded). 

495. The cash bond market closed at 17.30, and it was clear from the Replay graphs that there were few lots on the stack, and from the Spread graphs that traded volumes were very low.  This was the least liquid time of day.

496. Mr Urra placed Large Orders late in the day in F30 (a Multi Trader Instance) where he placed a Large Order for 500 lots at 17.39.45, in F31 he placed a Large Order to buy 499 lots at 17.45.22 and in F32 he placed a Large Order to buy 450 lots at 18.12.07.  Mr Urra agreed it was “very rare” to be transacting at this time of the day and that he would “definitely not” have wanted to leave the position open over the weekend if it had traded.  There was no evidence to support an expectation that another market maker had traded €45m and would be needing to buy Futures to hedge their position.

497. The latest time in the Specified Instances at which Mr Sheth placed a Large Order was for 500 lots in F55 at 16.55.42.  The lack of volume was less stark at this time - 818 lots had been traded in the five minutes before the Instance, but volumes were lower during the Instance and the stack was light. 

Lack of documentary record of operation of strategy

498. There was no contemporaneous record of the occasions on which Mr Urra or Mr Sheth deployed the strategy.  They were not recording, or sharing with each other, the clients where they decided to test the strategy, or the outcomes (including, eg, where the cancellation decision was driven not by having reached a conclusion on whether the client was splitting its order but by market events).  This was in circumstances where their evidence was that they were not collaborating and were each operating independently, both tested the strategy against ICBC and Sigma, and Mr Sheth was expecting to ask for Mr Urra’s help to exit the position if the Large Order traded and he was not able to unwind it himself within a few seconds.

499. The lack of any documentation in relation to the operation (and any results) of any strategy is illustrated by the fact that Mr Urra was not able to provide any evidence based on recollection or contemporaneous notes of the trade on 12 May 2016 on which Mr Sheth relied, the three occasions on which his Large Orders in the Instance Pool traded in part, or the three Non-Instance large orders which traded (two in part, one in full).  The Tribunal considers it reasonable to expect that, if these were examples of the successful implementation of the strategy - or at the very least those on 12 May 2016 and where the Non-Instance large order traded in full - Mr Urra might have written down, or shared with others at the time, the identity of the relevant clients or triggers for the use of the strategy.

Timing for which Large Orders were live and timing of cancellation

500. The Large Orders were often live for just a few seconds, and then cancelled shortly after the Small Order filled. 

501. Mr Urra’s Large Orders in the ten Specified Instances that were Single Trader Instances were live for 1.899 seconds (in F7) to 46.3 seconds (in F47).  The order that was live for 46.3 seconds in F47 is an outlier; it is the first of two Large Orders placed in F47, and when it was cancelled the Small Order had not traded; the second Large Order in that Instance was live for 2.13 seconds.  The range without that first Large Order in F47 is from 1.899 seconds (in F7) to 4.815 seconds (in F12).

502. Most of the Specified Instances that were Single Trader Instances for Mr Sheth include Multiple Large Orders, and he would cancel the one closest to the touch first, and that is the one that had been placed most recently.  In F55, F67 and F125 (each of which involved just one Large Order), the Large Orders were live for 1.646, 2.66 and 1.433 seconds respectively.  There were three Large Orders in F103, live for 9.331, 6.679 and 1.478 seconds respectively.  Illustrating this another way, F104 and F153 both involved Mr Sheth placing two Large Orders - the time from placing the first, to cancelling the second was 4.581 and 4.02 seconds (ie at that point the first Large Order was still live, and cancelled shortly afterwards, but this is the period of time for which Mr Sheth says he was testing the theory that a client was splitting its order).

503. We are not persuaded that these short periods of time were sufficient to test a theory that a Posited Trader might be looking or willing to trade in volume at that time, and reach a conclusion that there was no such Posited Trader.  This is particularly apparent where, as in F104, there had already been a gap between the RFQ which triggered the use of the strategy and the placing of the Large Order - if it was thought that it was worth placing a Large Order around 45 seconds after the RFQ was received, it is inconsistent to then conclude that the Large Order need only be on the order book for less than five seconds. 

504. This is supported by Mr Creaturo’s evidence that the variation in the length of time the Large Orders were left on the book is not explained by a desire to test whether clients were splitting orders.  Instead, his opinion was that the linking factor in what drives the cancellation of the final Large Orders in the Instances seems to be the fill of the Small Orders on the opposite side of the book. 

505. In the Single Trader Instances which were Specified Instances, the time from the fill of the Small Order to cancellation of the Large Order for Mr Urra ranged from 1.339 seconds (in F7) to 4.646 seconds (in F12).  As noted above, the first Large Order in F47 was cancelled before the Small Order filled; the second Large Order in that Instance was cancelled 1.797 seconds after the Small Order filled.  Mr Urra drew attention to the slightly longer cancellation times in the Instance Pool as a whole.  However, from the Tribunal’s review of further Instances, the pattern remains the same, in that the length of time for which the Large Order was live varied significantly more than the cancellation time after the Small Order filled.

506. For Mr Sheth, the time at which he cancelled his Large Orders (or the Large Order closest to the touch where there were Multiple Large Orders) after the Small Order filled ranged from 0.14 seconds (in F175) to 2.529 seconds (in F176).

507. In his witness statement Mr Sheth referred to his activity in F52 (a Multi Trader Instance involving Mr Lopez and Mr Sheth, which was not a Specified Instance).  Mr Sheth had placed a Large Order to sell 500 lots whilst Mr Lopez had a Small Order to buy seven lots, had reduced the price and then cancelled the Large Order, at a point at which Mr Lopez’s Small Order had not started to fill.  Mr Sheth said his Large Order was placed for his strategy, and that if the Authority’s theory was correct he would not have cancelled the Large Order at that time. 

508. We are not persuaded that the activity in this Instance supports this submission.  Mr Sheth’s Large Order had been live at its amended price of 140.74 for 13.392 seconds.  Mr Lopez also had a Large Order to sell 200 lots for 140.73.  The Spread graph shows a spike in traded volumes and the market moved towards the Large Orders.  Mr Sheth’s Large Order was at the touch, and he immediately cancelled the order, as did Mr Lopez.  They cancelled their Large Orders within 0.2 seconds of each other.  Whilst we recognise that Mr Lopez’s most recent Small Order had not traded, Mr Sheth cancelled his Large Order just at the moment when a sharp price movement meant he could have traded.  Instead, he decided to cancel.     

509. Giving evidence, Mr Sheth referred to F94 (a Multi Trader Instance involving Mr Lopez and Mr Sheth) as an Instance where he cancelled the Large Order he had placed before Mr Lopez’s Small Order had filled; and had also cancelled a Large Order after Mr Lopez’s Small Order had been filled.  He sought to rely on this to show there was no correlation between the cancellation of the Large Order and the fill of the Small Order.

510. We do not find the activity in this Instance supportive of Mr Sheth’s explanation.  There is various activity in this Instance:

(1)        Mr Sheth had placed a Small Order to buy ten lots and a Large Order to sell 500 lots.  The Small Order filled and he cancelled the Large Order.

(2)        Mr Lopez placed a Small Order to buy five lots, which filled.

(3)        Mr Lopez then placed a Small Order to sell five lots.  Mr Sheth placed a Large Order to buy 500 lots and cancelled that order more than six seconds later (without any fill on the Small Order).  However, Mr Sheth’s next action was to place a new Large Order to buy 500 lots (at a higher price).  Mr Lopez’s Small Order filled and Mr Sheth cancelled the Large Order just over one second later.

511. So here, the cancellation of the first (buy) Large Order by Mr Sheth, whilst that did take place when the Small Order had not traded, was followed by the placing of a new Large Order of the same size and in the same direction.  We do not place any weight on the decision to cancel that first (buy) Large Order.

512. Mr Urra’s and Mr Sheth’s evidence was that in any of the Instances they may have cancelled the Large Order as the market was moving away from that order; and the fill of the Small Order and the cancellation of the Large Order may have been triggered by the same market event; and Mr Sheth also gave evidence that he may have cancelled the Large Order if the cash trade he had lined up to hedge his Futures position disappeared. 

513. This is an issue where MTS data and market information from the Relevant Period would offer the potential to assist with explaining the timing of the decision to cancel the Large Orders (according to the evidence of Mr Urra and Mr Sheth). 

514. However, even allowing for the possibility that some market information would support their explanations, the Tribunal considers there are difficulties:

(1)        There are various Instances where, having decided to cancel a Large Order, the Trader then placed another Large Order within a short period of time, eg F64 (Mr Urra), F94 (Mr Sheth in the Multi Trader Instance above) and F103 (Mr Sheth). 

(2)        In other Instances, Mr Urra and Mr Sheth cancelled the Large Orders once the market seemed to be moving towards that Large Order, eg in F47 (Mr Urra), F52 (Mr Sheth), F104 (Mr Sheth) and F125 (Mr Sheth).

(3)        Three of the Large Orders in the Instance Pool traded in part.  These were Mr Urra’s Large Orders in F16 (three lots out of 444), F30 (39 lots out of 500) and F167 (42 lots out of 444).  In F16, Mr Urra amended the price of the Large Order away from the touch 0.531 seconds after it started to fill; once the Small Order traded he amended it further away again and then cancelled.  In F30 he cancelled the Large Order 0.948 seconds after it started to fill, and in F167 he cancelled the Large Order 2.131 seconds after it started to fill.  Mr George put it to Mr Creaturo that Mr Urra may have cancelled because these orders had not traded in one go, ie they were trading because of market movement and not a Posited Trader buying significant volume in one shot.  This submission is difficult to accept - the reality is that Mr Urra would have no way of knowing who was acquiring the position or the size in which they were buying.  Whilst volumes varied significantly, there were always other visible lots ahead of the Large Order at the time at which it was placed, with the uncertainty of how many iceberged lots could also be available.  A trader looking to buy, eg, 400 lots, would trade through the visible and invisible lots on the stack, and may then only trade, eg 30 or 80 lots from the Large Order.

Placing of new Large Orders shortly after cancellation and switching of sides

515. The Tribunal has considered the activity of Mr Urra and Mr Sheth in sequential Instances that are in close proximity in time to each other, and within Instances where multiple Large Orders were placed that were not concurrent.  Both Mr Urra and Mr Sheth gave evidence that the Large Orders were placed pursuant to the Information Discovery Strategy, and that the timing of the cancellation of the Large Orders may have been driven by events in the market in relation to which there is no longer information available. 

516. The Tribunal focused on the Specified Instances, but did also look at the Instance Pool as a whole (alongside the data in relation to cash bond transactions, Electronic RFQs and daily price graphs) to see whether the activity might provide some support for the position being put forward by Mr Urra and Mr Sheth.

517. From the very beginning of the Relevant Period, Mr Urra would place a Large Order, decide to cancel then place a new Large Order shortly afterwards.  As noted above, there is no evidence in relation to market events, but there is evidence of pricing (from both the daily price graph and the Spread graph during each Instance).

(1)        F1 and F2 took place on 7 June 2016.  Mr Urra placed a Large Order to sell 444 lots for 141.09 at 8.56.00.393, and cancelled that order at 8.56.06.604.  Mr Urra then placed a Large Order to sell 490 lots for the same price at 8.56.45.116, cancelling it at 8.56.47.206.  The orders were live for 6.211 seconds and 2.09 seconds respectively.  Mr Urra’s case is that on two occasions within 30 seconds of each other there were two cash market actions which led him to place an offer in the Futures market in anticipation of a trade hedge being entered by a Posited Trader, who would then execute in the Futures market to the gain of Mr Urra’s Large Orders; and it was a coincidence of timing that on both occasions Mr Urra, having placed his Large Order shortly after Mr Lopez placed a Small Order in the opposite direction, cancelled his Large Order within two seconds of the Small Order being filled.

(2)        F79 and F80 took place on 20 June 2016.  Both of these Instances involved Mr Lopez placing a Small Order on the buy-side.  During the course of these two Instances, Mr Sheth placed three Large Orders, each to sell 500 lots at 140.86.  They were not concurrent - the first was placed at 9.57.32.754 and cancelled at 9.57.44.609, the second was placed at 9.57.53.165 and cancelled at 9.58.00.899, with the third placed at 9.58.09.700 and cancelled at 9.58.12.174.  In isolation, it is difficult for the Tribunal to conceive of a legitimate strategy that, within the space of 40 seconds, would see the same order, at the same price, placed and cancelled three times.  The timing instead appears to be linked to the placing and filling of Mr Lopez’s Small Orders.

518. We were frequently referred to F176, in which Mr Sheth placed a total of seven Large Orders (all in the same direction).  He cancelled his sixth (and at that time final) Large Order to buy 500 lots at 15.31.16.40, less than three seconds after his Small Order of 80 lots filled.  However, he then placed another Large Order to buy 500 lots at 15.31.37.500.  That seventh Large Order was live for less than two seconds.  It had been placed almost four seconds after he had placed a second Small Order to sell ten lots, and was cancelled 1.73 seconds after the second Small Order filled.

519. Both Mr Urra and Mr Sheth placed Large Orders in different directions within a short period of time:

(1)        F15 is an Instance to which Mr Urra referred in his witness statement.  It is a long Instance, taking place over 40 minutes.  However, within that time Mr Urra placed a Large Order to sell 450 lots for 141.72 at 16.01.05.469, cancelled it at 16.01.08.667, then placed a Large Order to buy 444 lots for 141.72 at 16.03.44.664, which he cancelled at 16.03.46.655.  Mr Urra’s third Large Order in this Instance is a Large Order to sell 444 lots at 141.64, but that is placed 17 minutes later.  All three of these Large Orders overlapped with, and were in the opposite direction from, a Small Order.  However, at the beginning of the Instance Mr Urra had placed two Small Orders, each to buy ten lots, which remained live until the very end of the Instance and were therefore on the same side as his second Large Order.  This is unexplained.

(2)        In F158 Mr Sheth placed a Small Order to sell 17 lots, then placed two Large Orders to buy 500 lots (at 14.54.25.109 and 14.54.27.450), the Small Order traded, he cancelled the two Large Orders and then he placed a Small Order to buy 17 lots and two Large Orders to sell 500 lots (at 14.54.49.370 and 14.54.50.550), which were then cancelled after the Small Order filled.  (Mr Sheth identified that the first Small Order would have been a hedge for a cash trade with ICBC, and that it was this cash trade that triggered him to deploy the strategy.)  The cancellation of the Large Order closest to the touch was on both occasions within two seconds of the Small Order being filled.

Prospect of a profitable position

520. We have already reached our conclusions in the context of plausibility as to the prospect of the filling of the Large Order resulting in a profitable position.  There are only three Instances in the Instance Pool in which any of the Large Orders filled, and this is not sufficient to enable us to seek to draw any specific conclusions in those Instances.    

Overlap between the Small Orders and the Large Orders

521. The overlap between a Large Order and a Small Order in the opposite direction is one of the Criteria.  The Authority challenged Mr Urra’s and Mr Sheth’s explanation that the two orders were often prompted by the same client activity, and that the overlap was thus a feature of the Information Discovery Strategy.  This was on the basis that:

(1)        the client activity that has been identified does not justify the size of the Large Orders;

(2)        they have not identified triggers for many of the Specified Instances, and such failure cannot be explained by reference to missing information, as the list of actual trades that would lead to a Small Order is comprehensive; and

(3)        the Small Orders were not always prompted by a client RFQ or trade in any event, eg in F174 Mr Sheth identified a €5m cash trade that Mr Lopez had executed as the trigger for his Large Order, but Mr Lopez’s evidence (as is clear from the cash trade data) was that this was not a client trade but Mr Lopez buying a bond for his book on MTS.  There was no “client” who could be splitting orders.

522. The Tribunal considers that the above submission raises several different issues.

523. We agree with the Authority on the first of their reasons - whether it be the identity of the client, the size of the RFQ, or the prospect of there being size in the market at the relevant time of day, the “triggers” identified by Mr Urra and Mr Sheth (or suggested by Mr Kasapis and agreed by them to be plausible) are inherently implausible.  They would involve a client splitting an order on a bizarre basis, unseen in the market.

524. We consider the Authority’s second reason in support of its submission does not fully address the explanations provided by Mr Urra and Mr Sheth as to Large Orders being prompted by Voice RFQs or Bloomberg/Chat RFQs which traded away, and thus where there was no cash trade executed by MHI.  However, if a Large Order was placed to test the theory of clients splitting orders in relation to an RFQ which had traded away, this itself raises a different question as to the overlap between the Small Orders and the Large Orders, including as to the very fact of the overlap (separate from the coincidence in timing of cancellation of the Large Orders).

525. As to whether they have identified a trigger, Mr Sheth has identified the activity which he says could have triggered his use of the strategy for most of the Specified Instances.  It was Mr Urra who has not reconstructed this information, and only considered that three of the cash bonds identified by Mr Kasapis were plausible.  It was, however, Mr Sheth who had tested the strategy having identified that Mr Lopez had bought a bond on MTS, as in F174, where there had been no client RFQ.  The fact of an overlap in that situation can be attributed to Mr Lopez placing a Small Order to hedge that cash transaction; however, we consider it almost incredible that in this situation Mr Sheth would hypothesise from this that a client might be splitting orders.  This is made even more unlikely by the fact that this was one of the two Instances identified by Eurex at the end of July 2016, so Mr Sheth would have been well-placed at that time, having been interviewed by Compliance on 3 August 2016, to have investigated the activity to be able to provide as much evidence as possible to Compliance at his second interview on 10 August 2016.

526. In addition, Mr Sheth identified that the Authority had challenged a Large Order he had placed in an Instance that did not overlap with a Small Order.  In his witness statement Mr Sheth referred to his placing of two Large Orders in F151, both of which were relied on by the Authority.  The second Large Order in that Instance was placed 2.655 seconds after Mr Lopez’s Small Order filled.  He said this was not consistent with the Authority’s case theory (and drew attention to Mr Lopez’s Small Order to buy 97 lots already having started to trade when he placed the first Large Order).  Mr Sheth’s review of the cash trades led him to suggest that the Large Orders may have been placed to test the theory against Banca Carige, who had submitted a sell-side RFQ for €25m ten-year bonds which had been won by Mr Urra.

527. In F151 the first Large Order did overlap with the Small Order, and the stack at the time that Large Order was placed shows about 15 lots visible at the touch on the sell-side, when Mr Lopez still had 73 lots remaining of his Small Order.  Mr Shivji submitted that Mr Lopez’s Small Order was a hedge for Mr Urra’s book and submitted that the market had been rallying significantly and Mr Sheth’s orders were trying to cool the market down.  Mr Sheth’s response in cross-examination was that it was preposterous to suggest that the two Large Orders would have caused the market to cool down by 20 cents.  We agree with Mr Sheth that, whilst the remaining activity in this Instance meets the Criteria and fits with the pattern relied upon by the Authority, the second Large Order, which is relied upon as a misleading order, does not overlap with an order on the opposite side of the book, and there was no evidence in support of Mr Shivji’s submission that the placing of this order may have been to cool the market in case there was more to do.    

Amendment of price of Large Orders

528. Both Mr Urra and Mr Sheth amended the price of their Large Orders during the course of the Instances:

(1)        Mr Sheth was generally amending the price to bring his Large Order closer to the touch (separate from his activity in other Instances where he placed new Large Orders at a price closer to the touch).  This could be seen in, eg, F201.

(2)        Mr Urra made the most price amendments, and would not only move the Large Order closer to the touch, but made multiple price amendments, towards the touch and away again, and sometimes back again.  This could be seen in, eg, F31 and F47.   

529. The Authority challenged the price amendments by Mr Urra in Instances where he amended his price multiple times, towards the touch and away again, submitting that this made no sense as he had already tested whether there was a counterparty at the levels to which he was amending.   

530. The Authority’s position was that these price amendments were intended to apply pressure to the order book by bringing the Large Order closer to the touch.  In particular, the speed of the price adjustments made by Mr Urra allowed no time for thought or reflection, and it was not about getting an attractive price for the order.

531. The Tribunal does not consider that the mere act of amending the price is inconsistent with the pursuit of an Information Discovery Strategy.  We do, however, consider that the adjustments by Mr Urra which are made at pace, teasing towards the touch, away, and then back again give the appearance of applying pressure to push the market away whilst ensuring that the Large Order does not remain at the touch long enough to trade.  Mr Sheth had not used this approach in the Specified Instances.

Reduced use of strategy over the Relevant Period

532. Mr Urra said he became less enthusiastic about the Information Discovery Strategy and decreased the number of times he used it as the Relevant Period went on.  Mr Sheth’s evidence was similarly that he had tried the strategy intensively over three weeks in June 2016, with further attempts in July.  He believed it could work, but lost interest as it was not working.

533. The majority of the Instances relied upon by the Authority against Mr Urra and Mr Sheth were in June 2016.  Mr George and Mr Bailin both submitted that this pattern is consistent with a trading strategy that proved ineffective in June and was used less frequently in July; it was inconsistent with a market abuse scheme that supposedly worked well in June but was then used less in July.

534. The Tribunal has found that 179 of the 233 Instances occurred in June 2016 and the remaining 54 in July 2016.  The evidence in relation to the activity of Mr Urra and Mr Sheth was not presented on an equivalent basis - 74.2% of the Instances on which the Authority relied against Mr Urra were in June 2016, with 25.8% in July 2016 (and these involved Mr Urra placing 96 and 39 Large Orders respectively); Mr Sheth placed 184 out of 202 large orders (this including both Large Orders and Non-Instance large orders) in June 2016 (with the first Large Order placed by Mr Sheth being in F36 on 13 June 2016), although the split between the number of Instances relied upon against him is less stark.

535. Mr George referred to the split between Electronic RFQs and cash bond trades, whereas Mr Shivji drew attention to the total number of traded lots each month.  The data included:

 

June 2016

July 2016

% by which June exceeds July

Number of Electronic RFQs received by Desk for BTPs

3377

2660

26.9%

Number of executed BTP trades by Desk

849

632

34.3%

Total traded lots on Exchange

3,399,472

1,567,151

116.9%

 

536. It can be seen that there was a clear difference between the level of market activity in June and July.  Mr Creaturo reminded the Tribunal that June is half-year end and very active, whereas July is the beginning of the summer period and the number of trades and volumes drop off at this time.

537. The Tribunal is not persuaded by Mr George’s and Mr Bailin’s submissions that the occurrence of these Instances supports a conclusion this was a legitimate strategy that was not working, and considers the position to be neutral.  There was a reduction in all activity in July 2016 compared to June 2016, whether in the market as a whole or for the Desk, and there are still Instances throughout July 2016 involving Mr Urra and/or Mr Sheth (with the focus on whether Mr Sheth placed Large Orders in those Instances being somewhat distortive, as it is the Authority’s case that the Traders were involved in an Instance if they had placed the Small Order that was opposite the Large Order), and looking at the list of Instances relied upon by the Authority it is clear that all three Traders were involved in the activity even in the last week in July, up to and including 28 July 2016.

Conclusions on the Information Discovery Strategy

538. The Tribunal is not persuaded that such a strategy, even if it were being pursued, explains the pattern of trading activity by Mr Urra or Mr Sheth within the Instance Pool.  The Tribunal accepts that both plausibility and the (agreed) lack of success of the strategy are not necessarily determinative, and the burden of proof is on the Authority.  However, they do raise the question as to the likelihood that either or both of Mr Urra and Mr Sheth were in fact pursuing such a strategy. 

Anticipatory Hedging Strategy - plausibility

539. We have set out Mr Lopez’s explanation of the Anticipatory Hedging Strategy above. 

540. In his witness statement Mr Lopez had sought to use the information that is available to reconstruct his trading activity in the Instances (setting out where his Small Orders were hedges for cash bond transactions or adjustments to his curve position, the cash bonds he bought on his own initiative rather than in response to an RFQ, and his opinion on market movements at particular times by reference to the price graphs).  Both in his witness statement and in cross-examination Mr Lopez referred to the information that is no longer available, with his evidence being that his decisions on the placing of large orders (both the Large Orders and the Non-Instance large orders), and the timing of cancellation, would have been a result of the information that is and is not available, including other market events or market movements.  We record that, in contrast to Mr Urra and Mr Sheth, Mr Lopez did not say that he was reducing the use of his strategy - he said he was being meticulous and trying to trade, but was not (as a result of the Eurex Letter and the instruction given to the Desk not to place orders away from the touch) given time to pursue his strategy.

541. The Authority submitted that the Large Orders placed by Mr Lopez were not placed in pursuance of a pre-positioning or anticipatory hedging strategy.  Mr Shivji submitted that Mr Lopez has reverse-engineered a purported rationale and presented it as a recollection (in his interview with HR and initially in the interview with the Authority) and now as a reconstruction. 

Use of terminology of pre-positioning and anticipatory hedging

542. The Authority addressed the use of terminology by the Traders and Mr Kasapis:

(1)        “Pre-hedging” was described as being a hedge in anticipation of a specific client trade, “knowing that there is a high likelihood that the trade will execute”.  As Mr Urra’s Reply went on to describe it: “Pre-hedging in these circumstances involves taking a position before the client order is executed to mitigate the risks associated with execution. Hedging in advance aims to protect the trader from market volatility that may occur whilst the client order is being executed.”  

(2)        The wider concept of “anticipatory hedging” is not linked to a client enquiry or trade, but involves “taking a position in a security or derivative instrument that is expected to offset the potential losses that may arise from market movements”.  Mr Urra’s Reply gave the example of a trader holding inventory and anticipating a market-wide decrease in BTP prices due to changes in interest rates, and taking a short (sell) position in Futures to offset the risk.  

543. The different terminology was used by the Traders at different times, and there was sometimes a degree of inconsistency.  The Tribunal accepts that “pre-hedging”, or “pre-positioning” (the term sometimes used by Mr Creaturo), is a hedge in anticipation of a specific client trade, and it is a form of “anticipatory hedging”.  The latter term can be used in a broader sense, whether the prospect of an incoming client order or anticipated market movements based on market intelligence. 

544. Mr Lopez’s description of his strategy was one of anticipatory hedging.  The Tribunal accepts that a strategy of anticipatory hedging can be a legitimate trading strategy.  Mr Jaffey referred to Mr Heiberg having accepted in his interview with the Authority that anticipatory hedging is the only way where you can have a competitive desk, referring to traders having a portfolio ready to sell to clients.  However, Mr Heiberg was being asked about trading cash bonds off the back of client enquiries, where he was making clear that the Desk may also be acquiring bonds that they anticipated they could sell to a client, and the need to try and predict what may be interesting for an asset manager.  He was not talking about acquiring Futures positions in this context.

545. The Desk’s Mandate permits anticipatory hedging, but sets out the meaning which applies for that purpose, providing that transactions in hedging instruments must not be conducted where they increase the Desk’s market risk exposures, other than “in anticipation of a highly likely near term exposure to risk, where there is a sound risk management rationale for such anticipatory hedging”.  We address separately whether we consider that the Anticipatory Hedging Strategy was within the scope of anticipatory hedging as permitted by the Mandate.

Presentation of evidence by Mr Lopez

546. The Authority criticised Mr Lopez for his presentation of evidence in various ways:

(1)        Following the Eurex Letter Mr Lopez said he did not go back to look at his trading for occasions when the large orders overlapped with small orders.

(2)        As referred to at [116] above, in his email to HR of 19 October 2016 he had sought to offer an explanation for F174, saying he can provide “some details to the best of what I can remember”, ie presenting this as recollection and identifying an RFQ for a client selling bonds.  But this RFQ was not received until later and could not have been the originating thought-process.  This was pointed out to Mr Lopez in his interview with the Authority and he has subsequently identified different RFQs.  The Authority submitted that this shows his efforts to find evidence to construct his case, rather than an honest account of what happened.

(3)        Mr Lopez had sought to reconstruct the rationale for his specific trading activity in his witness statement.

547. We do agree that Mr Lopez has had every opportunity to obtain the evidence on which he wishes to rely in relation to F174.  To the extent that the explanations now provided in relation to F174 are inadequate or unsupported by any evidence, this would seem to be a substantive deficiency in Mr Lopez’s explanation.  The difficulty for Mr Lopez is that, having made a mistake about the timing of the RFQs which had been received by the Desk, which mistake he made after the Eurex Letter was received, or certainly in advance of the disciplinary interview with HR in October 2016, that mistake was not identified until January 2018, by which point it would, as with all of the Instances, have been harder for Mr Lopez to find supporting evidence beyond the details of cash bond transactions and Electronic RFQs.

548. However, to the extent that the Authority makes a wider challenge about reconstruction based on available evidence, we consider such challenge is unfair (and this applies equally in respect of Mr Sheth’s reconstructions).  It is the Authority’s case that all of the reconstructions are fabricated in any event; but the relevant point is that it would be impossible for the Traders now, and we consider even in October 2016, to be able to explain for any particular order they had placed the events that were happening at that time, whether market movements, RFQs or otherwise.  The only exception would be if the event in question were a one-off, such that it might have an unusual background or explanation which they could recall (even then potentially still without details).  The reconstruction is an attempt to demonstrate how decisions would have been taken; the question is then for the Tribunal whether we accept that explanation.

Responsibility for increasing success rate in medium-sized RFQs

549. Mr Lopez’s evidence was that he had been recruited to focus on increasing MHI’s success rate in winning medium-sized RFQs, ie cash bonds in the range of €20-30m.  Mr Shivji challenged this in cross-examination, referring to the evidence showing that both Mr Urra and Mr Lopez had responded to such RFQs in the Relevant Period.

550. We have found that Mr Lopez had been recruited with a view to focusing on the medium-sized tickets.  Mr Heiberg had repeatedly referred in his interview with the Authority to wanting Mr Urra and Mr Hill to handle the large tickets, with Mr Lopez and Mr Barouti then focusing on the medium-sized tickets, and having a clear division of roles and responsibilities.  The context in which he explained this was in large part that of a challenge to the operation of the Desk at the time and Mr Urra’s approach to supervision and allocation of responsibilities.

551. We nevertheless accept that Mr Urra and Mr Lopez had both responded to RFQs of this size during the Relevant Period.

Placing of anticipatory hedges at a beneficial price

552. The Tribunal accepts that in principle  it is legitimate for a market maker to try to identify the range within which products will trade and to try to buy or sell Futures at a beneficial price based on the range they have identified.   

 Approach to increasing the hit ratio and winning these RFQs

553. Mr Lopez explained that he had identified predictable patterns of RFQs from key potential counterparties, and was seeking to pre-position himself to offer competitive prices and win this business.  Mr Jaffey submitted that there were clear patterns, referring to Banca d’Italia (which sent an RFQ most days for MHI to sell €25m of bonds within a defined maturity range, often June 2026) and Banco Popular.  These client orders would have required hedging with 200 to 400 lots. 

554. The Authority submitted that Mr Lopez’s alleged strategy for increasing the hit ratio was illogical.  It was speculative and risky, and inconsistent with MHI’s position as a third-tier market-maker.  Rather than improve Mr Lopez’s chances of winning client business, if one of these Large Orders had executed, there was every chance it would prejudice MHI from obtaining business by locking MHI into a directional position which meant it was not then able to stand ready to quote prices on both sides of the market in response to any client enquiries.  It was, the Authority submitted, accordingly, unlikely that Mr Lopez was pursuing this strategy.

555. Mr Shivji submitted that this strategy was more likely to undermine his ability to win more business than it was to advance it:

(1)        even as a matter of theory it would increase his directional risk, contrary to a market maker’s usual preference for staying flat.  This would render him competitive only in one direction (even assuming the market had not moved against him) and lock him into that position for a material period of time.

(2)        The strategy was flawed in any event: even if Mr Lopez could somehow predict the direction in which he should position himself, by placing large, uniceberged orders on the Exchange, he was more likely to push the market away from him than he was to get his Large Order done.

556. The Tribunal has the list of the 93 RFQs and the evidence of Mr Kasapis and Mr Creaturo (in respect of which Mr Jaffey submitted that Mr Creaturo had accepted that these RFQs were from regular clients with predictable size, direction and duration).  Whilst the Authority’s submissions on the 93 RFQs were primarily focused on whether they could support a conclusion that they were “highly likely” for the purposes of the Mandate, the Tribunal has considered these submissions in the context of Mr Lopez’s evidence that he was predicting patterns and seeking to win this business and thus increase MHI’s hit ratio, alongside the Authority’s submissions as to the risks posed by the Large Orders and whether such a strategy could be effective.

93 RFQs and seeking to win this business

557. Mr Lopez was responsible for seeking to win this business in a very competitive market and with an information disadvantage compared to larger market makers.  MHI only saw a small portion of medium-sized RFQs which were being sent by clients to market makers, and had rarely been successful in winning medium-sized RFQs.

558. Mr Creaturo had expressed his opinion in his report that the Anticipatory Hedging Strategy described by Mr Lopez appeared highly speculative and unsupported by the Desk’s usual client flow.  Mr Jaffey’s cross-examination of Mr Creaturo focused on this opinion and the 93 RFQs.

559. The Tribunal makes the following findings in relation to the 93 RFQs:

(1)        Banca d’Italia sent 47 of these RFQs, all of which were for MHI to sell, they were received throughout the Relevant Period, and on most (but not all) of the trading days;

(2)        Banco Popular sent 15 of these RFQs, 13 of which were for MHI to sell, two for MHI to buy;

(3)        71 of the RFQs were for MHI to sell to the client, with the remaining 22 being for MHI to buy; and

(4)        these RFQs would all require hedging with more than 200 lots of Futures. 

560. Mr Jaffey put it to Mr Creaturo that there was a predictable pattern (of size, direction and duration), particularly from Banca d’Italia and that Mr Lopez would want to put himself in a position to provide a competitive price for these trades.  Mr Creaturo accepted that he could see this backdrop to the client flows and that he could understand the concept of anticipatory hedging in this regard to try to get ahead of these flows, anticipate the next client trade and be properly positioned to meet it (although he did not accept that this was “pre-positioning” as that is generally with a specific order in mind).  However, Mr Creaturo made clear that this acceptance of a predictable pattern was subject to caveats:

(1)        Being asked about the predictability of the RFQs being sent by Banca d’Italia, Mr Creaturo immediately asked about the times of day at which they were sent.  Giving evidence, Mr Creaturo repeatedly emphasised that we don’t have (on the list of 93 RFQs) the timing of the trades (and the Tribunal records that he wasn’t taken to the timing at any point by Mr Jaffey).

(2)        He said he was trying to understand why it was, if these RFQs were predictable and Mr Lopez had identified a support and resistance level, there was never a trade, why Mr Lopez never took duration, even of smaller size.

(3)        As well as seeing no trade, he saw the cancellation time on the Large Orders coinciding with the Small Order being filled.

561. The Tribunal did have information as to the timing of these RFQs on the full list of Electronic RFQs received during the Relevant Period, but it was not set out on the list of 93 RFQs itself to which Mr Creaturo was taken during cross-examination.  The Tribunal has identified that the RFQs sent by Banca d’Italia were generally received between 11.00 and 13.00.  There were occasional outliers, eg at 10.10.53, 10.59.31, 13.24.04 and 13.33.21, but this range was apparent from the beginning of the Relevant Period, which would have been the data that Mr Lopez had available to him when he was starting to consider whether there were patterns.

562. Based on this evidence, the Tribunal finds that the RFQs the Desk received from Banca d’Italia were predictable in size, duration and direction, and that Banca d’Italia sent RFQs to MHI most days (and sometimes two or three times in a day), but there was a wide variation in the timing of these.  We accept Mr Creaturo’s evidence that it is the timing of the next client trade that might justify the risk of taking a directional position, but each day Mr Lopez could not know if the next client trade is coming in five minutes, 15 minutes or an hour.  There was no data supporting imminent client activity at any particular point in time, and in this situation it would be taking a speculative position in the market to take directional risk based on the possibility of these one, two or three RFQs per day.

563. The RFQs from Banca d’Italia represented about half of the 93 RFQs.  Whilst Banco Popular also regularly sent RFQs of this size, the Tribunal is not satisfied that the fact that RFQs were received of this size nearly every day (none were received on 10 June 2016) of itself meant that it could be said there was any predictable pattern overall, particularly in a situation where not only the timing of the receipt could not be known but more importantly the direction of the potential client trade.

Directional risk and remaining competitive

564. A cash BTP market maker is required to stand ready to purchase and sell cash BTPs and quote in both directions competitively.  The most effective way to do this is to minimise (not significantly exaggerate) the directional risk of the book; this also protects from significant moves in the market, better enabling the market maker to quote whatever the current market conditions.

565. The Tribunal finds that Mr Lopez managed the risk on his book (both directional risk and curve risk) very carefully throughout each day.  Mr Lopez had described his approach in this way, but this is also borne out by his trading activity.  He placed 1,139 small orders over the Relevant Period, 103 of which were for one lot and 954 of which were for one to 19 lots.  If one of the Large Orders traded, this would have increased dramatically Mr Lopez’s directional risk.  This would have been at times when he was placing Small Orders for curve management purposes, eg in F56 and F63 (both of which involved a Small Order to sell five lots, iceberged to three, and a Large Order to buy 200 lots) and F132 (a Small Order to buy one lot and a Large Order to sell 200 lots).

566. Taking directional risk in this way would only have assisted him in quoting for client flows in one direction if, having put on the directional risk, the market then moved in his favour.  This positioning would have served to make him significantly less competitive in the other direction.  Yet Mr Lopez could not know in which direction he would be required to quote in the immediate future:

(1)        Client enquiries could come in at any time, for any size or duration, or for any amounts.  This is shown by the list of all Electronic RFQs.  

(2)        MHI was a lower tier market maker with little visibility over client flow and an information disadvantage.  Mr Kasapis accepted that only market makers with deep institutional order books would have a good idea of client flow.  Mr Lopez recognised this (sometimes using it as an explanation for ignoring the direction of client flow).  Mr Creaturo’s opinion was that other factors cited by Mr Lopez in his witness statement, such as comparing the Bund/BTP Futures spread and anticipating liquidity from auctions, were “ultimately speculative”; a bank such as MHI could not reliably predict whether client orders would materialise at all, let alone of the size and direction of Mr Lopez’s Large Orders. 

(3)        The list of 93 RFQs, and the repeated buying interest from Banca d’Italia, were not as predictable as Mr Jaffey submitted, and the uncertainty over timing, and indeed whether they would come at all, did not justify this directional risk.  Positioning the entire book with significant directional risk for the sake of a narrow set of business and one client’s RFQs (and thereby rendering his pricing uncompetitive for other business) would be irrational for a market maker.  It would not only likely result in a lower hit ratio overall (contrary to what he was trying to achieve) but if medium-sized tickets came in from clients in the other direction, Mr Lopez had effectively counted himself out of winning them.  It made no sense for MHI to be able to quote an occasional good price for certain trades (if the enquiry came in for the right bond at the right time in the right size and the right direction) whilst being uncompetitive for all trades in the other direction. 

(4)        Even if Mr Lopez did correctly predict the direction of client flow, it was speculation whether that would in fact enable him to quote competitively for those clients.  The timing of both the market movements and any client enquiries were unpredictable.  The market could well move against him, such that (for example) he could have bought at a far higher price than he needed to.  This could be seen in, eg, F42, where the market moved 20 ticks against him from the time he placed the Large Order to the time the first Banca d’Italia RFQ came in, with the result that the anticipatory hedge would have cost €40,000 more than a hedge taken out at the time of the RFQ.

(5)        MHI had a very low hit ratio in respect of these medium-sized RFQs.  Of the 93 RFQs, MHI responded to about half of them, and the Desk won seven of them (Mr Urra won six and Mr Sheth won one).  Mr Urra did win three of the RFQs from Banca d’Italia.    Mr Lopez may have been seeking to improve this position, but he would have been taking a risk for what is a speculative prospect of winning the cash trade in the situation where MHI’s record was poor, and where the timing of that cash trade was uncertain.

567. We have found that the average size of Electronic RFQs for cash BTPs received by MHI during the Relevant Period was €5.1m.  These 93 RFQs to which Mr Lopez refers, and which he relied on heavily as explaining his Anticipatory Hedging Strategy, were a very small portion of the total number of Electronic RFQs received by the Desk in the Relevant Period (of which there were 6,972 in total).  There would, in addition, have been Voice and Bloomberg/Chat RFQs, which we approximated as a further 2,500 RFQs.  Mr Lopez responded to 4,100 RFQs for BTPs during the Relevant Period, of which 104 were valued from €15-25m.  Of these 104, 42 of them were for bonds with maturities of less than five years which would not typically be hedged with Futures.  Mr Lopez only won ten of these 104 RFQs, and only three of the ten had a remaining maturity greater than five years, ie would justify being hedged by Futures.

568. The Tribunal considers that whilst in principle an Anticipatory Hedging Strategy may help with positioning the Desk to meet expected client flow, the reality is that Mr Lopez was not regularly trading cash bonds with clients in the size that would justify anticipatory hedging in the size of the Large Orders, and the level of client demand was not sufficiently predictable to explain or justify taking this amount of directional risk.

Whether placing of large, uniceberged, orders was less likely to achieve Mr Lopez’s aims

569. Mr Shivji submitted that placing large orders away from the touch in full visible size and assuming or hoping that the market comes to you is fundamentally flawed logic.  All other things being equal, the most likely effect of Mr Lopez placing a Large Order close to but offset from the touch would be to move the market away from him.

570. The key point in the Authority’s submission in this context is that Mr Lopez did not iceberg his Large Orders.  This contrasts with Mr Lopez’s trading style in respect of the Small Orders, where he would iceberg the Small Orders even for, eg, five lots which he iceberged to three in F56.

571. There are two different issues which arise in this context:

(1)        the impact of orders of visible size of 200 lots or more on the market; and

(2)        Mr Lopez’s evidence in relation to his reason for not iceberging the Large Orders.

572. The Tribunal has set out its conclusions on market impact at [393] above, which included that the Large Orders were each likely to give the impression or signal of significantly increased supply or demand, and that the most likely market reaction would be for the market to move in the opposite direction, away from the Large Orders.  This conclusion applies to Mr Lopez’s Large Orders of 200 lots, which did create a volume skew in the Specified Instances.    

573. Mr Lopez’s evidence was that iceberging would slow down the execution of an order, as each slice that is entered as the order trades will be placed at the back of the queue at that price point.  He would typically iceberg his small hedging orders as he was happy for the order to execute slowly, as this provided an opportunity for a client cash order to come the other way (which may have repositioned his book or altered his hedging needs) and he could amend the price of the hedge as it executed.  Mr Jaffey also submitted that latency in MHI’s computer systems could further delay the additional slices being sent to the queue.

574. Whilst the Authority accepted Mr Lopez’s description of the mechanics of how the slices of an iceberged order are entered, Mr Shivji submitted that this proposition was turning the mechanics of placing an iceberged order (which was agreed) into its rationale, and that an iceberged order could execute very quickly. 

575. Mr Kasapis and Mr Creaturo both gave evidence on the rationale for iceberging an order, the impact on visible liquidity, expectations of market participants and execution.

576. Mr Creaturo’s evidence was that the advantage of iceberging is that this prevents other market participants from seeing the full size of the order and being able to take advantage of this information, resulting in the market moving in the opposite direction from the order.  He said you iceberg orders so they fit with general liquidity at the time you wish to trade so it does not stand out.  We accept that this is an advantage of iceberging orders and a reason that traders would iceberg an order. 

577. Of itself, this does not necessarily preclude there being additional benefits of iceberging an order, and Mr Creaturo accepted that there can be an “element of optionality” to iceberging, allowing you to make adjustments to the order.

578. The Tribunal finds:

(1)        There is no evidence before us to conclude that any latency in MHI’s systems means that the process of re-submitting orders as each slice trades would be slow.  There was no direct evidence before us as to the speed of MHI’s trading systems in the Relevant Period. 

(2)        The data on trading activity in the Instance Pool did show that slices of an iceberged order could be placed on the market and trade quickly, eg in F62, F63 and F190.  In F190, Mr Lopez was selling 45 lots, iceberged to six lots.  He had amended the price at 14.14.18.919 and it started to trade, filling at 14.14.20.501, ie within about 1.6 seconds.    

(3)        The Traders traded manually, so would need to amend orders manually (albeit with a single click).  If an iceberged order did in fact trade slowly, the trader would be able to amend the price. 

(4)        A trader placing an iceberged order would not know if their order would execute quickly or slowly once it started to trade.  (This also applies to orders that were not iceberged.)  They would know, if they placed an order away from the touch, that it would not start to trade until the market had moved through the stack towards it. 

579. On this basis, we do accept in principle that iceberging an order may result in it trading more slowly than an uniceberged order; and that avoiding this may be a reason not to iceberg an order.

Anticipatory hedging under the Mandate

580. The Authority submitted that Mr Lopez’s alleged strategy would have placed him in breach of the Mandate.  His strategy did not involve hedging against a “highly likely near term exposure to risk”, nor was it underpinned by a “sound risk management rationale”. The Authority submitted that this, too, made it unlikely that Mr Lopez was pursuing this strategy.

581. Mr Jaffey submitted that the Anticipatory Hedging Strategy is expressly recognised as legitimate in the Mandate; and the MHI Compliance Report had concluded that Mr Lopez’s trading was not in breach of the Mandate.

582. Mr Lopez did not seek the evaluation and approval of the Anticipatory Hedging Strategy by Mr Urra or Mr Heiberg in accordance with the requirements of the Mandate.  Mr Lopez’s evidence was that his strategy was “completely common” and “entirely normal practice”.  Whilst the Tribunal accepts that anticipatory hedging is often undertaken by market makers, the difference is that Mr Lopez was, on his own account, placing orders of 200 lots or more to seek to position himself for RFQs that represented only a small portion of the Desk’s business.  We would expect that, at the very least, he would have discussed with Mr Urra to confirm that his proposed strategy did not require approval under the Mandate.

583. The Mandate defines what it means by “anticipatory hedging” for its purposes.  Anticipatory hedging was only permitted under the terms of the Mandate where there was a “highly likely near term exposure to risk” and where there was “a sound risk management rationale for such anticipatory hedging”.  We do not place any weight on the conclusions expressed in the MHI Compliance Report in this regard.  Whilst Compliance had interviewed Mr Lopez, the report seems only to have identified the explanations which had been provided by Mr Urra and does not expressly record that Mr Lopez said he was pursuing a different strategy.  In addition, the conclusion reached was expressed as “to the extent that these were placed with a view to establishing a hedge against expected near term customer trades or to establish a basis position for future sale to clients, would fall into the category of “anticipatory hedging”…”, ie by prefacing with “to the extent that…”, it assumes the required elements are met.

584. We accept that the phrase “highly likely near term exposure to risk” is not defined or explained further by the Mandate, and that it does not require that the exposure be highly likely to materialise within any specified period of time.  However, based on the conclusions reached at [557] to [568] above in relation to the 93 RFQs and directional risk, the Tribunal is not persuaded that there was a “highly likely near term exposure to risk”.  Mr Lopez was not regularly trading cash bonds with clients in the size that would justify anticipatory hedging in the size of the Large Orders, and the level of client demand was not sufficiently predictable to explain or justify taking this amount of directional risk.  We accept Mr Creaturo’s opinion that hedging in these circumstances was “purely speculative”.

585. For the same reasons, we have significant doubts as to whether there could be said to be a “sound risk management rationale” for the Anticipatory Hedging Strategy.  The Large Orders were not being placed for risk management reasons; even on Mr Lopez’s account they were being placed so that he could obtain anticipatory hedges at beneficial prices and offer more competitive prices to RFQs once received. 

586. However, we do take account of Compliance’s apparent focus on the Desk’s limits and accept that each of Mr Lopez’s Large Orders were of a level where they could be held overnight without breaching the Desk’s limits, and Mr Lopez did not place concurrent Large Orders.  Mr Shivji did refer us to F31 where Mr Lopez placed a Large Order of 400 lots very close to the end of the day, at 17.46.21.136, a time at which Mr Urra also had a Large Order of 499 lots on the market, and if they had both traded this would have taken them over the Desk’s PV01 limit.  That is correct as a factual matter, but not relevant when assessing, on a standalone basis, the Anticipatory Hedging Strategy as put forward by Mr Lopez.

587. The Tribunal therefore finds that the Anticipatory Hedging Strategy would result in an increase in market risk and was not within the exception for anticipatory hedging as set out for the purposes of the Mandate. 

Conclusions on plausibility

588. Anticipatory hedging can be a legitimate trading strategy.  However, the Anticipatory Hedging Strategy as explained by Mr Lopez involves taking directional risk based on what Mr Lopez says were predictable patterns of RFQs in circumstances where those RFQs were a small part of the Desk’s market making business and where, although they were fairly regular, they were not predictable in timing.  The result would have been to take directional risk, albeit of a size that would not, viewing each Large Order on its own, breach the Desk’s limits, where Mr Lopez’s trading activity was otherwise very precise, with adjustments frequently being made by a single lot order.  As a strategy it is at odds with both MHI’s business and Mr Lopez’s own approach to managing the risk on his book.  These features mean that whilst the Tribunal considers that the Anticipatory Hedging Strategy could in theory be plausible, there are reasons to doubt that Mr Lopez would have been pursuing such a strategy.

Anticipatory Hedging Strategy - operation by Mr Lopez

589. The Authority submitted that Mr Lopez’s behaviour across the Relevant Period is inconsistent with the Anticipatory Hedging Strategy being his true purpose, or is otherwise unexplained by it.  We have considered this in detail in the light of all the evidence of the trading activity of Mr Lopez in the Specified Instances, and his other activity to which we were referred (both within and outside the Instance Pool).

Speculative nature of anticipatory hedge orders

590. The Authority contrasted Mr Lopez’s otherwise diligent and cautious approach, being someone who managed his risk very carefully, with what they described as the speculative anticipatory hedges of 200 lots or more on the market, with no certainty over what client flow or the market would do next. 

591. We do not see this point as being anything more than a challenge to whether the strategy existed.  The examples relied upon by the Authority - F56, F63 and F132 (where the Small Orders were for five lots in the first two Instances and one lot in the third) do illustrate a submission made by Mr Jaffey namely that on the Authority’s case Mr Lopez was pursuing an abusive strategy in circumstances where the amounts at stake in the Small Orders, and thus to be gained, are very small, as the spread of one tick meant it would cost an additional €10 per tick for each lot to cross the spread and trade.

Timing of placing the Large Orders

592. The Authority submitted that Mr Lopez placed Large Orders at illiquid or potentially volatile times of the day, when he was (i) unlikely to get any client interest, and/or (ii) liable to be stuck with a risky position overnight or come close to his limits if the Large Order filled.

593. We agree with the Authority that Mr Lopez cannot realistically have anticipated imminent client interest at the time he placed some of his Large Orders.  Liquidity dropped dramatically towards the end of the day.  Mr Creaturo explained in his evidence that very few client orders or event driven trades take place after the official closing at 17.30.  Mr Kasapis also regarded this sort of time as one of the most illiquid parts of the day.  This could be seen from the volumes on the stack in the Replay graphs. 

594. Looking at all 93 RFQs, only three of them arrived after 14.00, and only 14 of them arrived after 13.00.  None of them were received after 17.00.  Across the Relevant Period, the Desk only did four cash market trades with a 2021 maturity or higher later than 17.30, and only two of these occurred on a Friday. 

595. In F31 Mr Lopez placed a Large Order to buy 400 lots at 17:46.21.136 on Friday 10 June 2016, which would have required an RFQ of close to €40 million ten-year BTPs.  He described this as being placed to “service client buy orders that I expected to receive in the immediate future”, but we agree with the Authority that he cannot reasonably have expected any client interest before Monday morning.    Mr Lopez also placed the following Large Orders late in the day:

(1)        F56 - he placed a Large Order to buy 200 lots at 17.02.21.324 on 15 June 2016;

(2)        F132 - he placed a Large Order to sell 200 lots at 17.07.53.566 on 23 June 2016; and

(3)        F177 - he placed a Large Order to buy 200 lots at 17.21.37.682 on 29 June 2016.

596. F132 took place on the date of the Brexit referendum, and Mr Lopez explained his Large Order as being that he thought other market participants would seek to square their positions ahead of the vote, ie de-risk their books.  His own apparent decision to take this level of risk at the end of the day is not only inconsistent with his own approach to managing risk on his book, but with what he says he was expecting other market makers to do.

597. The Authority submitted that placing these Large Orders at this time of day was also exposing him to considerable directional risk at the worst time of day, and said that if the Large Orders in F31 and F56 had traded and he held the position overnight he would have lost €112,000 and €60,000 respectively.  We accept this point on directional risk; but recognise that the size of these orders were each within the Desk’s limits.  

None of the Large Orders traded

598. The Authority submitted that Mr Lopez’s failure to achieve any sort of anticipatory hedge in this period is inconsistent with a genuine desire to do so.  On his case, he placed 88 large orders over the Relevant Period pursuant to this strategy and none of them traded.  55 of these were Large Orders within the Instance Pool, and five of them were Lone Large Orders.

599. Mr Creaturo regarded this as a significant point in determining whether Mr Lopez’s trading activity was consistent with his alleged Anticipatory Hedging Strategy.  He said that if he was managing a trader in that position, he would have encouraged them to be more aggressive - iceberg the order, try smaller, nudge the order closer to the touch, keep the order on for longer - and put the trade on.  Mr Jaffey put it to Mr Creaturo that such an approach would have involved greater risk, but Mr Creaturo’s answer to this was that Mr Lopez’s starting point was that he was at the support level he had identified.

600. The Tribunal accepts Mr Creaturo’s evidence that Mr Lopez’s failure to execute any of the Large Orders (or any of the other large orders he placed) is significant.  We recognise that the Large Orders were placed away from the touch and therefore less likely to execute.  However, as referred to by Mr Creaturo, Mr Lopez’s evidence was that he was carefully analysing all available market information, including the Bund/Futures spread and pricing levels in the Futures, trying to identify support and resistance levels, yet took an “all or nothing” approach to his Large Orders.  He explained why he did not iceberg them; but when they were repeatedly not executing, he not only did not adjust the price (which may in particular Instances be explained by his determinations on support and resistance levels) but did not place an order in a smaller size, or place an order of 200 lots but iceberged to, say, 25, to see if he could get any of it filled.  We note in this regard that Mr Lopez did place some Large Orders for more than 200 lots, generally earlier in the Instance Pool - it was no part of his evidence that the reduction in size was to try to get the orders done; instead he referred to his greater risk appetite at the beginning of June and that 200 lots was sufficient size for the medium-sized RFQs.

601. Furthermore, the outcome of none of these Large Orders trading contrasts with his approach elsewhere, in particular his decisions as to purchases of cash bonds on his own initiative: 

(1)        In the context of F74, Mr Lopez explained that he had bought €2m March 2030 bonds from Credit Agricole on MTS, a transaction which he executed as he wanted to make a small adjustment to the position of his book.  This was an anticipatory position in that he anticipated he would be able to sell the bonds later at a better price, in line with the client buying view he held at that time.  

(2)        In F174 the Small Order was placed to hedge Mr Lopez’s purchase of €5m May 2023 bonds from MPS Capital Services Banca per le Imprese SpA (“MPS Capital Services”), a transaction which he executed on his own initiative as he believed the market would continue its upwards trend and there was going to be client buying interest.     

602. Mr Lopez was thus actively taking positions, albeit in much smaller sizes, based on his views of client flow at the relevant time; yet whilst relying on the same conclusions as to client buying interest as the reason for placing his Large Orders he did not execute any of them.

Approach to determination of anticipated buying or selling interest

603. Mr Lopez’s approach to the factors in this thinking was often inconsistent.  Mr Lopez often relied on RFQ data to support his assertions that he had anticipated client buying or selling interest (according to the direction of the Large Order) but his approach to this was inconsistent:

(1)        If he did not receive any RFQs that would have required hedging with Futures orders of 200 lots or more, the absence of RFQs was said to justify the Large Order of 400 lots, eg F31. 

(2)        If he had received a number of RFQs in a particular direction that morning, this was said to justify his Large Order, eg F42.

(3)        For one of his Lone Large Orders, he referred to the client flow in the opposite direction but said that it was irrelevant because MHI was a lower-tier market maker and did not have a full picture.

(4)        He sometimes selected RFQs from the list of Electronic RFQs to rely upon which supported his position whilst ignoring other RFQs which did not.   In F177, he relied upon an RFQ from Banca d’Italia seeking to buy €10m bonds which had been received two hours previously.  At that time, the market had been about 25 ticks lower than the time of F177.  Moreover, at the time of the Instance, his counterparty was a seller of €10m bonds.    

604. As we have already addressed, Mr Lopez frequently cited potential buying interest from Banca d’Italia as a reason for why he had taken a view of the market.  The Authority submitted that it made no sense for Mr Lopez to be trying to position himself for a single counterparty whose business he never won (and that the Desk rarely won) whilst ignoring the prospect that other business (from other clients with whom they had a stronger relationship) might come to the Desk which he did have a good chance of winning.  We agree with this submission, particularly in circumstances where there was insufficient predictability as to the timing of Banca d’Italia coming with an RFQ.

Time for which Large Orders were live, amendments to price and cancellation decisions

605. Whilst Mr Lopez sought to set out a picture of the careful and thorough analysis he would have undertaken to determine pricing and positioning of his Large Orders, they were generally live for a short period of time, and he sometimes amended the price very quickly after placing the order, or cancelled an order on the basis of what he says was a decision that it was not going to execute but then placing another order (at a less attractive price) shortly afterwards: 

(1)        The time for which Mr Lopez’s Large Orders were live in the Single Trader Instances that were Specified Instances ranged from 3.267 seconds (in F56) to 16.269 seconds (in F190).  They were cancelled shortly after the Small Order in the opposite direction filled, with the range being from 0.58 seconds (in F177) to 5.213 seconds (in F190).

(2)        Mr Lopez had placed Large Orders in three of the Multi Trader Instances.  There was a greater variation in the length of time for which these orders were live, namely one minute 52 seconds in F31, 7.16 seconds in F174 and 50.926 seconds in F205.  However, the consistency or coincidence was in the cancellation times.  Mr Lopez cancelled his Large Order 0.995 seconds and 2.324 seconds after his Small Order filled in F174 and F205 respectively.  Mr Urra’s Small Order was only partially filled (36 out of 39 lots) in F31, but Mr Lopez cancelled his Large Order within 0.135 seconds of Mr Urra’s cancellation of his Large Order.

(3)        Where Mr Lopez amended the price of the Large Order, it was live at its new level for less than five seconds in F42, F62 and F63 (and in F63 it was cancelled within two seconds of the second amendment to the price).

(4)        The Instances show that having cancelled a Large Order, he placed a new Large Order in the same direction at a less attractive price shortly afterwards - if he had kept the first order on the market, he could have bought at that price:

(a)        F62 and F63 took place on 16 June 2016.  In F62 he placed a Large Order to buy 200 lots for 139.76 at 11.04.52.596, amending this to 139.77 about three seconds later.  He cancelled the order at 11.04.58.313, 1.582 seconds after his Small Order filled.  In F63 he placed a new Large Order at 11.13.30.597 for 139.81, increasing the price to 139.83 during the course of the Instance.  That order was live for 13.797 seconds, and was cancelled 1.073 seconds after his Small Order filled.  Less than an hour later, when hedging for Mr Urra, Mr Lopez bought at 139.96 during F64. 

(b)        F74 took place on 17 June 2016.  Mr Lopez had placed a Lone Large Order to buy 200 lots for 140.26.  That order was cancelled after three seconds.  F74 took place 40 minutes later, during the course of which he placed a Large Order to buy for 140.39, which he increased to 140.40

606. There is no consistency to the length of time that the Large Orders were live in the market.  In contrast, there does appear to be a degree of consistency between the trading of the Small Order and the cancellation of the Large Order.

Overlap with Small Orders

607. The rationale of the Anticipatory Hedging Strategy does not involve any connection with small orders that were placed in the opposite direction.  Mr Lopez’s explanation for the overlap was based on his having entered 1,139 small orders over the Relevant Period, whilst also relying on the large orders that were outside the Instance Pool.

608. Of the 88 large orders Mr Lopez placed in the Relevant Period, only five did not overlap with small orders placed by one of the three Traders.  Two of those five were placed very close to a small order in the opposite direction, including the Lone Large Order he placed on 17 June 2016 (to which he referred in his witness statement).  These Lone Large Orders are addressed further at [729] to [736] below. 

609.   We agree that a level of overlap is to be expected.  The Tribunal notes however the extent of the overlap between large orders and small orders, and, for the 55 Large Orders in the Instance Pool, the repeated coincidence in timing of cancellation being shortly after the Small Order filled.

Size of the Large Orders

610. The Authority submitted that the size of Mr Lopez’s Large Orders was inconsistent with the size of business that he was typically doing, and that had Mr Lopez genuinely been interested in anticipatory hedging, there was no reason not to apply that strategy to (much more common and likely) smaller cash trades and “get in” at the price he considered advantageous.  That was not the pattern observed: instead, Mr Lopez regarded it as 200 lots or nothing.

611. Mr Lopez only placed nine large orders of 250 lots or more.  In all ten of the Single Trader Instances that were Specified Instances, the Large Order placed by Mr Lopez was for 200 lots (although he did place Large Orders of other sizes, including a Large Order of 400 lots in F31, which is one of the Multi Trader Instances).

612. Mr Lopez’s explanation was that he was trying to win more business of larger size than the cash trades he was currently winning.  We accept that in such a situation it would be entirely logical that the size of the Large Orders would not be consistent with the size of the business he was currently winning - he was trying to do something different. 

613. We consider that this submission raises again the point made by Mr Creaturo, namely the reason for the absence of a trade.  We accept that all of the 93 RFQs would, if the Desk had won those RFQs, have resulted in a cash trade which would have required hedging with more than 200 lots of Futures.  The list does, however, show the extent to which the approximate number of lots varied significantly.  All were over 200, the range was 200 to 768 lots; and 87 of them would require 200 to 400 lots.  So the use of 200 lots as a standard size was itself not precise.  Given this, this raises the question as to why, if he wanted to take an anticipatory position, Mr Lopez did not try an order of a different, smaller, size, or iceberg his order (even if that did slow execution and result in less being done). 

Conclusions on the Anticipatory Hedging Strategy

614. We accept that Mr Lopez was seeking to win more €20-30m RFQs, and that to do this he would need to offer competitive pricing.  A strategy of anticipatory hedging would make sense, on the assumption that there is a predictable pattern of RFQs (with a predictable size, direction and time) and this would not hinder the Desk’s other activities.  In this context, we can accept that orders of 200 lots were targeted at this size of RFQ, and were of a size that were within the Desk’s risk limits.  This is a strategy which could explain not only the Large Orders in the Instance Pool, but also the Non-Instance large orders (including the Lone Large Orders). 

615. We have the following key difficulties:

(1)        The pattern of RFQs was not as predictable as Mr Lopez or Mr Jaffey tried to make out, nor as predictable as Mr Lopez needed it to be to ensure that he was not holding a directional position (with the associated risk) for a period of time which both ran the risk of going underwater and precluded him from quoting competitively in both directions in the meantime.

(2)        This cannot explain the Large Orders that were placed late in the day.

(3)        Mr Lopez’s explanation was that the timing of cancellation could be explained by reference to market movements, ie the market moved away and the desired push-back was not going to happen promptly or market news.  We recognise that he does not have access to all the information that would have been available to him at the time.  The difficulty is that Mr Lopez, an experienced trader, is saying that every single time he placed a large order the market moved against him and yet he did not change his approach to try to get some of his Futures order done as an anticipatory hedge, in circumstances where he did act to buy cash bonds to take an anticipatory position.

(4)        There is a repeated, obvious, and striking pattern of coincidence of the timing of the cancellation of the Large Orders shortly after the Small Order filled.

Placing of concurrent Large Orders

616. There are different ways in which concurrent Large Orders appear in the Instance Pool:

(1)        the placing of Multiple Large Orders by Mr Sheth, eg, in F103, F176 and F201, where the orders were for the same number of lots (usually 500) but subsequent orders were placed closer to the touch and which he said were a mistake and that they were a result of him wanting to amend the price of his existing Large Order;

(2)        Mr Urra placed concurrent Large Orders within a single Instance.  In F82 he placed two Large Orders of 450 lots each.  He agreed this was not a mistake; and this is confirmed by the way in which he made price amendments to both of them in the second half of the Instance.  Mr Urra thus had concurrent Large Orders for a total of 900 lots, and the second of those Large Orders was placed within a second of Mr Sheth placing a Large Order in the same direction for 500 lots.  Mr Urra also placed concurrent Large Orders in F59 and F220, totalling 900 lots on both occasions; and 

(3)        there were Instances where more than one Trader placed a Large Order in the same direction.  In addition to F82 above, this appeared in:

(a)        F31 where Mr Lopez’s Large Order of 400 lots overlapped with Mr Urra’s Large Order of 499 lots;

(b)        F121 where Mr Sheth placed six Large Orders of 400 lots each, not all of which overlapped, but whilst two of these orders for 400 lots were live Mr Urra placed a Large Order of 450 lots; and

(c)        F209 where Mr Urra’s Large Order of 450 lots overlapped with Mr Sheth’s Large Order of 250 lots.

617. The Tribunal considers that all three of these variations present somewhat inconsistent possibilities, even before we consider in more detail Mr Sheth’s explanation for the Multiple Large Orders. 

618. The placing of concurrent Large Orders by a single Trader (including Mr Sheth’s Multiple Large Orders) could support the Authority’s position that the Traders were applying increasing pressure to the order book according to the levels of supply and demand needed to be shown at a particular time (a submission that was also made by Mr Shivji when explaining that, contrary to Mr Jaffey’s submissions in relation to the size of Mr Lopez’s Large Orders, the Authority did not accept that an order of 200 lots would always be sufficient to facilitate the execution of a Small Order but it depended on price and the total volume of lots being placed).  However, if there was an abusive scheme, and its operation required more than one Large Order to be placed, it is not clear why these would then be placed by different Traders, ie why Mr Urra and Mr Sheth would both be placing Large Orders, rather than one of them simply placing more than one Large Order.  Yet if Mr Urra and Mr Sheth were pursuing an Information Discovery Strategy, and they say they were not coordinating or discussing their activity during an Instance, they were pursuing it rather ineptly if they both decided to test the theory of a client splitting an order at the same time without telling each other (whether before, during or afterwards).  It is notable that there are concurrent Large Orders involving all three Traders, including Mr Lopez, who on his account was pursuing a different strategy and yet was doing the same thing by placing Large Orders in the same direction and very close in time to Mr Urra and Mr Sheth.

619. The Multiple Large Orders are the overlapping Large Orders which were placed by Mr Sheth in a single Instance, eg in F103 on 21 June 2016 (which was the first date on which he placed Multiple Large Orders) he placed a Large Order to sell 500 lots for 141.14 at 16.29.57.570, cancelled it, then placed a Large Order to sell 500 lots for 141.13 at 16.31.33.702, with his next action being to place another Large Order to sell 500 lots for 141.12 at 16.31.37.776.

620. Mr Sheth’s evidence was that the Multiple Large Orders were “poor practice” and a mistake, arising as a result of his inexperience and lack of training in operating Bloomberg Escalator, and arose where he wanted to amend the price of his Large Order.  The details of his explanation are considered further below. 

621. The MHI Compliance Report had concluded that this was “poor practice”, and Mr Heiberg had described Mr Sheth’s explanation as a “plausible explanation” in his interview with the Authority. 

622. In its Amended Statement of Case for Mr Sheth at [64] the Authority’s position was “Mr Sheth has also suggested that in some cases, where he placed multiple, overlapping Large Orders, these were in fact intended to be amendments to an existing Large Order.  Whether or not this is true does not affect the Authority’s assessment of the purpose of the placement of a Large Order on the opposite side of the order book to a small order.”  At the hearing Mr Shivji set out the Authority’s position that it did not accept that these subsequent Large Orders were a mistake, or anything other than deliberate behaviour.

623. In support of Mr Sheth’s position, Mr Bailin submitted that the placing of Multiple Large Orders only served to draw attention to the trading activity during the Relevant Period.  This could be seen in, eg F176, one of the Instances questioned by the Exchange,  where Mr Sheth placed a total of seven Large Orders and at one point four of them were on the market at one tick intervals, thus showing a total of 2,000 lots on the buy-side.

624. Mr Sheth has consistently described these Multiple Large Orders as being a mistake, from the first time he was interviewed by Compliance.  He said he wanted to amend the price of his Large Order and move it closer to the touch.

625. Mr Sheth acknowledged that he did know a way to amend the price of a Large Order within Escalator, but described the relevant steps as “slow and cumbersome”, and that in a dynamic trading environment it was far quicker to place a new Large Order.  That Mr Sheth knew how to amend the price is clear from F103 itself, as Mr Sheth had amended the price of the Small Order before placing the first Large Order in that Instance.  Mr Sheth had initially explained that he forgot to cancel the original order and did not intend to have more than one Large Order in the market at any point; but at the hearing he explained that in his mind the primary focus was the order at the best price, and he considered that the market would not get to his earlier Large Orders which were further away from the touch.  His own explanation also makes it clear that he knew that he had not cancelled the first Large Order before placing the second at a better price.  The different behaviours are also illustrated in F103, where he cancelled the first Large Order before placing the second, but did not cancel the second before placing the third.  Mr Sheth’s focus on the order closest to the touch was then illustrated by his consistency in the order of cancellation of the Large Orders in the Instances involving Multiple Large Orders, as he cancelled that which was closest to the touch first.

626. The difficulties we have with the terminology of mistake are:

(1)        Mr Sheth knew the difference between amending the price of an order and placing a new order at a different price (he accepted this in evidence and this could be seen in the Instances, including in F103);

(2)        he did sometimes place a Large Order, cancel it, then place a new Large Order at a different price (eg in F94, which took place on 21 June 2016); and

(3)        he knew he had not cancelled the first order which meant that it remained live (he accepted this in evidence, explaining his focus on the order closest to the touch). 

627. The result of these actions, and even based on Mr Sheth’s own evidence, is that at the time he placed the second (and subsequent) concurrent Large Orders in any Instance (and this would be the placing of the third Large Order in F103) he no longer had an intention to trade the previous Large Orders which had been placed and not cancelled (this would be the second Large Order in F103). 

628. The Tribunal considers the issue to be better described as whether this was poor practice (rather than a mistake) or Mr Sheth deliberately increasing the number of lots on the order book to apply further pressure on the book.  We accept that placing a new order rather than amending an existing order could be the poor practice of a kind a more junior trader might make. 

629. Mr Bailin submitted that these Multiple Large Orders are potentially significant:

(1)        The first Instance involving Mr Sheth’s trading activity was on 13 June 2016, and the first Multiple Large Order did not occur until 21 June 2016.  We accept that this raises a question as to whether there was a trigger for this change in behaviour.  Mr Sheth’s evidence was that he had spoken to Mr Urra about his frustration with the strategy not working, and that Mr Urra had encouraged him to move his orders closer to the touch.  The Authority’s case includes that any explanations now about these discussions on the strategy are fabricated, as they were not pursuing a legitimate strategy, yet the Tribunal acknowledges that in cross-examination by Mr Jaffey (in the context of Mr Lopez’s trading), Mr Creaturo had said that if a trader came to him saying they wanted to trade but their orders were not trading he would encourage them to move the order closer to the touch and iceberg it.  (We recognise that these questions were put to Mr Creaturo against the background of Mr Lopez’s evidence that he was looking to take directional risk, having sought to analyse the support and resistance levels within the market.)

(2)        Mr Sheth placed these Multiple Large Orders in Multi Trader Instances (as well as in Single Trader Instances), which raises a question as to why, if these orders were a mistake or poor practice and the Traders were acting together to commit market abuse, Mr Urra and Mr Lopez did not notice these orders on the stack and check with Mr Sheth that they were not his to ensure he did not draw unwanted attention to their actions:

(a)        In F121 Mr Sheth placed a Small Order to buy 22 lots.  Mr Sheth placed six Large Orders in this Instance and Mr Urra placed one Large Order.  During that Instance there is more than one point in time where two of Mr Sheth’s Large Orders were on the market at the same time (and on the first occasion they also overlapped with Mr Urra’s Large Order).

(b)        In F174 Mr Lopez had placed a Small Order and both Mr Lopez and Mr Sheth placed Large Orders.  Mr Sheth placed three Large Orders, of 500 lots each, which were on the market at the same time (including concurrently with Mr Lopez’s Large Order for 200 lots).  By the end of the Instance there were Large Orders of 1,700 lots, and Mr Sheth’s final Large Order was at the same price as Mr Lopez’s Large Order.

The volume of the Multiple Large Orders must have been seen by Mr Lopez in F174 as three of them overlap.  It is less obvious in F121 as Mr Sheth had a pattern in that Instance of placing the second then cancelling the first, so they overlapped for a shorter period of time.  Mr Shivji submitted that this occurrence in Multi Trader Instances is not enlightening as the Multiple Large Orders were not a mistake and were consistent with the practice which can be seen in other Instances of the Traders placing concurrent Large Orders.

630. The Tribunal is not persuaded that the Multiple Large Orders placed by Mr Sheth should necessarily be assessed entirely separately from the other Large Orders.  Mr Urra had also placed concurrent Large Orders within an Instance, and amended the price of his Large Orders closer to the touch. 

Collaboration

631. The Authority’s position is that the Traders were collaborating in real-time, in person, during the Instances to commit an abusive scheme, and, in the Multi Trader Instances, this involved updating each other on the progress of trading of the Small Order (including not trading or going slowly) and giving sufficient information for the other Trader(s) to place a Large Order in the opposite direction, and then confirming once the Small Order had filled (or mostly filled) at which point the other Trader(s) would cancel the Large Order(s).

632. The Traders did not share a common blotter, so if one Trader was placing a Small Order and another Trader was going to place a Large Order in the opposite direction to facilitate the execution of the Small Order, they would need to be told the direction in which to place the Large Order, when to do so and when to cancel (even on the assumption that pricing and the nudging of the Large Order closer to the touch were decisions taken alone by the Trader placing that Large Order).

633. We heard evidence from the Traders as to how they worked together on the Desk, during the course of which, as well as denying that they were collaborating on an abusive scheme, they each gave evidence that they did not discuss their specific trading activity with each other:

(1)        Mr Urra said that “opportunities for collaboration were very limited”  and “they rarely discussed specific trades, and they primarily focussed on their own books”.

(2)        Mr Lopez said that he “did not trade collaboratively with either Mr Urra or Mr Sheth” and that “[w]e did not have a common strategy on the EGB Desk as to how we were going to approach our market making activities”; even where he was hedging for Mr Urra, “I work completely independently”. 

(3)        Mr Sheth said that the Traders “would tend to trade independently, albeit pursuing the wider strategies and objectives of the EGB Desk”. 

634. We have found that there is no documentary, electronic communications or witness evidence of the Traders collaborating to commit market abuse.  Indeed, the Tribunal considers that the evidence goes beyond this (negative) finding:

(1)        There is positive evidence from Mr Hill and Mr Barouti that they saw and heard nothing which led them to suspect that the Traders were engaged in anything untoward.

(2)        The MHI Compliance Report describes its communications review and found no issues from its review of Bloomberg messages and chats, email communications and telephone records.

(3)        The Authority has had access to at least some of the electronic and voice communications between the Traders, namely telephone calls on fixed office lines, audio recordings from the dealer board system, emails and Bloomberg messages.  The Authority reviewed the communications data for the rationale for placing of the Large Orders, and have not disclosed that data, ie they do not rely on it themselves nor in the opinion of the Authority might it undermine the decision to take the referred action. 

(4)        Mr Lopez was on holiday for the first two weeks in August 2016, during which time Mr Urra and Mr Sheth were both interviewed by Compliance.  Yet there was no evidence of communications between the Traders during this period.

635. The Authority submits that this absence of evidence is unsurprising - there was no need for the Traders, who sat next to each other, to document their plan, the acts required to instruct each other to place and cancel orders at particular prices and times would have sounded like ordinary trading activity, and Mr Hill and Mr Barouti were focused on their own trading.  Further, and in contrast to the Traders’ denial of any form of collaboration (even about what was agreed to be genuine trading activity) Mr Shivji submitted that there would be reasons to have specific conversations of the type that is denied by the Traders about particular trading - avoiding the risk of errors (by inadvertently trading with each other on the Exchange or breaching the Desk’s limits), when hedging for each other, and the responsibility of Mr Urra and Mr Lopez to supervise when they delegated tasks to Mr Sheth.  Mr Shivji also relied on the activity and the coincidences of timing in the Multi Trader Instances themselves.

636. On the basis of all the evidence, the Tribunal concludes that the Traders were understating the extent to which they did work together and communicate in relation to specific trading activity:

(1)        We accept Mr Creaturo’s evidence that trading desks are dynamic environments with frequent communications between the traders, not just about general market information but more specific discussion about what the traders were doing at the time; and that it would be highly unusual for a trading desk not to communicate with one another.

(2)        They decided who would respond to particular RFQs, with there being no fixed allocation of RFQs to each of them.  Mr Urra would sometimes respond to medium-sized RFQs where he could offer a better price, a decision which itself would require communication as they identified what price they could each offer.

(3)        They hedged for each others’ books, and would need, at the very least, to communicate size and direction.  In F84 it can be seen that Mr Sheth placed the initial amount of the hedge, with Mr Lopez placing an order for more (we infer once the exact amount required was identified).

637. In addition, we find that Mr Sheth’s evidence was inconsistent in its approach as to the level of communication on the Desk.  His evidence was that if his Large Order had traded, he would have tried to manage it himself first for a few seconds, then sought the assistance of Mr Urra to exit the position.  But he also said that not only did he not discuss his use of the Information Discovery Strategy with Mr Urra in relation to particular clients, but that he did not check with Mr Urra whether it would be a good time to try using the strategy, ie whether Mr Urra would be available to assist.  This seems to present a difficult scenario, particularly where the Tribunal can see the level of activity on the Desk in short bursts of time and that there were occasions where Mr Lopez and/or Mr Sheth were hedging for Mr Urra yet somehow Mr Sheth was expecting that Mr Urra would have time to assist.  We do accept that the size of the Large Orders and the risks to the Desk, proportionate to the size of the Desk’s usual business, means that Mr Urra would have assisted; but consider it surprising that Mr Sheth would have proceeded on such a basis.

638. We find that the Traders would have been able to carry out the abusive scheme alleged by the Authority in person.  We agree with the Authority that, once the Traders came up with a plan (which was in essence very straightforward), they would be able to give any instructions and share information quickly in person in real time; that Mr Urra sat in between Mr Lopez and Mr Sheth would not preclude such coordination between Mr Lopez and Mr Sheth in Multi Trader Instances which did not include orders being placed by Mr Urra (eg F94 and F174) in circumstances where Mr Urra was aware of the abusive scheme.  The question is whether or not they did.

639. Mr Shivji drew attention - both in his submissions and during cross-examination of the Traders - to the occasions on which the Traders placed, amended or cancelled Large Orders so close in time to the activity of another Trader as to be “virtually simultaneous”.  The Traders said this was not a coincidence, as it was driven by market information.  Mr Shivji submitted that this was no explanation at all - on the Traders’ cases, they were pursuing different strategies, prompted (they say) by completely different triggers and with different aims.

640. The Tribunal addresses here the activity in the Multi Trader Instances relied upon by the Authority as Specified Instances.  The Traders say these are not multi trader instances at all, on the basis that (as we have found) there is no common blotter showing live orders on the Exchange, and (the Traders submit) the Traders were not collaborating and therefore did not know of the orders being placed by another Trader. 

641. We set out below the orders placed by the Traders in the Multi Trader Instances:

 

Small Orders

Large Orders

 

Trader

Size

Trader

Size

 

 

 

 

 

F30

Mr Urra

9 lots

Mr Urra

500 lots (39 filled)

 

Mr Lopez

5 lots

 

 

 

Mr Urra

39 lots (not filled)

 

 

 

 

 

 

 

F31

Mr Urra

39 lots

Mr Urra

499 lots

 

 

 

Mr Lopez

400 lots

 

 

 

 

 

F64

Mr Lopez

160 lots (122 filled)

Mr Urra

480 lots

 

 

 

Mr Urra

480 lots

 

 

 

Mr Sheth

500 lots

 

 

 

 

 

F82

Mr Sheth

20 lots

Mr Urra

450 lots

 

Mr Sheth

17 lots

Mr Sheth

500 lots

 

 

 

Mr Urra

500 lots

 

 

 

 

 

F84

Mr Sheth

200 lots

Mr Urra

490 lots

 

Mr Lopez

55 lots

 

 

 

 

 

 

 

F94 Part I

Mr Sheth

10 lots

Mr Sheth

500 lots

 

Mr Lopez

5 lots

 

 

 

 

 

 

 

Part II

Mr Lopez

5 lots

Mr Sheth

500 lots

 

 

 

Mr Sheth

500 lots

 

 

 

 

 

F121

Mr Sheth

22 lots

Mr Sheth

400 lots

 

 

 

Mr Urra

450 lots

 

 

 

Mr Sheth

400 lots

 

 

 

Mr Sheth

400 lots

 

 

 

Mr Sheth

400 lots

 

 

 

Mr Sheth

400 lots

 

 

 

Mr Sheth

400 lots

 

 

 

 

 

F174

Mr Lopez

25 lots

Mr Sheth

500 lots

 

 

 

Mr Sheth

500 lots

 

 

 

Mr Lopez

200 lots

 

 

 

Mr Sheth

500 lots

 

 

 

 

 

F205

Mr Urra

10 lots

Mr Urra

450 lots

 

Mr Urra

10 lots

Mr Lopez

200 lots

 

Mr Lopez

5 lots

 

 

 

Mr Lopez

5 lots

 

 

 

 

 

 

 

F209

Mr Lopez

35 lots

Mr Urra

450 lots

 

Mr Lopez

4 lots

Mr Sheth

250 lots

 

F30 at 17.39.34.225 and F31 at 17.45.10.137 on 10 June 2016

642. F30 and F31 involved Mr Urra and Mr Lopez.  The Authority’s case was that in F30 Mr Urra had placed a Large Order to help facilitate the trading of Mr Lopez’s Small Order, was “caught out” when 39 lots of his Large Order traded, then tried to sell those lots unsuccessfully.  Mr Urra then tried again to sell those lots, with them being the Small Order in F31, at which point Mr Lopez had helped by placing a Large Order.

643. The Tribunal was repeatedly referred to these two Instances, including in the context of the time of day at which the Large Orders were placed, that Mr Urra’s Large Order started to fill in F30 and he cancelled it and immediately tried to sell the 39 lots he had acquired, and the coincidences in timing. 

644. The only Large Order in F30 was placed by Mr Urra.  The trading activity was as follows:

(1)        At 17.39.34.225, Mr Urra placed a Small Order to sell nine lots, iceberged to five, for 141.92.

(2)        That did not trade, and at 17.39.45.444 Mr Urra placed a Large Order to buy 500 lots for 141.89. 

(3)        The Small Order filled.

(4)        At 17.39.54.403, Mr Lopez then placed a Small Order to sell five lots, iceberged to one, for 141.93. 

(5)        Mr Urra amended the price of his Large Order, increasing it to 141.90, and then increasing it further to 141.91 at 17.39.56.110. 

(6)        Mr Lopez’s Small Order filled at 17.39.59.350.

(7)        Mr Urra reduced the price of his Large Order to 141.90 at 17.39.59.393, 0.043 seconds after Mr Lopez’s Small Order filled. 

(8)        39 lots of Mr Urra’s Large Order traded, and he cancelled the remaining 461 lots at 17.40.01.104.

(9)        At 17.40.26.885 Mr Urra placed an order to sell 39 lots for 141.91, reduced the price to 141.90 and then cancelled the order.

645. Mr Shivji submitted that:

(1)        Mr Urra had not intended to buy Futures at this time of day - having placed an order to buy 500 lots, he cancelled that order as soon as it started to trade and then placed an order to sell the 39 lots which had traded in that Instance; and

(2)        Mr Urra’s amendments to the price of his Large Order, moving it away from the touch as soon as the Small Order traded, must have been coordinated.

646. Mr Urra and Mr Lopez denied any collaboration - Mr Urra’s evidence was that something was happening in the market here, and Mr Lopez said all he could see was his Small Order to sell five lots and he could not see Mr Urra’s activity.

647. F31 then started at 17.45.10.137 and involved Mr Urra placing a Small Order (to sell the 39 lots from F30), with Mr Urra and Mr Lopez both placing Large Orders (of 499 and 400 lots respectively).  It was a long instance, of 9 minutes and 50 seconds.  The trading activity was as follows:

(1)        At 17.45.10.137 Mr Urra placed a Small Order to sell 39 lots, iceberged to four, for 141.93.

(2)        Mr Urra placed a Large Order to buy 499 lots at 141.80 nearly 12 seconds later.  Mr Urra adjusted the price of this Large Order to 141.82, 141.83, 141.84, then back down to 141.83, 141.82, 141.81 then 141.80.  These price changes take place across 27 seconds.

(3)        At 17.46.21.136 Mr Lopez then placed a Large Order to buy 400 lots for 141.84. 

(4)        Less than 20 seconds later, Mr Urra amended the price of his Small Order, to 141.92 then 141.91.  The second price reduction is at 17.46.40.705.

(5)        At 17.46.45.359 Mr Urra then started a series of amendments to the price of his Large Order, to 141.81 and up to 141.88 and then reducing back to 141.81, all in increments of 0.01. 

(6)        At 17.47.17.885 Mr Lopez increased the price of his Large Order to 141.85 and then to 141.86 almost eight seconds later.

(7)        The following second Mr Urra then started to amend the price of his Large Order again, to 141.82, 141.83, 141.84 and then 141.85.

(8)        The Small Order started to trade at 17.47.33.255 at the price of 141.91.

(9)        At 17.47.39.633 Mr Urra amended the price of the remaining 15 lots of the Small Order to 141.92.

(10)     Less than two seconds later Mr Urra started amending the price of his Large Order, increasing it to 141.86, 141.87, 141.88, 141.89, 141.90 and then back to 141.89.

(11)     The Small Order started to trade again.

(12)     Mr Urra continued reducing the price of the Large Order, to 141.88, 141.87, 141.86, 141.85 and 141.84.

(13)     Mr Urra reduced the price of the Small Order, which continued to fill, and at 17.48.11.360 the final three lots of the Small Order were placed on the market. 

(14)     Mr Lopez and Mr Urra cancelled their Large Orders at 17.48.12.755 and 17.48.12.890 respectively.

(15)     Mr Urra increased the price of his Small Order and it filled at 17.55.00.136.

648. Mr Urra and Mr Lopez denied collaborating, with their evidence being that they had each placed their Large Orders for their own Trading Strategies, they did want to trade their Large Orders, with Mr Lopez adding that something was going on in the market, and it may have been linked to an OECD report. 

649. The Tribunal records that:

(1)        Mr Urra and Mr Lopez were, on their account, pursuing different Trading Strategies.  Mr Lopez was placing orders with the aim of predicting MHI’s client flow, whereas Mr Urra was going against MHI’s client flow in anticipation that the market maker with whom the client had traded would then be selling Futures.  Yet the pursuit of these different strategies led to them both placing Large Orders in the market in the same direction at the same time.

(2)        The price adjustments to Mr Urra’s Large Order were not keeping the order at a certain level away from the touch; he was moving the Large Order towards it and then away again, with the “peak” getting closer to the touch each time.

(3)        The result of this activity was that at 17.48 on a Friday evening the Desk had orders in the market for 899 lots.  There was very little visible volume on the Replay graphs, particularly on the buy-side.  On the basis of all of the evidence before us, we do not accept that Mr Lopez could reasonably have expected that he would win a cash trade of €40-50m of ten-year bonds at this time of day; or that Mr Urra could reasonably have anticipated that another market maker would be needing to sell a large number of Futures following a cash trade.

(4)        If Mr Lopez had bought 400 lots at this price, he would have been sitting on a loss of €112,000 on Monday morning’s opening. 

(5)        The two Large Orders were cancelled 0.135 seconds apart. 

F84 at 11.24.53.106 on 20 June 2016 

650. The activity involves all three Traders, but it is only Mr Urra who placed a Large Order (of 490 lots). 

651. Mr Lopez and Mr Sheth had placed Small Orders in this Instance.  We find that both of the Small Orders were a hedge for Mr Urra’s book.  Mr Lopez had accepted this (putting it forward as the likely explanation in his witness statement); Mr Sheth said it is possible that the Small Order was a hedge for a Spanish bond or could have represented him rebalancing the risk on the book generally, but on the basis of the evidence of recent cash trades and the size of the Small Orders, we conclude that Mr Urra had asked Mr Sheth to place this hedge.  Both of the Small Orders were placed at Best Offer.

652. The Instance length was 26 seconds; and the Instance involved a large number of transactions.

(1)        Mr Sheth placed a Small Order to sell 200 lots for 141.22, iceberged to 25, at 11.24.53.106.

(2)        Mr Lopez placed a Small Order to sell 55 lots for 141.22, iceberged to six, at 11.25.02.342. 

(3)        Both of these Small Orders started to trade.

(4)        At 11.25.07.805 Mr Urra placed a Large Order to buy 490 lots for 141.20.  (This was one tick from Best Bid.)

(5)        The Small Orders continued trading.  There were no price adjustments.  Mr Sheth’s Small Order filled at 11.25.13.231 and Mr Lopez’s Small Order filled at 11.25.18.120. 

(6)        Mr Urra cancelled his Large Order at 11.25.19.341. 

653. Mr Shivji submitted that collaboration between the Traders explains the timing of the cancellation of Mr Urra’s Large Order, and why Mr Lopez and Mr Sheth did not cross the spread and hit Mr Urra’s bid (which was only two ticks away from their orders). 

654. On the basis of this trading activity:

(1)        Mr Lopez and Mr Sheth had both placed Small Orders to sell at 141.22 to hedge for Mr Urra’s book.  This price was Best Offer, and both of these orders were already trading.  (Mr Lopez’s Small Order filled in about 16 seconds, and Mr Sheth’s in 20 seconds.)

(2)        If the Traders were collaborating in an abusive scheme in this Instance, Mr Lopez and/or Mr Sheth would need to have told Mr Urra that the trading had stalled.  They would not need to have told him direction, as he would have known this on the basis of his own cash trade.  Yet this is then a situation in which Mr Urra had asked Mr Lopez and Mr Sheth to hedge for his book, we infer on the basis that he was busy, but Mr Urra nevertheless was at his desk and had time to assist with the abusive strategy by placing a Large Order.

(3)        Mr Urra’s Large Order was one tick away from Best Bid, ie just two ticks away from the Small Orders.  The Small Orders were for a total of 255 lots.  The Replay graph shows there were only about five lots visible at Best Bid, and then about 100 lots visible ahead of Mr Urra’s Large Order one tick away.  If Mr Lopez had crossed the spread to get the hedge done, there is a real risk that he would have hit Mr Urra’s bid (depending on the invisible liquidity at the touch). 

(4)        The Large Order was live for 11.536 seconds, and cancelled within 1.221 seconds of the Small Orders filling.

F174 at 12.58.50.334 on 29 June 2016

655. This Instance was one of the two Instances questioned in the Eurex Letter.  Mr Lopez had placed a Small Order, he also placed a Large Order and Mr Sheth placed three Large Orders.

656. Mr Lopez had bought a €5m May 2023 bond on MTS from MPS Capital Services at 12.58.10.  This was on his own initiative, and not related to an RFQ.  He placed the Small Order to sell 25 lots partially to hedge this bond transaction.  

657. The trading activity was as follows:

(1)        At 12.58.50.334 Mr Lopez placed a Small Order to sell 25 lots, iceberged to three, for 142.15.

(2)        At 12.58.50.686 Mr Sheth placed a Large Order to buy 500 lots for 142.09.  Mr Sheth placed another Large Order to buy 500 lots for 142.10 at 12.58.52.674.  (The first Large Order remained live.)

(3)        Mr Lopez’s Small Order started to trade - there was a part fill of two lots.

(4)        At 12.58.57.448 Mr Lopez placed a Large Order to buy 200 lots for 142.11.

(5)        Four more lots of the Small Order traded.

(6)        At 12.59.00.588 Mr Sheth placed a third Large Order to buy 500 lots for 142.11. 

(7)        The Small Order continued to trade. 

(8)        At 12.59.00.785 Mr Lopez amended the price of his Large Order, increasing it to 142.12.

(9)        The Small Order continued to trade, filling at 12.59.03.536.

(10)     All four Large Orders were cancelled.  Of his three orders, Mr Sheth cancelled his order which was at the highest price first (ie the one that had been placed third, for 142.11), then those further away from the touch.  The cancellations were at 12.59.04.531 to 12.59.06.704.

658. The explanations given for the Large Orders were as follows:

(1)        In his witness statement and in cross-examination Mr Lopez explained how he may have considered that a support level could have formed at around 141.85, anticipated client buying interest based on RFQs which had been received by the Desk, and may have seen that the Bund/BTP Futures spread wanted to tighten.  He therefore placed his Large Order in anticipation of client buying interest.

(2)        Mr Sheth explained that he would have seen that Mr Lopez had bought the bond from MPS Capital Services and would have deployed the Information Discovery Strategy in relation to this.  In cross-examination, when presented with Mr Lopez’s explanation (ie that the cash bond was not related to an RFQ), Mr Sheth put forward an alternative possibility, namely that he had seen that MPS Capital Services, a market maker, was selling, and might have hypothesised that it might be on the back of what MPS treasury would be doing.

659. The coincidences in this Instance are:

(1)        Mr Sheth placed his first Large Order 0.352 seconds after Mr Lopez placed the Small Order.  Mr Lopez’s cash bond trade had been done 40 seconds previously, so at this point in time it is unlikely that Mr Sheth would have been responding to that information on the MTS screen.

(2)        Mr Sheth had placed Large Orders priced at 142.09 and 142.10.  Mr Lopez’s Large Order was priced at 142.11.  Mr Sheth placed his third Large Order at that price and 0.197 seconds later Mr Lopez increased the price of his Large Order to 142.12.

(3)        The whole Instance takes place over 56.37 seconds from placing of the Small Order to the final cancellation of the Large Orders.  Yet, within this, once the Small Order is filled, all four Large Orders are cancelled within 3.168 seconds.

(4)        The strategies relied upon by Mr Lopez and Mr Sheth would suggest that they are doing different things - Mr Lopez is trying to position himself for client buying interest (ie which would involve MHI selling bonds) whereas Mr Sheth is testing whether another market maker needs to sell Futures having bought bonds from a client splitting a sell-side RFQ.  Yet both have decided to place a Large Order to buy Futures within a very short period of time.

660. Both Mr Lopez and Mr Sheth denied collaborating with each other.  Mr Lopez’s evidence was that the timings were not a coincidence; the timing can be explained by the market information which would have been on the MTS screen, showing the movements in the cash bonds.  They both denied knowing of the orders which had been placed by the other Trader.

661. The Traders relied on the activity in this Instance, in particular Mr Sheth having placed Multiple Large Orders in a Multi Trader Instance, submitting that the Authority’s case relies on Mr Lopez and Mr Sheth both knowing what the other is doing, yet Mr Sheth had already placed two Large Orders, totalling 1,000 lots, at the time Mr Lopez placed a Large Order of 200 lots.  If they were coordinating, Mr Lopez would have known of these Large Orders placed by Mr Sheth, would not have seen the need to place another 200 lots, and would have seen that Mr Sheth was failing to amend the price of his Large Orders but instead placing new Large Orders at a different price and failing to cancel the earlier Large Order(s).  Essentially, the Traders submitted that, if the Traders were involved in an abusive scheme, Mr Lopez would have seen that Mr Sheth was deploying this poorly and risked drawing attention to their activity.

F209 at 10.12.49.319 on 22 July 2016

662. This trading activity took place over 38 seconds and involved Small Orders being placed by Mr Lopez and a Large Order being placed by both Mr Urra and Mr Sheth:

(1)        At 10.12.49.319 Mr Lopez placed a Small Order to sell 35 lots for 143.75, iceberged to six.    

(2)        At 10.12.56.113 Mr Urra placed a Large Order to buy 450 lots for 143.71, and at 10.12.56.217  Mr Sheth placed a Large Order to buy 250 lots for 143.71. 

(3)        The Small Order started to trade.

(4)        When 12 lots of the Small Order had traded, Mr Lopez placed another Small Order to sell four lots at the same price of 143.75, iceberged to three.

(5)        Mr Sheth cancelled his Large Order and Mr Lopez amended the price of both of his Small Orders, reducing them to 143.74, at 10.13.10.874 and 10.13.11.591. 

(6)        At 10.13.17.555 Mr Urra increased the price of his Large Order to 143.72. 

(7)        Mr Lopez’s Small Orders started to trade again.

(8)        Mr Urra reduced the price of his Large Order back to 143.71.

(9)        The last slice of Mr Lopez’s Small Orders filled at 10.13.22.278.

(10)     Mr Lopez then placed a small order to buy one lot for 143.74, which filled.

(11)     At 10.13.26.859 Mr Urra cancelled his Large Order.

663. Mr Lopez’s evidence was that the Small Orders were hedges of a cash bond he had just bought, being €5m December 2025 bonds from Cassa Centrale Banca. 

664. The coincidences of timing are:

(1)        Mr Lopez’s Small Order did not trade for more than six seconds, after which Mr Urra and Mr Sheth placed a Large Order in the same second (0.104 seconds apart).

(2)        After Mr Lopez placed his second Small Order there was a pause of almost nine seconds of no activity (as the touch has moved away).  Mr Sheth cancelled his Large Order in the same second as Mr Lopez amended his Small Orders towards the touch.

(3)        Mr Urra amended his Large Order towards the touch when there was a pause in trading; and moved away once the Small Orders started to trade again.

(4)        Mr Urra’s Large Order was on the market for 30 seconds, and cancelled 0.081 seconds after Mr Lopez’s Small Orders filled.

665. Mr Shivji submitted that these coincidences in timing are not explained by the Traders’ rationale.  Other market participants in a dynamic market would simply see the size at various price points, and that there was significant skew on the buy-side as a result of the Large Orders.  He also submitted that the trading in this Instance shows that the trading of the Small Orders was nudged by the placing and amendment of the Large Orders.

666. The Tribunal observes that this Instance shows the extent to which Mr Lopez was carefully managing the risk on his book - he had initially placed a Small Order to sell 35 lots, but 12 seconds later placed a further Small Order to sell four lots; yet at the very end of the Instance (and we record that Mr Urra’s Large Order was still live at this time) he bought one lot.

Conclusions

667. All three Traders feature in different combinations in the Multi Trader Instances, as shown by the table of the orders placed in the ten Specified Instances.  They each placed both Small and Large Orders, and this is relevant when assessing whether individually they were pursuing an abusive strategy (as we recognise that the placing of Small Orders with no knowledge of the Large Order(s) being placed by another Trader in that Instance would not of itself mean that the Trader placing the Small Order was engaged in an abusive strategy). 

668. We do, however, consider that the coincidences in timing in the activity in the Multi Trader Instances, including but not limited to those described in detail above, strongly supports a conclusion that the Traders were collaborating.  If that is the case, then the placing of concurrent Large Orders by the Traders is relevant both to whether the Multiple Large Orders placed by Mr Sheth were poor practice or a deliberate intent to put more pressure on the order book; and also to the submissions as to whether the Large Orders were “large”, and this shows that it is not a standalone question for each individual Large Order but the total volume of the concurrent Large Orders by reference to the market at that time.

669. When assessing collaboration, we consider that the strongest submissions and evidence in favour of the Traders are not ones based on the specific activity, but instead:

(1)        If the Traders were collaborating to commit an abusive scheme, and had planned how to achieve this, we would expect that they would have come up with an explanation for their activity, yet they have set out two different explanations.

(2)        The abusive scheme alleged by the Authority involves three Traders and they have managed to avoid committing any evidence of this scheme to any form of electronic communications, whether before, during or after the Relevant Period.  This is in circumstances where, whilst the scheme could be conducted in real time in person, Mr Lopez was on holiday at the time of Compliance’s initial investigation when Mr Urra and Mr Sheth were both interviewed by Compliance.  

Plausibility of Authority’s case that the Traders conducted an abusive scheme

670. The challenges put by the Traders to the Authority’s case included:  the abusive scheme would not have worked (looking at market impact, liquidity and volatility) and would not therefore have facilitated the trading of the Small Orders; the scheme would not have benefitted the Traders; it would have been easily detected; and it does not explain all the features of the Traders’ trading activity during the Relevant Period. 

Whether the abusive scheme would have worked

671. Mr George submitted that the Authority’s case is that the Traders’ true purpose when placing the Large Orders was to facilitate the execution of the Small Orders on the opposite side of the book, rather than to execute the Large Orders.  This argument can only succeed, he submitted, if it can be shown that the Traders hoped or believed that the Large Orders would in fact facilitate the execution of the Small Orders - if they did not have any such understanding, then there would be no rationale for placing any of the Large Orders unless they intended to trade them.

672. We have concluded that the Large Orders were each likely to give the impression or signal of significantly increased supply or demand, and the most likely market reaction would be for the market to move in the opposite direction to the Large Orders, ie towards the Small Orders.  We have also concluded that a Trader placing the Large Order in these circumstances would have known of this likelihood. 

673. Our analysis and conclusions on market impact focused on the impact of the Large Orders.  We address here the position of the Small Orders in the Specified Instances, and the plausibility of the Authority’s case that the Traders would commit market abuse to facilitate the trading of these orders in the market at the time at which they were placed.  The Traders drew attention to various features of these Small Orders in different contexts, and made submissions in relation to:

(1)        the number and size of the Small Orders;

(2)        the pricing of the Small Orders; and

(3)        whether the market was already moving towards the Small Order, and, in some Instances, the Small Order was already trading.

Number and Size of the Small Orders

674. The Small Orders in the Instances are only a small proportion of the orders of fewer than 200 lots placed and executed by the Traders in the Relevant Period. 

675. Mr Kasapis suggested in his report that this demonstrated that the Small Orders in the Instance Pool were “independent” as “the larger order was clearly not necessary to cause the smaller orders to trade”.  We do not accept this - the Authority has not suggested, nor do we consider it to be plausible, that every time that the Traders placed an order on the Exchange they would commit market abuse to facilitate the execution of that order.

676. That the Traders are alleged to have committed market abuse to facilitate the execution of some of their small orders does give rise to a question as to the decision to place large orders on the opposite side of the book to these particular small orders. 

677. The Authority’s approach was to identify small orders as being those of under 200 lots, but modified this to include two orders of 200 lots or more which were iceberged to 22 and 25 (relying on the smaller size being shown to the market).  However, there was a wide range in the size of the Small Orders that were within the Instance Pool, and this was reflected in the Specified Instances:

(1)        The Small Orders in the Specified Instances for Mr Urra were from eight to 90 lots, those for Mr Lopez were from one to 90 lots (with three of the ten Specified Instances involving Small Lots of five lots or fewer) and those for Mr Sheth were from five to 80 lots (where the Small Order of 80 lots was placed on the same side and at the same price as another Small Order he placed for ten lots).    

(2)        In the Multi Trader Instances the Small Orders included five Small Orders of five lots each (all of which had been placed by Mr Lopez, although these were sometimes concurrent with additional Small Orders on the same side), but in F64 the Small Order (placed by Mr Lopez) was of 160 lots and in F84 the Small Orders were of 200 and 55 lots (placed by Mr Sheth and Mr Lopez respectively).  

(3)        The Small Order in F132 (Mr Lopez) was an order to buy one lot.  This is a Specified Instance selected by the Authority.  Whilst it is an illiquid time of day (Mr Lopez placed a Large Order to sell 200 lots at 17.07.53.566), and being at the end of the day we are not persuaded that Mr Lopez could reasonably have expected that there would be imminent client buying interest of this size (which goes to the plausibility of the Anticipatory Hedging Strategy), the Small Order was placed at Best Bid.  Looking at the Replay graph, including the position of the Small Order in the queue and the minimal volume showing as available at the touch on the sell-side, we consider that there must be a question as to the likelihood of the Small Order filling.  However, as Mr Jaffey emphasised in his submissions, crossing the spread to get this Small Order done would have cost just €10. 

(4)        Across the 40 Specified Instances, there were seven orders of 50 lots or more.

678. Logically, it might be expected that if a Trader is to commit market abuse in relation to only some of the small orders they were placing in the Relevant Period, they would not do so when the small order was tiny but would focus instead on facilitating the execution of the “larger” small orders.

Market direction and Small Order already trading

679. We cannot know on the basis of the evidence before us whether there was any actual market impact; the submissions in relation to market direction and the Small Order trading are relevant to whether the Small Order would have traded in any event, and thus, potentially, the likelihood of a Trader deciding to commit market abuse to facilitate execution of the Small Order.  

680. The Tribunal has addressed Mr Kasapis’s Market Trend Analysis and agreed with the Authority that the rounding of minutes and the absence of marking the time at which the Large Orders were placed are significant weaknesses in the approach taken by Mr Kasapis.  The Tribunal also identified that in three Specified Instances for Mr Sheth the price graphs showed that the market was trending away from the Small Order until the Large Order was placed (in contrast to Mr Kasapis’s conclusion that the market was trending towards the Small Order in these Instances). 

681. We are not persuaded that this analysis bears the weight attached to it by Mr Kasapis (and observe that Mr Kasapis did not conduct this analysis for the Multi Trader Instances).  We do, however, recognise that there is a difference in the outcomes amongst the Traders, and in particular the conclusions he reached in respect of Mr Urra’s Small Orders in the Single Trader Instances.  Mr Kasapis identified that the market was trending towards Mr Urra’s Small Order in seven of the ten Specified Instances, and was neutral in one further Instance, ie there were only two Instances where the market was trending away from Mr Urra’s Small Order.

682. In 22% of the Instances, and 20% of the Specified Instances, the Small Order had already started to fill before the Large Order was placed.  We have looked at the activity in these Specified Instances:

(1)        F7 (Mr Urra) - 18 of the 40 lots had filled before the Large Order was placed.  However, these 18 lots had filled within one second of the Small Order being placed, and there was then a pause of more than ten seconds.  The Large Order was placed and within 0.2 seconds the Small Order started to trade again.  (This was an Instance where the visible liquidity throughout the stack, not only at the touch but also up to four to five ticks away was multiples of the actual size of the Small Order.) 

(2)        F64 (Multi Trader Instance) - Mr Lopez had placed a Small Order for 160 lots iceberged to six, 31 of which had traded over a period of almost 19 seconds, before Mr Urra placed a Large Order.  The Small Order continued to fill relatively slowly, with some part fills being of one and two lots.  122 lots filled in total over a period of 2 minutes 30 seconds (after which Mr Lopez cancelled the remaining 38 lots).

(3)        F84 (Multi Trader Instance) - Mr Sheth’s Small Order was 200 lots iceberged to 25 and Mr Lopez’s Small Order was 55 lots iceberged to six.  Both were selling at 141.22 and had started to trade - 25 lots for Mr Sheth, six for Mr Lopez.  They both filled in this Instance (in 20 seconds and 16 seconds respectively) and the trading was alternating between them as the slices of their iceberged orders were placed in the queue.  At the time the Large Order was entered, the visible volume at the touch on the buy-side was about ten lots.  The Large Order was just one tick from the touch (ie two ticks from the Small Orders) and there were about 90 visible lots ahead of it in the queue.      

(4)        F150 (Mr Urra) - The Small Order of 90 lots was not iceberged.  29 lots traded within 0.5 seconds of the Small Order being placed.  There was then a pause of more than seven seconds before the Large Order was placed.  At the time the Large Order was placed, the visible volume on the buy-side of the stack was about ten lots, with about 35 lots visible one tick away.  After the Large Order was placed, the Small Order started to trade again within 0.1 seconds. 

(5)        F194 (Mr Urra) - The Small Order was 25 lots iceberged to four.  Eight lots had traded before the Large Order was placed.  The first four lots had traded 0.741 seconds after the order was placed, the next slice was entered and there was a pause of more than 10 seconds.  The next part fill of four lots occurred before the Large Order was placed, but these two events were within 0.5 seconds of each other.  The Small Order continued to trade, and it filled in 12.774 seconds.  This is an Instance where there was better volume available by reference to the size of the Small Order - there were about 45 lots visible at the touch on the buy-side at the time the Large Order was placed.

(6)        F201 (Mr Sheth) - One lot of the Small Order of ten lots had traded within one second of the Small Order being placed.  It then paused and Mr Sheth amended the price of the Small Order 50 seconds later (increasing it twice).  Having amended the price, Mr Sheth placed his first Large Order more than eight seconds later.  Mr Sheth placed two Large Orders, cancelled both, made two further increases to the price of the Small Order, placed a third Large Order, then reduced the price of that order twice, and the remaining nine lots of the Small Order filled in one go.  These nine lots filled nearly four minutes after the Small Order had been placed.  The market was moving away from his Small Order throughout this Instance - whilst one lot had traded, at the time he placed the first Large Order the Small Order was two ticks from the touch at the back of the queue at that price point (with more than 80 lots ahead of it at that price, and around 110 lots better priced).

(7)        F232 (Mr Urra) - Mr Urra had placed a Small Order for 25 lots, uniceberged.  Five lots traded within 0.5 seconds of the order being placed, and a further single lot more than eight seconds later.  The Large Order was placed two seconds later.  The remaining 19 lots filled within 0.2 seconds.  That the Small Order had started to trade shows that the whole of the uniceberged order was at the front of the queue.  There were about 40 lots visible at the touch on the sell-side.

683. The above are all the Specified Instances in which the Small Order had started to trade before the Large Order was placed.  It can be seen that the position is different across the Traders (none of Mr Lopez’s Small Orders in the individual Specified Instances had started to trade, although all of those were at the touch when he placed the Large Orders).  It is notable that some of Mr Urra’s Small Orders were already trading.

684. However, taking account of the Market Trend Analysis (whilst bearing in mind the identified weaknesses), and looking at the actual trading activity in these Specified Instances where the Small Order was already trading, including the pace at which the Small Orders were filling, the pauses in trading before the Large Orders were placed, and the visible volumes on the stack, we are not persuaded that a Trader who had placed the relevant Small Order could necessarily have been confident of it filling swiftly at the price at which it had been placed.

Pricing of the Small Orders

685. In most of the Specified Instances for Mr Urra, the Small Order was at the touch at the time the Large Order was placed - this could be seen in F7, F12, F32, F60, F150, F152, F194, F203 and F232.  In F47 the Small Order was one tick from the touch at the time the first Large Order was placed and at the touch when the second Large Order was placed. 

686. In all of the Specified Instances for Mr Lopez, the Small Order was at the touch at the time at which he placed the Large Order.  Mr Jaffey drew attention to the activity in F190 where the Small Order to sell 45 lots had been placed one tick away from Best Offer, which then became the Best Offer, and he reduced the price thus crossing the spread and it started to trade.

687. In the Specified Instances for Mr Sheth, the Small Order was at the touch in five of them, one tick away in two Instances (although in F158 there were two Small Orders, one of which was at the touch one was one tick away) and two ticks away in three Instances.

Conclusions on facilitation of the trading of the Small Orders

688. The Traders referred to the spread generally being only one tick, and that therefore trading at Best Bid or Best Offer rather than crossing the spread would save only €10 per lot.  The Tribunal accepts that.  However, whilst that raises significant doubts as to whether the Traders would have been prepared to engage in an abusive scheme to ensure execution of a Small Order of just a few lots, this risks ignoring the need to execute hedges swiftly, and the fact that in some Instances the visible volumes were low.

689. By way of illustration, in F209, a Multi Trader Instance, Mr Lopez placed his Small Order to sell 35 lots iceberged to six at a price of 143.75 which improved Best Offer by one tick (this did not cross the spread as there had been a brief period where the spread was two ticks).  However, it did not trade for six seconds.  At the time at which the first Large Order was placed there was visible volume of 20 to 25 lots at Best Bid; the first slice of the Small Order was at the front of the queue but there were visible lots of about 30 lots behind it (so subsequent slices would be placed behind these).  The stack was slim - ignoring the Large Order, at the time the first Large Order was placed the visible volume did not exceed 70 lots at any price point on either side of the order book.

690. The Small Orders were hedges for their cash bond transactions, or repositioning of their books.  They were the trades that the Traders were actually transacting and needed to transact.  We take account of the importance for the Desk of trading these Small Orders.    

Abusive scheme would not have benefitted the Traders

691. The Traders submitted that the Authority has not identified any financial benefit to the Traders, with Mr Jaffey submitting that the Authority has no pleaded case on motive.

692. The Authority’s Decision Notices refer in their “Summary of Reasons” to:

(1)        the Desk’s role being to provide prices and liquidity in EGBs to clients and that the Traders would often hedge their trades with clients through Futures on Eurex, and

(2)        the abusive trading strategy alleged is described as having the objective of assisting the execution of a smaller genuine order they wished to trade on the opposite side of the order book.  Other market participants would likely have altered their trading strategies as a result of the false and misleading signals given by the Large Orders, eg when a Trader placed a large buy order it gave a false signal that there was a material buyer in the market and other buyers, anticipating that the market was likely to move higher, would likely act with more urgency in order to secure the execution of their buy orders.

693. The Decision Notices set out the “Facts and Matters” and made the following allegations under “Trader remuneration and performance of the Desk”:

(1)        The Desk often lost money as a result of trading with clients but it was strategically important for MHI to offer EGBs to clients of MHI’s other services.  Senior management had increased the hit ratio for the Desk, requiring the Traders to execute a higher proportion of client orders than previously. 

(2)        In order for the Desk to be successful, and to achieve the increased hit ratio, it was necessary to respond to clients quickly and with as competitive a price as possible.  Through the use of the abusive trading strategy, the Traders aimed to respond to clients’ RFQs more quickly, and make more competitive prices with increased certainty, in order to increase their hit ratios. 

(3)        Whilst the Traders were remunerated based on a range of weighted factors, the performance of the Desk was a significant factor when calculating the Traders’ bonuses.

694. In the context of setting out the Authority’s determination of the financial penalty, the Decision Notices recorded that the Authority had not identified any financial benefit that the Traders derived directly from the market abuse, but does also note that the aim of executing genuine orders more efficiently and managing better the risk on their book as a result of the placement of misleading orders would have improved the performance of the Desk which was a factor taken into account in determining the bonuses they were to receive.

695. In its Amended Statements of Case the Authority sets out the “Facts and Matters relied on by the Authority”, and they include:

(1)        To secure the order in response to an RFQ, it is generally necessary to ensure that the price provided is sufficiently competitive.  The price will also be determined by a number of other factors, including the trader’s assessment of the risk posed to his book.  In general, the risk is minimised by hedging as soon as possible after the trade because this minimised possible exposure to moving prices.

(2)        In April 2016 the target hit ratio of the Desk had been increased to 15%, requiring the Desk to execute a higher proportion of client orders than they needed to previously.  The hit ratio was likely to increase if RFQs were responded to quickly and with as competitive a price as possible.

(3)        The Traders’ remuneration package from MHI consisted of fixed remuneration (a salary) and variable remuneration (a discretionary bonus).  Discretionary bonuses were dependent on achievement of individual objectives, measured formally via a Balanced Scorecard Appraisal, which gave scores across a range of weighted factors according to objectives set each April. 

(4)        The abusive strategy involved the placing of misleading orders which the Traders did not intend to trade but instead intended or hoped to facilitate the execution of other genuine orders on the opposite side of the order book.

696. In setting out the alleged breaches of s118 FSMA 2000 and Article 15 of the Market Abuse Regulation, the Authority stated that the purpose of placing the misleading orders was to facilitate the execution of genuine orders at a more advantageous price, or on a more timely basis, than would otherwise have been achieved but for their having misled other market participants by the misleading orders.

697. The Tribunal considers that the Authority has throughout set out the reasons which it relies upon in support of its allegations that the Traders committed market abuse and, within this, the alleged benefits to the Desk and to the Traders of pursuing an abusive strategy.  This is a separate issue to whether these alleged benefits, or even any identified benefits, would be sufficient to form a reason for the Traders to commit market abuse. 

698. We have found:

(1)        Market making in EGBs is very competitive.  Most market makers, having traded in the cash bond market and who needed to hedge their risk, would hedge with Futures and would do so swiftly in order to manage their exposure to market movements.  The Desk would generally make less than one tick per lot on cash trades.

(2)        The Small Orders were the trades in Futures that the Traders were actually transacting and needed to transact.  They were important to the Desk.   

(3)         Mr Urra was an experienced trader, well-regarded in the industry and had recently been promoted to Head of European Rates Trading at MHI.  He was recognised as being a successful trader, and Mr Heiberg was satisfied with his personal performance.  He had received a total bonus award of £325,000 for 2015/2016.

(4)        Mr Lopez was an experienced trader and was considered to be a highly responsible individual with a strong moral compass by Mr Heiberg.  He was still in his probationary period at MHI during the Relevant Period.

(5)        Mr Sheth was the junior trader on the desk and had recently been promoted from Analyst to Associate Trader.

(6)        The 15% hit ratio introduced by Mr Heiberg was an increase on what the Desk as a whole had previously been achieving but Mr Heiberg considered that Mr Urra was already meeting this individually, it was an informal target and was not introduced as one of the defined performance objectives for any of the Traders.

(7)        The Desk had a target of £8.5m for 2016/2017 and had not met its target for the two preceding financial years.

(8)        The Traders had personal targets.  Whilst we accepted that Mr Sheth had not known his specific target during the Relevant Period, we found that Mr Sheth knew that he was expected to generate an overall profit.

(9)        The Traders’ remuneration included a discretionary bonus, the amount of which would be based on performance measured by reference to performance objectives which took into account several factors including their own P&L or revenue generation (and, for Mr Urra, the performance of the Desk).

(10)     The Traders’ proposed bonus awards for 2016/2017 were £400,000 for Mr Urra, £50,000 for Mr Lopez and £100,000 for Mr Sheth.  These were ultimately withheld by MHI.

699. The Authority is not required to establish that the Traders would have obtained significant (or any) benefit from the alleged market abuse, but in circumstances where a large part of the Authority’s case relies on inferences it submits should be drawn from the activity, we consider that it is relevant to consider why the Traders might have been motivated to commit market abuse.  This is particularly so in circumstances where the abusive scheme is alleged to have been committed deliberately and dishonestly on a multitude of occasions in a relatively short period of time and would result in serious consequences for the Traders. 

700. The Traders submitted that there is no basis for concluding that any of the Traders could have believed that the abusive scheme alleged by the Authority would have been worthwhile, whether by reference to the execution rates of the Small Orders, performance of the Desk or individual Traders, or have a meaningful effect on bonuses.

701. In this regard, the Tribunal considers that the Traders’ submissions have failed to acknowledge the fact that the Traders were market makers and that their job was to win RFQs submitted by MHI’s clients for the Desk by offering attractive prices in a competitive market and hedge the resulting exposure.  These hedges needed to be executed swiftly, and we consider that they were important, even if viewed in isolation some of them were small.  This was their routine business, a business they were seeking to improve.  We do not, therefore, disregard the importance to the Desk, and thus to the Traders who worked on that Desk, of being in a position to win RFQs and execute hedges swiftly (and at a good price).

702. For each of the Traders, our findings as to their experience and remuneration do also illustrate reasons which we accept might make it seem unlikely that they would commit this abuse - Mr Urra was successful, and was already receiving high bonuses, Mr Lopez had only recently joined MHI, Mr Sheth was junior and had recently been promoted.  However, we have found that the Desk performance and individual performance is a factor in determining the amount of discretionary bonuses, and the size of Mr Urra’s bonuses in prior years and the proposed amounts of the bonuses for 2016/2017 show the potential rewards for the Traders if they were found to be performing well.  It would be an error to dismiss this as of no consequence.

703. Further, we consider that this emphasis on the lack of (or minimal) rewards for an abusive strategy should not be assessed in isolation and must take account of the Traders’ expectations as to the risks involved, or the potential for their conduct to be detected. 

Absence of direct evidence of Traders collaborating to commit market abuse

704. Our conclusions in relation to the absence of direct evidence of the Traders collaborating to commit market abuse - including both our findings in relation to Mr Hill and Mr Barouti, and the absence of evidence from electronic communications - are relevant to assessing the plausibility of the Authority’s case. 

705. We place more weight on the absence of evidence from electronic communications.

Risk of detection

706. Mr George submitted that the alleged abusive scheme would have been easily detected.

707. We have found that:

(1)        The Desk was located in an open-plan working environment, and conversations between the Traders could be overheard. 

(2)        Each of the Traders expected and understood that their communications and online activities would be monitored and recorded by MHI and that their trading would be monitored by the Exchange. 

(3)         The Traders did not know the specifics of the monitoring that was conducted by MHI, save that they were aware that Mr Heiberg had access to their live positions throughout the day and that, as limits applied at the end of the day, Compliance would have access to their positions at the end of the day.   

(4)        The Traders did not, at the beginning of the Relevant Period, know that orders in Futures were not being monitored by MHI. 

708. Whilst the Relevant Period started on 1 June 2016, the first Instance was not until 7 June 2016 - there were seven Instances that day.  On that single day, Mr Urra placed seven Large Orders of between 400 and 490 lots, and Mr Lopez placed one Large Order of 300 lots, which was concurrent with a Large Order of 444 lots which had been placed by Mr Urra.  None of these orders filled, and all were cancelled.  No questions were asked by Mr Heiberg, Compliance did not request an explanation, and there was no immediate challenge from the Exchange.  The pattern of activity continued that week, with a further eight Instances the following day (although those Instances only involved Mr Urra’s trading activity).   

709. As a matter of fact, we do not agree with Mr George’s submission that this alleged abusive scheme would have been easily detected as the reality is that it was not.  The Eurex Letter is dated 26 July 2016 and related to two Instances of trading which had taken place on 29 June 2016, both of which had involved the Multiple Large Orders placed by Mr Sheth.  It was the various investigations prompted by the Eurex Letter that resulted in the identification of the Instance Pool.

710. The Tribunal considers that this raises a question as to what the Traders can be inferred to have learnt about the monitoring during the course of June 2016 in any event.  We found that the Traders did not, at the beginning of the Relevant Period, know that orders in Futures were not being monitored by MHI.  We consider it would have been reasonable for Mr Urra to have concluded, by the end of 10 June 2016 (at which point 33 Instances had occurred) that MHI and the Exchange were not monitoring, or flagging as potentially abusive, the placing and cancellation of orders of less than 500 lots.  

711. When we look at the size of the Large Orders, Mr Urra generally placed Large Orders of 400 to 499 lots (with one Large Order of 500 lots), most of Mr Lopez’s Large Orders were of 200 lots, and most of Mr Sheth’s Large Orders were of 500 lots.  None of the Large Orders were for more than 500 lots.  There were a number of Instances involving concurrent Large Orders, of 800 to 1,000 lots (with some of the Multiple Large Orders resulting in volumes of up to 2,000 lots). 

Authority’s alleged scheme cannot explain all trading activity

712. The Traders submitted that the Authority’s alleged scheme does not explain all of the trading activity of the Traders.  We assess this below.

Trading Activity of the Traders in the Relevant Period

713. The Traders have throughout submitted that the Authority’s failure to address all of the trading activity of the Traders in the Relevant Period is a significant defect in its approach, as they submitted that this is inconsistent with the Authority’s allegations of an abusive strategy but entirely consistent with the pursuit of what they submit are their legitimate Trading Strategies.

Amendment of price of Large Order in Instance Pool after Small Order filled

714. The Traders referred to the fact that in some Instances the price of the Large Order was amended after the Small Order had traded.  Mr Kasapis’s evidence was that these price amendments demonstrated that the Large Orders were independent from the Small Orders, as there would be no reason to adjust the Large Order and keep it live beyond the Small Order having traded if the purpose of the Large Order was to cause the Small Order to trade. 

715. Mr Shivji submitted that these Instances were not illuminating - the market was volatile, the Large Orders were kept away from the touch, and keeping the Large Orders (rather than cancelling them) did not necessarily show that they were genuine. 

716. The Instances identified by Mr Kasapis are set out below.  The first three columns are produced from the table in Mr Kasapis’s report.  The fourth column has been calculated and added to the table by the Tribunal.

Trading Instance

Time Large Order remained after price adjustment (seconds)

Trader

Time between Small Order filled and price adjustment (seconds)

F32

2.092

Mr Urra

0.026

F86

2.288

Mr Urra

1.018

F90

8.167

Mr Lopez

0.683

F161

5.282

Mr Lopez

1.077

F170

2.063

Mr Urra

0.559

F171

2.852

Mr Lopez

0.829

 

717. The possible significance (or otherwise) of these amendments was put to both of the experts during cross-examination:

(1)        Mr Shivji put it to Mr Kasapis that in some of these Instances the price of the Large Order was amended and then another Small Order was placed by one of the Traders.  However, whilst Mr Kasapis accepted this was possible, the Tribunal was not shown any occasion on which this had happened.      

(2)        Mr George put these Instances to Mr Creaturo and Mr Creaturo’s opinion was that as MHI had a manual amendment process, these Instances could simply be a case of a Large Order being amended manually at the same time at which the Small Order filled. 

718. In closing, Mr Shivji submitted that the Large Orders may have remained live and been kept away from the touch whilst the Traders considered whether to place further Small Orders on the opposite side; and referred to Mr Kasapis having recognised that a variant of spoofing is to place larger orders and manipulate the touch, and then place other orders afterwards.  However, we consider that such possibilities are speculative, were not pleaded by the Authority and we were not taken to any supporting evidence. 

719. Looking at the period of time between the Small Order being filled and the price of the Large Order being amended in the six Instances identified by Mr Kasapis, we agree with Mr Creaturo that in at least three of these (F32, F90 and F170) it is likely that the price amendment was being made at effectively the same time as the Small Order filled. 

720. All three of Mr Urra’s Large Orders were cancelled within three seconds of the price amendments being made (and within 3.5 seconds of the Small Order being filled).

721. Of the six Instances identified, it is F90 and F161 which are potentially the most interesting; both of these were Large Orders placed by Mr Lopez, and although in F90 the price amendment was in the same second as the Small Order filled, both of these Large Orders were live for several seconds after the price was amended.  The Large Order in F90 had been placed three ticks from the touch, and the amendment was to move it closer, to two ticks from the touch.  The Large Order in F161 had been placed three ticks from the touch, and during the Instance its price was improved so it was two ticks away; the amendment after the Small Order filled was to reduce the price of the buy-side order so it was three ticks away, and it was cancelled when the market started to move towards the order.  It is not therefore correct that the Large Order was amended away from the touch in both of these two Instances; however, we do agree that based on volumes there was no immediate likelihood of these Large Orders trading.

Lone Large Orders

722. The Tribunal has found that:

(1)        Mr Urra placed 15, Mr Lopez placed five and Mr Sheth placed 14 Lone Large Orders. 

(2)        There is nothing to distinguish the Lone Large Orders from the Large Orders in the Instance Pool, save for the fact that there was no small order on the opposite side of the order book. 

723. The Traders submitted that these Lone Large Orders are a real problem for the Authority - they are a meaningful number and the Authority has not put forward any explanation for them.  They are, they submitted, entirely consistent with the Trading Strategies as put forward by the Traders and demonstrate that, whatever the Traders were hoping to achieve by placing the Large Orders, it cannot have been to abusively facilitate the execution of the Small Orders.

724. The Authority’s position was:

(1)        The Lone Large Orders represent a very small proportion of the Traders’ large orders, and may be even smaller than identified by Mr Creaturo given that there were examples where the Lone Large Orders appear to have missed overlapping with a Small Order by milliseconds.

(2)        The Authority’s case is not that every single time that the Traders placed a large order they were intending to facilitate the execution of a small order.  It was not Mr Creaturo’s job to provide an explanation for these orders; this was not what the experts had been instructed to do.  Mr Creaturo’s view was that, although he could not explain why the Lone Large Orders were placed, they were not explained by the Traders’ Trading Strategies, because he did not accept that the Traders’ alleged strategies were plausible explanations for any of the large orders.  

(3)        The evidence put forward by the Traders as to the placement of these Lone Large Orders was subject to the same flaws identified in respect of the evidence in relation to their Large Orders in the Instance Pool.

725. The Tribunal considers that it is accurate to describe the number of the Lone Large Orders as being a small proportion of the total number of large orders placed by the Traders in the Relevant Period.  However, the numbers are not insignificant, and their existence does rather beg the question as to why they were placed.  This is a question that the Authority is not required to answer, as it has not pleaded that they were abusive, but as the Traders submit that all large orders they placed were in pursuit of their Trading Strategy we have considered the further explanations which have been provided.

726. The Traders submitted that they placed all of their large orders in pursuit of their Trading Strategy (ie the Large Orders within the Instance Pool and the Non-Instance large orders which included the Lone Large Orders).     

727. Mr Urra did not provide further evidence in his witness statement as to his placement of these 15 Lone Large Orders.  Mr George did rely on the fact of their existence. 

728. Mr Lopez and Mr Sheth did both give evidence in relation to particular Lone Large Orders they had placed in their witness statements on which they were cross-examined.  The Authority submitted that, even for these, the Traders have not offered a plausible explanation. 

Lone Large Orders placed by Mr Lopez

729. Mr Lopez placed five Lone Large Orders.  He provided further evidence in relation to two of these in his witness statement.

730. Mr Lopez placed a large order to sell 200 lots for 140.19 at 9.40.26 on 16 June 2016.  That order was cancelled after 17 seconds.  (Later that day he placed Large Orders in F62 and F63.)  His evidence was that he had placed the Lone Large Order in anticipation of client selling, and hoped to catch a “pull back” to 140.19 (being one tick below where he considered a resistance level may be forming), and likely cancelled because he thought this pull back was not going to happen. 

731. Having looked at the Futures activity on the Desk the Tribunal accepts that there was no opposing small order in the immediate proximity of this Lone Large Order.  It was truly “Lone”.

732. On 17 June 2016 Mr Lopez placed a large order to buy 300 lots for 140.26 at 12.05.17.  That order was cancelled after three seconds.

733. Mr Lopez referred to identifying strong client buying interest - the Desk had received an RFQ from Banca d’Italia at 10.10.53 for €25m June 2026 bonds, and three further RFQs had been received from Banca d’Italia between 11.41.52 and 11.49.08 (each for different bonds, but none of which were of sufficient size to require hedging with at least 200 lots of Futures and so were not amongst the 93 RFQs). 

734. Mr Lopez explained that he had bought €10m of December 2019 bonds (in two clips of €5m each) from Credit Agricole on MTS on his own initiative, at 12.04.05 and 12.04.13.  He explained that he executed these bond transactions as he wanted to make a small adjustment to the position of his book; it was an anticipatory position in that he anticipated he would be able to sell the bonds later at a better price, in line with his buying view at the time.  At 12.05.10 he placed a small order to sell five lots of Futures for 140.31, iceberged to two, to partially hedge his cash trades; that order filled at 12.05.17.

735. After placing and cancelling the Lone Large Order, he continued to buy cash bonds in shorter maturities on his own initiative, and at 12.05.29 and 12.06.29 bought a further €10m of December 2019 bonds (in two clips of €5m each) from Credit Agricole on MTS; these were anticipatory orders in anticipation of client buying.  Mr Lopez said the client buying interest he anticipated did materialise - at 12.29.47 the Desk received an RFQ for a client buy order for €25m June 2021 bonds from Banca d’Italia.

736. Mr Lopez relied on this activity to explain his approach to anticipatory hedging and his placing of large orders, including the Large Orders.  The Tribunal does not consider this bears that weight: 

(1)        The Lone Large Order here was placed in the same second that his small order in the opposite direction filled, and he cancelled the Lone Large Order three seconds later.

(2)        Mr Lopez cancelled this large order to buy for 140.26, yet 40 minutes later, in F74, he placed a Large Order to buy 200 lots for 140.39, then increasing the price to 140.40.  That Large Order was live for 12.982 seconds and was cancelled less than two seconds after the Small Order filled. 

(3)        Of the RFQs identified by Mr Lopez in support of his view of client buying interest, the largest, or rather, that which would require hedging with more than 200 lots of Futures, was that from Banca d’Italia which had been received at 10.10.53, ie much earlier that day.  That was the order he was sizing for; yet he had missed this.

(4)        Mr Lopez explained that he bought the bonds at 12.04.05 and 12.04.13 on his own initiative as anticipatory positions, as with his subsequent purchases at 12.05.29 and 12.06.29.  He thus had a view of the market and pricing levels, and executed trades to reflect that view; he took a position.  These are smaller transactions; but they show his risk appetite and his willingness to execute a trade based on his view of the market and client buying interest.

Lone Large Orders placed by Mr Sheth

737. Mr Sheth addressed some of his Lone Large Orders in his witness statement. 

738. On 21 June 2016, Mr Sheth placed a large order to sell 500 lots for 141.10 at 16.56.04, which he cancelled seven seconds later.  Then, 49 seconds later, he placed another large order to sell 500 lots for 141.09, which was also live for seven seconds before being cancelled.  In his witness statement he identified an RFQ from Method Investments which had been received at 16.56.50 for a sell-side order of €1.5m of a six-year bond and suggested this might have prompted him to deploy the Information Discovery Strategy, and that the sales team might have had advance notice of this RFQ (hence why the Futures orders preceded the timestamp on the RFQ). 

739. This RFQ had traded away.  Method Investments subsequently came back with another RFQ on the same side.  Mr Shivji submitted that this was at the end of the day, and it might be hard to reverse out of the position without making a loss.

740. Mr Sheth identified a large order to buy 500 lots at 16.45.37 on 28 June 2016 (which is the same date as F158).  The order was live for two seconds.  Having reviewed the data, he believed two events may have triggered him to use the Information Discovery Strategy:

(1)        An RFQ for a €1.3m buy transaction with Method Investments, where he suspected that they were trading in larger sizes away from MHI and their full position was greater.

(2)        A €10m buy side cash trade with ICBC.  The timestamp for this cash trade was 28 minutes later, but this was a voice trade so may have been agreed at the time of the instance.

741. Mr Shivji submitted that this was conjecture about the long delay.  The RFQ from Method Investments was for a May 2021 bond, ie with less than five years’ maturity, which had a Futures equivalent of about two lots - the large order was 250 times this size, so he would have been assuming Method Investments was splitting €325m elsewhere.  Mr Shivji emphasised that the order was live for two seconds, and that this cannot have been long enough to test whether Method Investments was splitting its order.  Mr Sheth suggested in cross-examination that he may have cancelled the order after two seconds if his cash hedge had disappeared. 

742. On 30 June 2016 Mr Sheth placed a large order at 12.42.55 to buy 500 lots for 142.28, cancelled the order after 24 seconds and then placed another large order for 142.30 and subsequently cancelled that order after ten seconds.  He identified an RFQ which had been received by the Desk at 12.39.24, from Banca IMI SPA for a quote for a €500,000 buy-side order of a 30-year bond and said that the length of this bond meant it would require nearly 500 lots to hedge.  The RFQ is noted as “dealer time out”, ie MHI did not provide a price.  Mr Sheth accepted in cross-examination that this bond would in fact only require about ten lots to hedge. 

743. For the Lone Large Orders placed by both Mr Lopez and Mr Sheth, the Tribunal agrees with Mr Shivji that the explanations are no better than those for the Large Orders in the Instance Pool.  However, this does not necessarily operate in favour of the Authority - a poorly thought through and applied strategy is not abusive if there was an intention to trade.  What is more significant, at least for Mr Lopez, is the number of the Lone Large Orders.  We agree with the Authority that only three of the five Lone Large Orders were not placed in the same second as a small order being filled, and that he is therefore relying on the existence of three out of 88 large orders to support his position that they were not placed to facilitate the execution of a small order on the other side.  Mr Sheth placed a greater number, 14 Lone Large Orders out of 202 large orders (bearing in mind that these 202 large orders include all of the Multiple Large Orders). 

Small Orders which overlapped with (and on same side as) Large Orders

744. Mr Kasapis relied on the Overlapping Small Orders as a reason why the Large Orders were unlikely to have been placed with the intention of moving the market towards the Small Orders, as in doing so the Traders would have also moved the market away from these Overlapping Small Orders, making these orders less likely to trade. 

745. Mr Shivji submitted that many of these were very far from the touch and placed long before the relevant Trading Instance started - in other words, what he termed “legacy” orders that were not a priority for the relevant Trader.  Mr Shivji criticised Mr Kasapis for not having looked at the price levels of these orders in his analysis, submitting that the price and timing was fundamental to analysing the relevance of these orders - if they were not a priority for the Traders at the time of the Instance, there is no reason at all why they should be illuminating when analysing their behaviour during the Instance.

746. Whilst Mr Shivji criticised the exercise undertaken by Mr Kasapis, the Authority has not provided any additional analysis to support its description of the Overlapping Small Orders as being “legacy” orders.  We agree with Mr Kasapis that these have potential to be significant, and we have sought to assess what evidence we do have about those orders listed by Mr Kasapis in his report.     

747. The Tribunal identified that the check said to have been done by Mr Kasapis was whether the Overlapping Small Order was live whilst the Small Order was live on the opposite side of the book.  The Tribunal considered it more relevant whether the Overlapping Small Order was live whilst the Large Order relied upon by the Authority was live on the same side of the book, on the basis that the Authority’s argument was that the Large Order would be pushing the market away from that Overlapping Small Order (and that this was the intended effect of the Large Order).

748. The Overlapping Small Orders appear in a number of Instances and occur throughout the Relevant Period (from 10 June to 27 July 2016) and were placed by all three Traders, including those where the Trader placing the Large Order in an Instance had placed the order that became the Overlapping Small Order in the relevant Instance. 

F27 at 10.15.48.236 on 10 June 2016

749. Mr Urra and Mr Lopez both placed Small Orders to sell 15 lots each at 141.78.  Mr Urra placed a Large Order to buy 444 lots for 141.72, the price of which he amended. 

750.  There was an Overlapping Small Order to buy 40 lots, uniceberged.  This had been placed by Mr Lopez at 10.08.51 for 141.70, the price of which he had then reduced to 141.69 at 10.09.10.  The order sat at that price, away from the touch, throughout F27, but afterwards, at 10.23.37, Mr Lopez reduced the price to 141.68, it immediately started to fill and had filled at that price by 10.23.42. 

F40 at 14.16.34.477 on 13 June 2016

751. Mr Sheth placed a Small Order to sell 40 lots for 141.22 which started to trade.  Mr Lopez then placed a Large Order to buy 200 lots for 141.19.

752. The Overlapping Small Order was an order to buy 40 lots which had been placed by Mr Urra at 14.08.49.  It was at Best Bid at the time the Large Order was placed, although in the middle of the queue at that price point.  It filled at that price in one go at 14.50.23.

F48 at 11.01.18.775 on 15 June 2016

753. The Large Order was placed first, and it was an order placed by Mr Urra to buy 490 lots.  Mr Lopez then placed the Small Order to sell ten lots. 

754. The Overlapping Small Order was an order to buy five lots, iceberged to two, placed by Mr Lopez, for 140.60 which had been placed at 10.50.53.  Mr Lopez had amended the price at 11.00.11 to 140.62, suggesting that even shortly before the Instance it was still being given attention.  It was seven ticks from the touch at the time the Large Order was placed (whereas the Large Order was two ticks from the touch).  The order was cancelled at 11.10.06. 

F83 at 11.15.29.662 on 20 June 2016

755. Mr Sheth placed a Small Order to sell 20 lots for 141.27 and a Large Order to buy 500 lots for 141.23. 

756. The Overlapping Small Order was amended during the course of the Instance.  The order had been placed by Mr Lopez at 11.15.06, to buy 255 lots for 141.24.  The amendment was to the number of lots, not price, reducing the size to 155 lots.  At 11.16.49 Mr Lopez increased the price of that order to 141.29 and it started to trade.  84 lots filled, and at 11.17.06 Mr Lopez cancelled the remaining 71 lots.

757.   At the time the Large Order was placed, that Overlapping Small Order was one tick from the touch, albeit at the back of the queue, with more than 75 visible lots ahead of it in the queue at that price point. 

F106 at 10.03.19.849 on 22 June 2016

758. Mr Lopez placed a Small Order to sell 15 lots, iceberged to three.  This order is at the touch.  He then placed a Large Order to buy 200 lots, three ticks from the touch.

759. Mr Kasapis identified two Overlapping Small Orders in this Instance.

(1)        The order that is visible in the Replay graph was placed by Mr Sheth.  It was an order to buy five lots, uniceberged, that had been placed at 10.02.36 for 141.3, which was the same price as the Large Order.  It was ahead of the Large Order in the queue (although there were about 90 lots visible ahead of that Overlapping Small Order).  It filled at that price at 10.07.40. 

(2)        There was a second order, which had not been visible.  That had been placed by Mr Lopez at 9.38.07 to buy five lots, iceberged to two, at 141.1.  There were two price amendments within the next second, to 141.05 and then to 141.  He cancelled the order at 12.23.57. 

760. We accept this second Overlapping Small Order does not appear to have been a priority to Mr Lopez, based on the timing at which it was placed, that it was left outstanding and cancelled nearly three hours later.

F181 at 11.14.07.730 on 1 July 2016

761. Mr Sheth placed a Small Order to sell 20 lots, and two Large Orders to buy 250 lots. 

762. At the time Mr Sheth placed the two Large Orders, there was an Overlapping Small Order on the stack, two ticks from the touch.  This is the price point at which Mr Sheth had placed his second Large Order.  That order had been placed by Mr Lopez at 10.58.27 for 143.82, amended to 143.8 at 10.58.30, and was for two lots.  It traded almost three minutes after the two Large Orders were cancelled.

F203 at 12.36.16.793 on 19 July 2016

763. Mr Urra placed a Small Order to sell eight lots and a Large Order to buy 250 lots.  At the time he placed the Large Order (two ticks from the touch), his Small Order was at Best Offer.  The Overlapping Small Order had been placed by Mr Lopez at 12.33.08 and was five lots for 143.82.  It was one tick from the touch at the time the Large Order was placed, and was at the front of the queue at that price point.   

764. Mr Lopez amended the price after F203, at 12.40.37, increasing it to 143.84 and it filled at 12.40.42.  The amendment and the fill occurred within the time of F204 (but in that Instance the Large Orders were on the sell-side).

F222 at 11.19.50.290 on 27 July 2016

765. Mr Lopez placed a Small Order to sell five lots, then a Small Order to sell two lots, both at Best Offer of 144.04.  Mr Urra placed a Large Order to buy 490 lots for 144.

766. The Overlapping Small Order was to buy five lots and had been placed by Mr Lopez at 11.00.14 for 144.02, reduced to 143.9 at 11.00.33.  The order sat there at that price during the Instance (and was about 13 ticks from the touch at the time at which Mr Urra placed the Large Order), and at 11.29.20 Mr Lopez increased the price to 143.99, and it filled at 11.29.43.

Overlapping Small Orders that did not overlap with Large Order

767. Some of the Overlapping Small Orders identified by Mr Kasapis did not overlap with the Large Order in the relevant Instance.  We consider this is significant, and means that the placing of the Large Order cannot be said to have (unintentionally or deliberately) been pushing the market away from such order.  For this reason we consider that the existence of these Overlapping Small Orders is entirely neutral. 

768. This can be illustrated by the following Overlapping Small Orders:

(1)        F109 at 11.07.20.353 on 22 June 2016 - The order number identified by Mr Kasapis is that of Mr Lopez’s small sell order which had filled before the Large Orders were placed.

(2)        F180 at 8.42.46.542 on 1 July 2016 - The order number identified by Mr Kasapis was a buy order placed by Mr Urra that had filled before he then placed a Large Order to buy 444 lots.

(3)        F204 at 12.40.33.513 on 19 July 2016 - This involved Large Orders placed by Mr Urra and Mr Lopez to sell 444 lots and 200 lots respectively.  The order number identified by Mr Kasapis was that of Mr Urra’s Small Order in F203 which had filled before these Large Orders were placed.

769. These orders had been identified as “Overlapping Small Orders” as they overlapped with the Small Order in the Instance; but they did not overlap with the Large Order.

Other Overlapping Small Orders

770. Having considered the details of the remaining orders identified by Mr Kasapis in other Instances and looked at the timing at which they were placed, distance from the touch during the Instance and subsequent actions of the Trader who placed the relevant order, we agree with the Authority that the remaining Overlapping Small Orders can be described as not illuminating on the basis that they were legacy orders or not a priority and were unlikely to trade (irrespective of any impact of the Large Order). 

771. This conclusion can be illustrated by the following Instances:

(1)        F32 at 18.11.59.427 on 10 June 2016 - Mr Urra placed a Small Order to sell nine lots for 141.94, a Large Order to buy 450 lots, the Small Order traded, he amended the price of the Large Order and then cancelled the Large Order.  The Large Order was placed at 18.12.07.402 for a price of 141.91.  The Overlapping Small Order is visible on the stack at that time.  This order was put to Mr Creaturo by Mr George in cross-examination, who had said it is irrelevant as it is about ten ticks away from the touch.  (We count that it was 13 ticks away.)  The order was an order to buy five lots and had been placed by Mr Lopez at 17.53.19.213.  He had amended the order at 18.07.and the order was then cancelled at 18.14.54.

(2)        F47 at 10.55.35.870 on 15 June 2016 - This was not identified by Mr Kasapis, but this was a Specified Instance and the Overlapping Small Order was visible in the Replay graph.  (It is the only Instance to which we refer which was not on Mr Kasapis’s list.)  At the time Mr Urra placed his Large Orders, there was an Overlapping Small Order on the stack.  We do not know which Trader placed this order, and it is 16 ticks away from the touch but had a visible size of around 30 lots. 

(3)        F53 at 16.38.46.120 and F54 at 16.42.28.541 on 15 June 2016 - Mr Urra placed a Small Order to sell ten lots and a Large Order to buy 400 lots.  At the time the Large Order was placed, as well as the Small Order on the stack, there was another MHI sell-side order with a visible size of about 20 lots that was 13 ticks from the touch at 140.83.  In addition, there was an Overlapping Small Order which was an order to buy ten lots which had been placed by Mr Sheth at 16.21.52.  These orders were both visible on the stack in F54, which also involved a Small Order to sell ten lots and a Large Order to buy 400 lots being placed by Mr Urra.  Mr Sheth cancelled the Overlapping Small Order at 16.55.39.    

(4)        F166 at 17.10.59.129 on 28 June 2016 - Mr Sheth placed a Small Order to sell 30 lots and two Large Orders to buy 500 lots.  There was an Overlapping Small Order on the buy-side for 141.46.  It had been placed by Mr Sheth at 16.58.10 and was an order to buy ten lots.  It was 15 ticks from the touch at the time the Large Order is placed.  He cancelled it at 17.24.03.

(5)        F168, F169 and F170 from 9.17.18.332 to 9.28.19.828 on 29 June 2016 - These Instances were close in time and involve all three Traders at different points.  All of the Large Orders in these Instances were on the buy-side:

(a)        In F168 Mr Sheth placed a Small Order to sell 12 lots and a Large Order to buy 500 lots, and Mr Lopez placed a Large Order to buy 200 lots. 

(b)        In F169 Mr Sheth placed a Small Order to sell 23 lots and a Large Order to buy 500 lots. 

(c)                   In F170 Mr Urra placed a Small Order to sell 19 lots, then a Large Order to buy 444 lots which he then cancelled (before the Small Order had traded).  Mr Sheth placed a Small Order to sell five lots, then a Large Order to buy 500 lots, after which Mr Urra placed a second Large Order to buy 444 lots.  Mr Lopez had placed an Overlapping Small Order to buy one lot at 9.10.47 for 142.34, which was then amended to 142.3 at 9.11.36.  It was at that price during F168 and F169.  Mr Lopez cancelled that order during the course of F170.  The order had been about 17 ticks from the touch when Mr Urra placed that Large Order (which had been just over 30 seconds before Mr Lopez cancelled the one lot order).

772. The distance of the Overlapping Small Orders from the touch in these Instances means that we consider that these orders were very unlikely to trade at that price around the time of the Instance (such that they cannot be said to have been prejudiced by the placing of the Large Order). 

Conclusions on the Overlapping Small Orders

773. Whilst we did not find the reference to the mere existence of these orders helpful or supportive of Mr Kasapis’s opinion, we were also not persuaded by Mr Shivji’s description of all of them as legacy orders or not illuminating without additional information.

774. We have identified eight of them (F27, F40, F48, F83, F106, F181, F203 and F222) which do not appear to be legacy orders, whether because they had only been placed or amended shortly before the Instance, they were close to the touch, or were amended after the Instance, suggesting that the Trader who had placed the relevant order still wanted it to trade.  We accept that the Overlapping Small Orders may not have been an immediate priority at the time of the relevant Instance.  Some of these orders do show that where a Trader was prepared to wait they could, in the right conditions, still trade these orders without improving the price, eg the first Overlapping Small Order in F106 and in F181.

775. We considered F40 to be interesting.  The Authority’s position is that Mr Lopez had placed a Large Order to facilitate the execution of Mr Sheth’s Small Order of 40 lots (some of which had already traded) whilst at the same time prejudicing the likelihood of his manager’s order of 40 lots from trading (or at least trading at the price at which it had been entered).  Mr Urra’s Overlapping Small Order did in fact trade at that price after the Instance.

Conclusions on Market Abuse

776. The Authority alleges that the Traders engaged in an abusive trading strategy to commit market abuse, which involved them placing Large Orders that gave or were likely to give a false or misleading impression and/or signal as to the supply of or demand for Futures; and they did not intend to trade these Large Orders.  The Authority alleges that this conduct was deliberate, dishonest and lacked integrity.

777. We have set out the relevant law, including the focus in Burford on the subjective intentions of the relevant trader.  The law relating to market abuse does not contain any requirement that the conduct be dishonest.  In the Decision Notices the Authority’s conclusion was that the Traders had deliberately engaged in market abuse, and this conduct was dishonest and lacked integrity, and this has been the Authority’s pleaded case throughout these references.  The Tribunal therefore reaches conclusions not only on whether each Trader committed market abuse, but also whether they were dishonest by reference to the test in Ivey, namely ascertaining (subjectively) the actual state of each Trader’s knowledge or belief as to the facts; and determining whether their conduct was honest or dishonest by applying (objective) ordinary standards.

778. We have reached our conclusions on the basis of all of our findings of fact and our evaluation of all of the evidence and submissions before us.  We have not found it necessary to refer to all of the evidence or all of our evaluation of that evidence in setting out our conclusions.  The Tribunal is mindful of the seriousness of the allegations which have been made by the Authority and the consequences for the Traders of its decision; and that the burden of proof is on the Authority.

779. Before setting out our conclusions in respect of each of Mr Urra, Mr Sheth and Mr Lopez (in that order for this purpose), we consider it important to reiterate the following:

(1)        The Authority’s pleaded case is in relation to the placing of the Large Orders in the Instance Pool; those Large Orders are said to have been placed to facilitate the execution of the Small Orders.  The Traders have criticised the Authority throughout for not taking account of their full trading activity in the Relevant Period, ie the activity outside the Instance Pool which includes the placing of Non-Instance large orders (including the Lone Large Orders).  This submission needs to be considered in context.  The sizing used for the categorisation of orders as being large or small (or “Large” or “Small” if within the Instance Pool) means that, when taken together, they capture all of the orders placed by the Traders during the Relevant Period.  To put this another way, there is no category of “medium-sized” orders in Futures for this purpose. 

(2)        The Large Orders and the Non-Instance large orders are all the orders of 200 lots or more placed by Mr Lopez, or 250 lots or more placed by Mr Urra and Mr Sheth (except where they placed orders in Multi Trader Instances in which case the lower threshold of 200 lots was applied) in the Relevant Period.  Where we have referred to the execution rates of such orders placed by the Traders in comparison with those of other market participants, where other market participants placed 2,406 large orders of 200 lots or more, 72.28% of which fully traded, this is a comparison being drawn with all of the Traders’ orders of such size in the Relevant Period.  Only six of Mr Urra’s 175 large orders traded (five in part and one in full); and none of Mr Lopez’s 88 large orders or Mr Sheth’s 202 large orders traded at all. 

(3)        The small orders (including the Small Orders) placed by all three Traders were sized by reference to the market making business which the Desk was actually doing.  They were (partial) hedges for cash trades (recognising that Futures were not a perfect offsetting hedge to a cash trade) and/or re-positioning the Traders’ books. 

(4)        We have concluded that the Large Orders were each likely to give the impression or signal of significantly increased supply or demand, and the most likely market reaction would be for the market to move in the opposite direction to the Large Orders, ie towards the Small Orders.  We found that a Trader placing the Large Order in these circumstances would have known of this likelihood.

(5)        On the basis of the trading activity in the Instances, in particular the overlap between the Small Orders and the Large Orders, the length of time for which the Large Orders were live and the timing of the cancellation of the Large Orders after the Small Orders filled, the Tribunal accepts that it could reasonably be inferred that the Large Orders were placed to facilitate the execution of the Small Orders. 

(6)        However, there are aspects of the trading activity during the Relevant Period which the Authority’s case does not explain, including the Lone Large Orders and the Overlapping Small Orders.  This activity could be consistent with the Trading Strategies as put forward by the Traders.

(7)        We have throughout sought to assess the plausibility of the explanations which have been put forward by Mr Urra, Mr Sheth and Mr Lopez and the Authority by reference to all of the evidence which is available and taking account of the information that is no longer available.

Mr Urra

780. Mr Urra’s evidence was that all of his large orders, including the Large Orders, were placed in pursuit of the Information Discovery Strategy.

781. The Tribunal had some difficulties with accepting the credibility of Mr Urra’s evidence as a witness.  We considered that his explanations in relation to the questions raised by the Eurex Letter (and in particular his insistence that it was about the Multiple Large Orders which had been placed by Mr Sheth) and his account of the steps he took immediately after receiving it (namely speaking to Mr Sheth but being reluctant to accept that he would have spoken to Mr Lopez, whose trading activity had also been expressly identified in the Eurex Letter albeit by trader code rather than by name) were not credible. 

782. The Tribunal has concluded that the Information Discovery Strategy put forward by Mr Urra as the explanation for the Large Orders was not plausible.  We do not accept that either of the two “benefits” of the strategy were likely to be achieved.  Mr Urra was an experienced and successful trader.  For this reason we are satisfied that Mr Urra knew this, and do not accept that Mr Urra would have designed such a strategy, pursued it himself or encouraged others to pursue it.  We considered it notable that the evidence before the Tribunal of the two occasions on which Mr Urra had had “success” in his placing of large orders, namely where they traded in full, was provided by Mr Sheth (in respect of the large order for 450 lots in May 2016) and Mr Kasapis (in respect of the Non-Instance large order) yet Mr Urra gave no evidence of the information which had been discovered on either of these occasions. 

783. Furthermore, we are not persuaded that such a strategy, even if it were being pursued, could explain his trading activity in the Instance Pool.  By way of example:  

(1)        The Large Orders were often live for just a few seconds, with insufficient time to test any theory.  They were routinely cancelled shortly after the Small Order filled (not just in the Specified Instances but throughout the Instance Pool).

(2)        Mr Urra placed concurrent Large Orders which could not be explained by such a strategy (as only one Large Order of sufficient visible size would, on Mr Urra’s case, be required to test whether a Posited Trader was keen to transact in size and prepared to pay a premium to do so).  

(3)        Mr Urra switched sides when placing his Large Orders in relatively short periods of time, eg in F15. 

(4)        Mr Urra’s management of the Large Orders supports a conclusion that he did not intend them to trade - for the three Large Orders that did start to trade (F16, F30 and F167), they were either cancelled immediately or moved away from the touch.  In F30, when the part fill resulted in him having bought 39 lots, he then placed an order to sell those lots.  In addition, Mr Urra made multiple price amendments to the Large Orders, towards the touch and away again, eg in F31 and F82; and in F47 he cancelled the Large Order when it was at the touch. 

(5)        Mr Urra placed Large Orders late in the day on 10 June 2016 (in F30, F31 and F32) in circumstances where he cannot reasonably have expected that another market maker had traded in sufficient size that they would need to hedge their position in Futures and be prepared to pay a premium to do so.

784. The timing of the activity in the Multi Trader Instances strongly supports a conclusion of collaboration with both Mr Lopez and Mr Sheth.  This could clearly be seen in F30 and F31 (with Mr Lopez) where in F30 Mr Urra moved the price of his Large Order away from the touch at almost exactly the same time as Mr Lopez’s Small Order filled, and in F31 he and Mr Lopez (who were, on their accounts, pursuing different strategies) had Large Orders in the market in the same direction at the same time, cancelling them 0.135 seconds apart.  In F84, the two Small Orders (placed by Mr Lopez and Mr Sheth) were hedges for Mr Urra’s book, and they were selling a total of 255 lots (in two Small Orders, both of which were iceberged).  Mr Urra’s Large Order was one tick away from Best Bid, was live for 11.536 seconds and cancelled 1.221 seconds after the Small Orders filled.

785. Mr Bailin submitted that the placing of Multiple Large Orders in Multi Trader Instances was supportive of the Traders’ position that they were not pursuing an abusive strategy, as those orders only served to draw attention to the trading and, if the Traders were collaborating, would have been seen by Mr Urra and Mr Lopez who would have told Mr Sheth of his error.  In F121, the Instance to which we were referred which involved Large Orders being placed by both Mr Sheth and Mr Urra, whilst Mr Sheth placed a total of six Large Orders during the course of the Instance, each of which were for 400 lots, only two were ever live concurrently.  Indeed, the activity in that Instance involved Mr Sheth placing a Small Order to buy 22 lots and then, after six seconds, a Large Order to sell 400 lots, which was then followed by Mr Urra placing a Large Order to sell 450 lots and Mr Sheth placing a second Large Order to sell 400 lots.  These three Large Orders were then cancelled (and Mr Urra’s trading did not feature again in this Instance).  Mr Sheth then amended the price of his Small Order and placed four further Large Orders later in that Instance.  Although we considered this to be an intriguing submission, ultimately we are not persuaded that it bears the weight which the Traders sought to attach thereto.  Mr Urra had also placed concurrent Large Orders, and two of Mr Sheth’s Large Orders being live concurrently was not inconsistent with Mr Urra’s own trading.

786. Whilst Mr George submitted that Mr Kasapis’s Market Trend Analysis showed that in 70% of Specified Instances the market was moving towards Mr Urra’s Small Order, and Mr Urra’s Small Order was already trading in some of the Instances (including the Specified Instances), such that there would be no need to take steps to facilitate the execution of these orders, let alone commit market abuse to do so, we are not persuaded that the position is as strong in favour of the timely execution of the Small Orders:

(1)        Where the Small Orders had started to trade, Mr Urra placed the Large Orders after such trading had paused for several seconds in F7, F150 and F194.  (In F7, the Large Order was live for less than two seconds, and cancelled 1.339 seconds after the Small Order traded.) 

(2)        The volumes on the stack show that it was not straightforward to get the Small Order done at the times they were placed.  In F12 (identified by Mr Kasapis as an Instance where the market was trending away from the Small Order) where Mr Urra was trying to buy 73 lots, he was at the back of the queue and there were only about 12 to 15 lots visible at the touch on the sell-side.  In F150 (identified as an Instance where the market was moving towards the Small Order) Mr Urra was trying to sell 90 lots (and it had started to fill) but there were only about 10 to 15 lots visible on the buy-side at the touch.

787. Mr Urra’s experience meant he would have known the potential significance of the decision as to whether (and to what level) to iceberg orders that he wanted to trade.  This was also evident from his trading activity.  His placing of the Small Orders shows the decisions he made - he placed Small Orders of 25, 73 and 90  lots that were not iceberged (in F232, F12, F150  respectively).  By contrast, in F31 he iceberged his Small Order of 39 lots to four, in F152 he iceberged his Small Order of 33 lots to nine and in F194 he iceberged a Small Order of 25 to four.  He adapted his actions to the market and his priorities at the relevant time and knew the likely effects of showing size to other market participants.

788. We do not consider that the three Instances (F32, F86 and F170) where Mr Urra amended the price of his Large Order after the Small Order filled are of particular assistance.  In F32 the price amendment was made 0.026 seconds after the Small Order filled, ie so close as to be regarded as the same time as the Small Order (which had been placed by Mr Urra) filled.  In F86 and F170 the price was amended 1.018 and 0.559 seconds after the Small Order filled.  In all three Instances, the Large Order remained live at the amended price for less than three seconds before being cancelled.

789. Mr George submitted that the Authority’s case was itself not plausible.  We have assessed these submissions above.  In particular, Mr George submitted that Mr Urra was a successful trader, who had recently been promoted to Head of European Rates Trading and would have no reason to pursue an abusive strategy.  We do not agree - Mr Urra was a successful, well-remunerated trader who was responsible for managing the Desk and improving performance, and we do not find it incomprehensible that he may have considered pursuing a strategy which involved committing market abuse to ensure he continued with his success.  This would particularly be the case if he considered his actions would be unlikely to be detected.  Here, the placing of Large Orders at the beginning of June 2016 without questions being raised by MHI or the Exchange could have given him comfort that Futures orders of 500 lots which were placed and cancelled without execution were not being flagged by the monitoring tools.

790. Furthermore, we were troubled by Mr Urra’s apparent lack of regard for the specific requirements of the Desk’s Mandate, notwithstanding that this Mandate was required by the Volcker Rule and in circumstances where, as Desk Head, he was responsible for the Desk’s adherence to the Mandate (and had signed it).  There can be no doubt that he would have known its importance.

791. The Tribunal does agree with Mr George that there are difficulties with the Authority’s case: 

(1)        The Authority’s position is that the Information Discovery Strategy is itself a fabrication, along with the discussions between Mr Urra and Mr Sheth recounted by Mr Sheth.  Yet not only is there no direct evidence of collaboration between any of the Traders to commit market abuse (in the form of electronic communications, or being overheard by Mr Hill or Mr Barouti), the Authority has not put forward a positive case as to how the Traders decided to commit market abuse, or how or when they came up with the explanations for the Trading Strategies that they have now put forward.  We have set out the timing of the receipt of the Eurex Letter, and the meetings of both Mr Urra and Mr Sheth with Compliance.  The meeting notes prepared by Compliance are not particularly clear and we concluded that they are not reliable.  However, we accept that they do appear to record similar explanations being given by both Mr Urra and Mr Sheth in the beginning of August 2016.    

(2)          The Lone Large Orders are not explained by the Authority’s case.  Mr Shivji did not accept that these Lone Large Orders were not abusive; his position was that the Authority was not putting forward a case that they were abusive.  We accept that these Lone Large Orders could be consistent with Mr Urra’s explanation of the Information Discovery Strategy.  However, we do not lose sight of the fact that we were not persuaded that such strategy explained Mr Urra’s trading activity in the Instance Pool.      

(3)        The occurrence of the Overlapping Small Orders raises a question as to why Mr Urra might have been operating against his own interests (or those of another Trader) by placing a Large Order which also pushed the market away from an Overlapping Small Order, or why another Trader would prejudice Mr Urra’s Overlapping Small Order.  Mr Urra placed Large Orders in F27, F48, F203 and F222 where there was an Overlapping Small Order on the same side, some (but not all) of which subsequently filled.  Furthermore, in F40 the Small Order was to sell 40 lots and had been placed by Mr Sheth, with a Large Order being placed by Mr Lopez (and that order was placed when Mr Sheth’s Small Order had already started to trade).  Mr Urra had an Overlapping Small Order to buy 40 lots which had been placed eight minutes before the start of the Instance and was (at the time the Large Order was placed) at Best Bid (although not at the front of the queue).  Mr Urra’s Overlapping Small Order was thus larger than the remaining size of Mr Sheth’s Small Order at the time that Mr Lopez placed the Large Order.  The Large Order would have been likely to prejudice the swift execution of that Overlapping Small Order; although it filled at that price point 30 minutes later. 

792. Assessing all of the evidence, the Tribunal has concluded that notwithstanding these difficulties with the Authority’s case, they are not fatal and they are heavily outweighed by the evidence in favour of the Authority.  The Tribunal concludes that the Authority has established that:

(1)        Mr Urra placed the Large Orders with the intention of facilitating the execution of the Small Orders and he did not have an intention to trade the Large Orders. 

(2)        The Large Orders were likely to give the impression or signal to other market participants of significantly increased supply or demand.  Mr Urra knew this and he knew that this would be likely to impact the trading activities of other market participants such that the most likely reaction would be for the market to move in the opposite direction to the Large Orders, ie towards the Small Orders on the opposite side of the order book. 

(3)        Mr Urra knew that the impression or signal given to other market participants in this situation was therefore false or misleading.  He intended to give such a false or misleading impression or signal.

793. The Tribunal concludes that this conduct amounted to market abuse within s118(1) FSMA 2000 and market manipulation within Article 15 of the Market Abuse Regulation.

794. On the basis of our conclusions as to Mr Urra’s actual knowledge and intentions, the Tribunal finds that his conduct was dishonest by ordinary standards.

Mr Sheth

795. Mr Sheth’s evidence was that his large orders were placed in pursuit of the Information Discovery Strategy which he had been shown by Mr Urra and that the Multiple Large Orders were a mistake (or poor practice).

796. In closing, when addressing issues relevant to collaboration between the Traders, Mr Shivji described there being, in the Authority’s submission, a “domino effect” amongst the Traders, which included that if we disbelieved Mr Urra then this would be a problem for Mr Sheth.  Mr Bailin submitted that for Mr Sheth the question is not about the plausibility of the Information Discovery Strategy, but about Mr Urra having had a profitable trade and having shared this with Mr Sheth, giving him permission to try the strategy.  Mr Sheth, he submitted, had no reason to believe this was anything but legitimate given that price discovery is a common market strategy and others at MHI had accepted that the Information Discovery Strategy was plausible.  Mr Sheth’s junior status was, Mr Bailin submitted, relevant to his state of mind. 

797. We have already referred (in the context of Mr Urra’s reference) to the Authority not having advanced a positive case as to how the Traders came up with their explanations for the Trading Strategies; and that the Authority’s position is that the Information Discovery Strategy is itself a fabrication, as is Mr Sheth’s account of his interactions with Mr Urra (both being shown the strategy and then subsequently expressing his frustration with it). 

798. The Tribunal accepts the Authority’s submission that Mr Sheth’s reference cannot be considered in isolation, and that the conclusions we have reached as to the Information Discovery Strategy not being plausible are also relevant to Mr Sheth.  However, the allegations made by the Authority require us to focus on the intentions of each Trader in placing the Large Orders, and the assessment of their honesty requires us to ascertain the actual state of their (subjective) knowledge.  This means that, whilst our conclusions as to the (lack of) plausibility of the strategy have the potential to be unhelpful to Mr Sheth, we do not accept that the conclusions reached in relation to Mr Urra are necessarily fatal to Mr Sheth.

799. We have found that:

(1)        Mr Sheth was the junior member of the Desk and was not an experienced industry professional. 

(2)        Mr Sheth had been told by Mr Urra that he, Mr Urra, had placed an order to buy 450 lots in May 2016 and that this had filled. 

800. This is not a finding that Mr Urra had told Mr Sheth of the Information Discovery Strategy at that time, or that Mr Sheth had been given permission to pursue it.  The only evidence we have as to the placing of this order by Mr Urra is the witness evidence of Mr Urra and Mr Sheth (where Mr Sheth’s witness statement included a copy of the trading activity for this order), and it is only Mr Sheth who said he had any specific recollection of these discussions (with Mr Urra accepting that such a conversation as recalled by Mr Sheth may have been more significant to Mr Sheth than it had been to Mr Urra).   

801. On the basis of the timing of Mr Urra’s and Mr Sheth’s involvement in the trading activity in the Instances, Mr Urra’s considerable experience and Mr Urra’s position as Desk Head and Mr Sheth’s manager, the Tribunal finds that Mr Urra was the driving force behind the activity being pursued by Mr Sheth, whether that was the pursuit of a legitimate trading strategy or an abusive scheme. 

802. The Tribunal has concluded that the Information Discovery Strategy was not plausible and that neither of its two “benefits” were likely to be achieved.  Mr Bailin submitted that others at MHI had, however, considered the strategy to be plausible, so why shouldn’t Mr Sheth.  We accept in principle that this may be relevant to Mr Sheth’s knowledge and understanding in the Relevant Period, and thus his intentions when placing the Large Orders but it is important to look further at what was actually said and understood by others.   

803. The opinions expressed by Mr Hill, Mr Heiberg and Mr Joshi about the Information Discovery Strategy included the following:

(1)        Mr Hill referred to showing good size on a less liquid contract and accepted there may be someone “desperate to trade” who would access that liquidity.  Mr Hill gave examples of maybe putting 100, 500 or 50 lots in, saying “it would always be different, depending on what you were doing at that particular time”.

(2)        Mr Heiberg said that building a basis position, hoping to be hit at a lower level if someone needs the liquidity, makes sense.  He also said that he was troubled by the frequency and extent of this activity.  Mr Heiberg was clear that he would not have approved this strategy being deployed whilst there were orders in the opposite direction.

(3)        Mr Joshi referred to “price discovery”, and being ready to give quotes and being more dynamic on the Futures side.

804. We have already concluded that these opinions were expressed in response to what seems to the Tribunal to have been a rather vague description or understanding at the relevant time of the explanation which was being put forward by the Traders.  Mr Hill certainly did not know the details of Mr Sheth’s trading activity (in particular the repeated use of 500 lots, with 250 or 300 lots only being used a few times and with no explanation of why - and certainly not by reference to a consideration of the time of day, possible counterparties or size of the cash trade being done).  Whilst Mr Heiberg and Mr Joshi would have been aware of more details related to the pattern of trading activity and the overlap with orders on the opposite side of the book, they did not know (from Mr Sheth’s interviews with Compliance or HR) of the identify of clients or the size of RFQs against which Mr Sheth said he was testing the theory of clients splitting the orders they were showing to MHI.  Furthermore, it is not clear whether, at the time at which Mr Heiberg and Mr Joshi were interviewed by the Authority in 2017, they were aware of the number of Instances now relied upon by the Authority. 

805. We are not persuaded that we should place weight on the opinions of Mr Hill, Mr Heiberg or Mr Joshi, and certainly not such as to support a conclusion that Mr Sheth might have had a subjective belief that the Information Discovery Strategy was plausible and legitimate. 

806. On Mr Sheth’s case, the Information Discovery Strategy had been shown to him by Mr Urra, but even if it appeared to make sense initially, or he was prepared to assume it must as it was being shown to him by his manager, we cannot ignore the reality that Mr Sheth knew the specifics of the decisions he was making and how he was implementing this strategy, which included the size of RFQs and identification of clients, the significant extent of the overlap with his own orders on the other side of the book (as well as those of other Traders), the time period for which the Large Orders were live (including many which were live for less than three seconds) and the proximity in time of the filling of the Small Order and the cancellation of the Large Order.  Mr Sheth did not have the level of experience of Mr Urra and Mr Lopez, but he was well-educated in financial risk, understood and could analyse data, had been working on the Desk since September 2014, had received training at MHI on market manipulation and market abuse, understood what spoofing was and he knew that spoofing (and placing any orders without an intention to trade) constituted market abuse.

807. In the context of assessing not only whether the Information Discovery Strategy had existed in the Relevant Period, but also whether Mr Sheth might have believed that he was pursuing it legitimately, the Tribunal identifies two further difficulties:

(1)        We regard it as significant that Mr Sheth did not keep any written records during the Relevant Period which would help to illustrate what he was learning or not learning from the relevant activity, eg that there was no interest in the Large Orders where they were placed in proximity of an RFQ received from ICBC, that he hadn’t tested against a particular client yet, that he had placed a large order but cancelled it before reaching a view as a result of market movements and his cash hedge disappearing.  He would have wanted this information for himself to improve his use of the strategy, and, if it had been shown to him by his manager, to report back to Mr Urra.  Such records may have been incomplete, but we would expect to see something.

(2)        We are not persuaded that Mr Sheth’s trading activity can be explained by the Information Discovery Strategy.  By way of illustration:

(a)        Mr Sheth placed Large Orders where the Desk had not received an RFQ, eg in F174 where Mr Lopez’s Small Order was a hedge for a cash bond Mr Lopez had bought on MTS on his own initiative.

(b)        Mr Sheth placed Large Orders in opposite directions within an Instance, switching sides in F158 where he placed two Large Orders to buy 500 lots, cancelled them and then placed two Large Orders to sell 500 lots.  For both pairs of Large Orders, the order closest to the touch was cancelled within two seconds of the Small Order being filled.

(c)        Mr Sheth placed more than one Large Order in some Instances, not all of which fall within his explanation for his Multiple Large Orders.  In F176 Mr Sheth cancelled his sixth Large Order to buy 500 lots at 15.31.16.40, less than three seconds after his Small Order of 80 lots filled.  He then placed another Large Order to buy 500 lots at 15.31.37.500.  That seventh Large Order was placed almost four seconds after he had placed a second Small Order to sell ten lots, and was cancelled 1.73 seconds after the second Small Order filled.

808. Mr Sheth had been given considerable autonomy in his trading, eg if he was hedging for another Trader’s book he would decide the size and price of the order.  Mr Sheth’s trading activity showed:

(1)        He took decisions as to whether or not to show the full size of his orders.  In the Specified Instances, Mr Sheth did not usually iceberg his Small Orders.  In the ten Specified Instances which were Single Trader Instances he placed Small Orders of five, six, ten, 15, 17 and 25 lots, none of which were iceberged.  He did iceberg his Small Order of 80 lots in F174, to show 40 lots.  In the Multi Trader Instances that were Specified Instances he did not iceberg his Small Orders of 10, 20 or 22 lots.  He iceberged his Small Order of 200 lots in F84 to show 25 lots - he and Mr Lopez were both hedging for Mr Urra’s book and Mr Lopez placed a Small Order at the same price to sell 55 lots, which he iceberged to six.  We infer that Mr Sheth knew the potential significance of showing size to the market, and made his decisions accordingly.  He took a different approach to Mr Lopez (whose orders were typically iceberged in small slices), even when they were both hedging at the same time for Mr Urra. 

(2)        He would have seen from his own trading activity (even if it was not otherwise obvious) that the Small Orders would not simply trade just because they were placed at the touch: 

(a)        In F84 Mr Sheth’s Small Order of 200 lots was iceberged to 25 and was placed at the touch.  It started to fill, and had a part fill of 20 lots in one go, but many of the fills were for just one or two lots at a time.  It took just over 20 seconds to fill. 

(b)        In F176 the Small Order of 80 lots was placed at the touch, iceberged to 40, but the market immediately moved away from him; by the time he placed his first Large Order nearly seven seconds later the Small Order was two ticks from the touch (and not at the front of the queue).

809. We consider that, viewed in the light of Mr Sheth’s training and trading activity, Mr Sheth’s junior status does not support a conclusion that Mr Sheth did not understand the likely market impact of placing the Large Orders.  The Tribunal concludes that Mr Sheth knew what he was doing when placing the Large Orders.   

810. We need to address the Multiple Large Orders placed by Mr Sheth.  The Authority’s position was that these were each placed to increase the pressure on the order book, in the same way that Mr Urra had sometimes placed concurrent Large Orders, or two Traders would place concurrent Large Orders in an Instance.  We have already stated that we are not satisfied that the Multiple Large Orders can be described as a “mistake”.  Mr Sheth knew the difference between amending an order and placing a new order; each Large Order was placed with the knowledge and intention that it would be live at the new price; and Mr Sheth knew whether or not he had cancelled the first order (illustrated by the fact that he did sometimes cancel a Large Order before placing another Large Order at a better price). 

811. If the Multiple Large Orders had only appeared in a very small number of Instances, or did not involve more than two Large Orders being live concurrently, the Tribunal would have concluded that Mr Sheth was intending to show the concurrent sizes of those Large Orders (in the same way that Mr Urra did in other Instances), even though we recognise that Mr Sheth was placing these orders at different (improved) prices.  However, having considered all of the Multiple Large Orders in the Specified Instances, and identified others in the Instance Pool, the Tribunal has concluded that it was inherently unlikely that Mr Sheth would have wanted to show size of, eg, 2,000 lots in F176.  This  volume is significantly higher than Mr Sheth had placed in other Instances; and there is no pattern that was apparent to the Tribunal of these larger volumes being used at particular times of day or when liquidity at or close to the touch was relatively high.  This leads the Tribunal to conclude that Mr Sheth placed each Large Order for the same reason (whether that be pursuant to a legitimate trading strategy or an abusive scheme), but at the point at which he placed subsequent concurrent Large Orders he no longer intended to trade his earlier Large Order(s).  His poor practice was thus placing the subsequent Large Order without cancelling the first, or failing simply to amend the price of his first Large Order.

812. The Tribunal has considered what Mr Bailin submitted was the implausibility of the Authority’s case, including:

(1)        There is no direct evidence of collaboration to commit market abuse; and in explaining their Trading Strategies, the three Traders have put forward two different explanations.  Mr Bailin submitted that if three Traders were conducting an abusive scheme, it is inherently unlikely that they would not also have devised a single explanation.  The Tribunal agrees that this is an unusual feature of the explanations.  The Authority has not put forward a positive case on this.

(2)        Mr Sheth had placed 14 Lone Large Orders.  Mr Sheth’s explanations for those he addressed in his witness statement are no more persuasive than the explanations he provided for the Large Orders in the Instance Pool; but we agree that they are also unexplained by the Authority.  At least one of the Large Orders in the Instance Pool, the second Large Order in F151, did not overlap with a Small Order.

(3)        The average size of the Small Orders placed by Mr Sheth in the Single Trader Instances that were Specified Instances was 17.5 lots.  Crossing the spread would have saved on average €175.  The largest Small Order in the Specified Instances was 80 lots, so crossing the spread would have cost €800.  We agree that, viewing each Instance in isolation, the amounts involved are relatively small.  However, the Authority is alleging that the Traders committed market abuse in 233 Instances, 104 of which involved Mr Sheth’s trading activity.   

813. Assessing all of the evidence, and notwithstanding these matters relied upon by Mr Bailin including Mr Sheth’s junior status, the Tribunal concludes that the Authority has established that:

(1)        Mr Sheth placed the Large Orders with the intention of facilitating the execution of the Small Orders and he did not have an intention to trade the Large Orders.  At the time at which he placed them, his Multiple Large Orders were placed with this same intention.

(2)        The Large Orders were likely to give the impression or signal to other market participants of significantly increased supply or demand.  Mr Sheth knew this and he knew that this would be likely to impact the trading activities of other market participants such that the most likely reaction would be for the market to move in the opposite direction to the Large Orders, ie towards the Small Orders on the opposite side of the order book.    

(3)        Mr Sheth knew that the impression or signal given to other market participants in this situation was therefore false or misleading.  He intended to give such a false or misleading impression or signal.

814. The Tribunal concludes that this conduct amounted to market abuse within s118(1) FSMA 2000 and market manipulation within Article 15 of the Market Abuse Regulation.

815. On the basis of our conclusions as to Mr Sheth’s actual knowledge and intentions, and taking account not only of Mr Sheth’s junior status and that Mr Urra was the driving force behind this activity but also Mr Sheth’s training and the autonomy he was granted in his trading activity, the Tribunal finds that his conduct was dishonest by ordinary standards.  We draw no distinction for this purpose between the Large Orders and the Multiple Large Orders; they were all placed with the intention of facilitating the execution of the Small Orders.  It is this placing of the Large Orders that was dishonest for each Large Order; that his failure to cancel the earlier Large Order(s) was poor practice does not thereafter ameliorate the position.  

Mr Lopez

816. Mr Lopez’s explanation was that he placed large orders as anticipatory hedges pursuant to his Anticipatory Hedging Strategy.

817. It was not clear to the Tribunal how this had been explained by Mr Lopez to Compliance in September 2016.  We found that the notes of Mr Lopez’s interview with Compliance do not present a clear picture of the explanations that were being given to Compliance by the relevant Trader.  The notes of Mr Lopez’s interview do refer to his experience at his previous employer, but they do not include any indication that Compliance had identified that Mr Lopez was providing a different explanation for his trading than that which had been given by Mr Urra and Mr Sheth.  The differences did, however, emerge from the notes of Mr Lopez’s disciplinary interview with HR in October 2016, and this was then consistent with his interview with the Authority in 2018.

818. We have accepted that Mr Lopez was seeking to win more €20-30m RFQs, and that to do this he would need to offer competitive pricing.  A strategy of anticipatory hedging would in principle make sense, on the assumption that there is a predictable pattern of RFQs (with a predictable size, direction and time) and this would not hinder the Desk’s other activities.   

819. However, we have expressed doubts about the Anticipatory Hedging Strategy as it was pursued by Mr Lopez.  In particular:

(1)        It involved taking directional risk based on what Mr Lopez says were predictable patterns of RFQs in circumstances where those RFQs were a small part of the Desk’s market making business and where, although they were fairly regular, they were not predictable in timing.  The result would have been to take directional risk where Mr Lopez’s trading activity was otherwise very precise, with adjustments frequently being made by a single lot order.  If the Large Order had traded he would have held a directional position which ran the risk of going underwater and would have precluded him from quoting competitively in both directions.  As a strategy, we consider that it is at odds with both MHI’s business and with Mr Lopez’s own approach to managing the risk on his book. 

(2)        This strategy cannot explain the Large Orders that were placed late in the day.

(3)        Mr Lopez’s reconstructions of his activity sought to show a detailed and thoughtful approach to the identification of resistance and support levels, and consideration of other market information, which led to his assessment of an attractive price at which to place an anticipatory hedge, yet there were several Instances in which, having conducted this analysis, he placed a Large Order, cancelled that order and then came back minutes later at a less attractive price (and could have bought at his initial price if he had kept the first Large Order live).

(4)        Mr Lopez’s reconstruction of Instances also showed an inconsistent approach to the relevance of RFQs which had or had not been received by the Desk at the relevant time.

(5)        The Anticipatory Hedging Strategy does not explain the level of overlap between the Large Orders and the Small Orders, with only five of his 88 large orders in the Relevant Period being Lone Large Orders.

(6)        None of the large orders (whether the Large Orders or the Non-Instance large orders) traded.  Mr Lopez’s explanation was that the timing of the cancellation of his Large Orders could be explained by reference to market movements, which meant he concluded that the anticipated “pull back” to a particular level would not happen.  On Mr Lopez’s evidence, this happened every single time he placed a Large Order and yet he did not change his approach to try to get some of his order done as an anticipatory hedge, in circumstances where he did act to buy cash bonds to take an anticipatory position. 

(7)        There was a repeated coincidence of the cancellation of the Large Order shortly after the Small Order filled.  In the ten Specified Instances for Mr Lopez, the timing of cancellation after the Small Order filled varied from 0.58 seconds (F177) to 5.213 seconds (F190), and it was less than three seconds in eight of these Instances.  There is a significantly greater variation in the length of time for which the Large Order remained live - from 3.267 seconds (F56) to 16.269 seconds (F190).

820. The above reservations means that we have significant doubts about Mr Lopez’s evidence as to the rationale for the placing of his Large Orders and his intention to trade those Large Orders.  

821. We are also concerned by Mr Lopez’s failure to comply with the requirements of the Mandate.  Mr Lopez said he did not have a specific recollection of being shown the Mandate but accepted that he would have known about, and would have been familiar with, the limits and concepts referenced therein.  We find his apparent approach surprising - Mr Lopez’s evidence throughout the hearing showed his level of attention to detail, and he would have known (from his previous experience at Credit Suisse and from his training at MHI) that the Desk’s activities were constrained by the terms of the Mandate such that he needed to know those constraints.  Yet whilst he repeatedly referred to the Mandate permitting anticipatory hedging, he had not sought the approval of Mr Urra or Mr Heiberg for the Anticipatory Hedging Strategy.  We do not consider that it is an answer to this to say that anticipatory hedging is entirely normal practice, as he was not pre-positioning for specific imminent client activity or by reference to the average size of the RFQs received by the Desk; or that Mr Heiberg had recruited him to focus on medium-sized RFQs and the orders he placed were sized by reference to such RFQs, as that addresses the outcome rather than the means by which it was to be achieved.      

822. Nevertheless, we do recognise the potential underlying merits of Mr Lopez’s explanations for the Large Orders, including not only the basic premise of anticipatory hedging but the link between the size of his Large Orders and the size of the cash trades he was trying to win, and that he placed fewer large orders than Mr Urra and Mr Sheth.  In addition, Mr Jaffey drew attention to the small number of lots in some of the Small Orders which had been placed by Mr Lopez, and where he crossed the spread to trade. 

823. In assessing whether the Authority has met its burden of proof, the Tribunal has also considered whether Mr Lopez’s trading activity has been (unfairly) caught up in the pattern of trading activity which results from the abusive scheme which was devised by Mr Urra and pursued by Mr Sheth.  We look in particular at the Small Orders placed by Mr Lopez, his Lone Large Orders and his activity in the Multi Trader Instances.

824. We have considered the Small Orders in the Specified Instances when assessing the plausibility of the Authority’s case.  It is notable that, of the ten Specified Instances (which were selected by the Authority) for Mr Lopez, the size of the Small Order in each of F56 and F63 was just five lots, and in F132 it was one lot.  Separately, in F190, the Small Order filled when Mr Lopez amended the price and crossed the spread.  We agree with Mr Jaffey that, in circumstances where the Authority’s position is that the Traders were pursuing an abusive strategy to facilitate the execution of their Small Orders, it might have been expected that a Trader would deploy this strategy on occasions where the small orders were for a greater number of lots (and certainly not in single figures). 

825. For each Trader, we have recorded that the Lone Large Orders are unexplained on the Authority’s case.  For Mr Lopez, we consider this has less weight.  He placed 88 large orders in the Relevant Period, only five of which were Lone Large Orders, and two of those were placed very close to a small order in the opposite direction (the one to which we were taken filled in the same second as Mr Lopez placed his order).  This leaves three Lone Large Orders which are, on the Authority’s case, unexplained.    

826. We have considered Mr Lopez’s involvement in the Multi Trader Instances.  We consider that some of these Instances cast significant doubt on Mr Lopez’s explanations and his denial of collaboration with Mr Urra and Mr Sheth:

(1)        We have described the trading activity in F30 and F31 above.  Mr Lopez placed a Small Order in F30 and the only Large Order in that Instance was placed by Mr Urra (to buy 500 lots, 39 of which filled).  We have concluded that Mr Urra did not have an intention to trade the Large Order, and that it had been placed to facilitate the execution of Mr Lopez’s Small Order.  That of itself need not mean that Mr Lopez knew of or was otherwise involved in the abusive strategy being pursued by Mr Urra.  The difficulty then arises from Mr Lopez’s involvement in F31, where he placed a Large Order for 400 lots at 17.46.21.136.  We are not persuaded that such an order at that time of day could have been placed in anticipation of receiving RFQs of the requisite size from clients.  Notably, Mr Lopez and Mr Urra cancelled their Large Orders within 0.2 seconds of each other.

(2)        F94 effectively involves two parts as, whilst Mr Sheth placed all three Large Orders in this Instance, he changed direction part way through the Instance: 

(a)        Mr Lopez placed a Small Order in Part I, in the same direction as Mr Sheth’s Small Order.  Mr Lopez’s Small Order did not overlap with Mr Sheth’s Large Order.  Viewed in isolation, this does not assist us with assessing whether he has been unfairly caught up in the activity of the other Traders.

(b)        In Part II, Mr Lopez placed a Small Order and Mr Sheth placed two Large Orders (the first of which was cancelled before the second was placed).  Mr Sheth’s second Large Order was cancelled 1.041 seconds after Mr Lopez’s Small Order filled.  No explanation has been given for Mr Sheth’s two Large Orders.  (Mr Sheth had suggested he was testing the strategy in respect of an RFQ from Iccrea Banca for which the first Small Order in Part I had been a hedge, but that cannot explain the second and third Large Orders as they are in the opposite direction.)

(3)        Mr Lopez placed a Small Order and a Large Order in F174, and Mr Sheth placed three Large Orders of 500 lots each, all of which were live concurrently.  Mr Lopez’s Large Order was live for 7.16 seconds and was cancelled 1.072 seconds after the Small Order filled.  He cancelled less than 0.1 seconds after Mr Sheth cancelled his Large Order that was closest to the touch.  It seems incredible that this would have happened without any communication between Mr Lopez and Mr Sheth.

827. The Multi Trader Instances show repeated coincidences of timing between the activity of Mr Lopez and the other two Traders but also, as illustrated by F31 and F174, show that notwithstanding that Mr Lopez was, on his evidence, pursuing a different Trading Strategy to Mr Urra and Mr Sheth, and one which was in the direction of MHI’s predicted client flow, whereas Mr Urra and Mr Sheth were trading against client flow (or, at least, against the current flow of clients who had sent RFQs to the Desk), the end result was that Mr Lopez placed Large Orders at the same time and in the same direction as both Mr Urra and Mr Sheth.  This is remarkable and potentially incredible.

828. Mr Bailin and Mr Jaffey relied on Mr Sheth’s Multiple Large Orders in F174 as illustrative of their submission that the Traders were not collaborating as Mr Lopez would have seen these three concurrent Large Orders and would have corrected Mr Sheth’s error as Mr Sheth’s activity risked (and indeed did) drawing attention to their trading.  In F174 Mr Sheth had placed two Large Orders, each to buy 500 lots, before Mr Lopez placed his Large Order to buy 200 lots.  Mr Sheth’s two Large Orders were four and five ticks from the touch by the time Mr Lopez placed his Large Order (three ticks from the touch) and created a significant volume skew.  After Mr Lopez had placed his Large Order, Mr Sheth then placed a further Large Order, this time at the same price as Mr Lopez’s Large Order.  We agree that, if the Traders were collaborating to facilitate the execution of Mr Lopez’s Small Order in F174, we would have expected Mr Lopez to see this activity and challenge Mr Sheth if it had not been intended that Mr Sheth would place total orders of this volume on the order book.  We do note that it is an extreme example of a Multi Trader Instance where Mr Sheth had placed Multiple Large Orders.

829. We have referred throughout to the absence of any direct evidence of collaboration between the Traders to commit market abuse.  It is striking that, in circumstances where Mr Urra, Mr Lopez and Mr Sheth are alleged to have collaborated in an abusive strategy, they have provided two different explanations for their trading activity.  

830. Having assessed all of the evidence, and having considered Mr Jaffey’s detailed submissions in relation to the plausibility of the Authority’s case, the Tribunal is satisfied that the Authority has established that:

(1)        Mr Lopez placed the Large Orders with the intention of facilitating the execution of the Small Orders and he did not have an intention to trade the Large Orders. 

(2)        The Large Orders were likely to give the impression or signal to other market participants of significantly increased supply or demand.  Mr Lopez knew this and he knew that this would be likely to impact the trading activities of other market participants such that the most likely reaction would be for the market to move in the opposite direction to the Large Orders, ie towards the Small Orders on the opposite side of the order book. 

(3)        Mr Lopez knew that the impression or signal given to other market participants in this situation was therefore false or misleading.  He intended to give such a false or misleading impression or signal.

831. The Tribunal concludes that this conduct amounted to market abuse within s118(1) FSMA 2000 and market manipulation within Article 15 of the Market Abuse Regulation.

832. On the basis of our conclusions as to Mr Lopez’s actual knowledge and intentions, the Tribunal finds that his conduct was dishonest by ordinary standards.

Prohibition orders

833. The Authority may decide to make a prohibition order under s56 FSMA 2000 “if it appears to it that an individual is not a fit and proper person to perform functions in relation to a regulated activity”.  

834. The part of the Authority’s Handbook entitled “The Fit and Proper test for Approved Persons” contains relevant guidance, including FIT 1.3.1B G, which provides that “the most important considerations will be the person’s: (1) honesty, integrity and reputation; (2) competence and capability; and (3) financial soundness”.  The Authority’s Enforcement Guide provides further guidance.  Paragraph 9.2.3 provides that the scope of a prohibition order will depend on the range of functions which the individual performs in relation to regulated activities, the reasons why he is not a fit and proper person, and the severity of risk which he poses to consumers or to the market.

835. In its Decision Notice in relation to each Trader the Authority set out its decision on the same basis for all three of them:

(1)        Fitness and propriety - The Authority considers that the relevant Trader’s conduct in deliberately engaging in market manipulation was dishonest and lacked integrity.  This dishonest conduct was highly likely adversely to impact other market participants and was repeated many times over a period of two months.  As a result, they are not a fit and proper person to perform any function in relation to any regulated activity carried out by an authorised person, exempt person or exempt professional firm.

(2)        Prohibition - The Authority has had regard to the guidance in Chapter 9 of the Enforcement Guide in considering whether to impose a prohibition order on the relevant Trader.  The Authority has the power to prohibit individuals under section 56 FSMA 2000.  The Authority considers that, due to their dishonesty and the fact that they have engaged in deliberate market abuse, each Trader is not a fit and proper person to perform any function in relation to any regulated activity carried out by an authorised person, exempt person or exempt professional firm, and that a prohibition order should be imposed on them under section 56 FSMA 2000.

836. The Authority decided to make the prohibition orders on the basis that each Trader’s conduct was “dishonest and lacked integrity”.  There can be no doubt that conduct that is dishonest must also lack integrity for this purpose.    

837. We have concluded that Mr Urra, Mr Sheth and Mr Lopez each engaged in an abusive trading strategy and their conduct was dishonest.  It is clear that the imposition of a prohibition order by the Authority under s 56 FSMA 2000 on the grounds that they are each not a fit and proper person to perform any function in relation to any regulated activity carried out by an authorised person, exempt person or exempt professional firm is a course of action reasonably open to the Authority on the basis of the findings that we have made.  We therefore see no basis on which we should interfere with the Authority’s decision in that regard.

Penalties

838. The Authority has power under s123 FSMA 2000 to impose a penalty “of such amount as it considers appropriate” on a person if the Authority is satisfied that they have engaged in market abuse.

839. The Authority’s policy on imposing a financial penalty is set out in the Decision Procedure and Penalties Manual (“DEPP”).

840. DEPP 6.1.2 states that the principal purpose of imposing a financial penalty is to promote high standards of regulatory and/or market conduct by deterring persons who have committed breaches from committing further breaches and helping to deter other persons from committing similar breaches, as well as demonstrating generally the benefits of compliant behaviour.  Financial penalties (and public censures) are therefore tools that the Authority may employ to help it achieve its statutory objectives.

841. DEPP 6.2.1 identifies a non-exhaustive list of factors that the Authority will consider when determining whether or not to impose a financial penalty.  They include: the nature, seriousness and impact of the suspected breach; the conduct of the person after the breach; and their previous disciplinary record and compliance history.  Relevant factors in respect of any conduct after a breach include the degree of cooperation the person showed during the investigation, and the likelihood that the same type of breach will recur if no action is taken.

842. DEPP 6.2.2 provides that, when deciding whether to take action for market abuse, the Authority may consider additional factors including: the degree of sophistication of the users of the market in question, the size and liquidity of the market and the susceptibility of the market to market abuse; as well as the impact that any financial penalty may have on the financial markets or on the interests of consumers.

843. DEPP 6.4 contains guidance on when the Authority should consider issuing a public censure, instead of imposing a financial penalty. Relevant factors include whether the person has made a profit from the breach, whether the breach is serious, and whether the individual has a poor record of compliance where, in each case, those factors would favour the imposition of a financial penalty and their absence may be indicative that the issue of a public censure is more appropriate. The application of the Authority’s policy on serious financial hardship may also result in a financial penalty being reduced to zero (DEPP 6.4.2(8)).

844. DEPP 6.5.2 states that the Authority’s penalty-setting regime is based on the following principles:

(1)        Disgorgement - a firm or individual should not benefit from any breach;

(2)        Discipline - a firm or individual should be penalised for wrongdoing; and

(3)        Deterrence - any penalty imposed should deter the firm or individual who committed the breach, and others, from committing further or similar breaches.

845. The Authority applies a five-step framework to determine the appropriate level of financial penalty.   The steps apply in all cases, but Steps 1 to 4 differ in market abuse cases against individuals.  In summary:

Step 1: Disgorgement

846. This involves the disgorgement of any financial benefit derived as a direct result of the market abuse, where it is possible to quantify this (DEPP 6.5C.1).

Step 2: The seriousness of the breach

847. DEPP 6.5C.2 provides that the Authority will determine a figure dependent on the seriousness of the market abuse, and whether it was referable to the individual’s employment, and sets out a formula for the determination of the penalty.

848. For penalties imposed on individuals for market abuse that are referable to their employment, the figure is the greater of:

(1)        a figure based on a percentage of the individual’s “relevant income”;

(2)        a profit multiple; and

(3)        for market abuse cases which the Authority assesses to be seriousness level 4 or 5, £100,000.  The Authority usually expects to assess market abuse committed deliberately as seriousness level 4 or 5.

849. The Authority will determine the percentage of relevant income, or profit multiple, which will apply by considering the seriousness of the market abuse and choosing a percentage between 0% and 40% and a multiple between 0 and 4.

850. There are five levels:

(1)        level 1 - 0%, profit multiple of 0;

(2)        level 2 - 10%, profit multiple of 1;

(3)        level 3 - 20%, profit multiple of 2;

(4)        level 4 - 30%, profit multiple of 3; and

(5)        level 5 - 40%, profit multiple of 4.

851. In deciding which level of penalty to apply, the Authority will consider a range of factors, including the impact of the market abuse, its nature, factors tending to show it was deliberate; and factors tending to show it was reckless.

852. Factors relating to the impact of the market abuse include the level of benefit gained or loss avoided, and whether the market abuse has had an adverse effect on markets.

853. Factors relating to the nature of the market abuse include: its frequency; whether the individual abused a position of trust; whether the individual caused others to commit market abuse; whether the individual has a prominent position in the market; and whether the individual is an experienced industry professional or held a senior position in the firm.

854. Factors tending to show market abuse was deliberate are identified as including: that it was intentional, in that “the individual intended or foresaw that the likely or actual consequences of his actions would result in market abuse”; the individual intended to benefit financially from market abuse; the individual knew that his actions were not in accordance with rules or procedures; the individual sought to conceal his misconduct or committed it in a way to avoid or reduce the risk of discovery; the individual thought his conduct would be difficult to detect; the individual’s actions were repeated.

855. Factors tending to show market abuse was reckless are identified as including: the individual appreciated there was a risk that his actions could result in market abuse and failed adequately to mitigate that risk; and the individual was aware there was a risk that his actions could result in market abuse but failed to check if he was acting in accordance with internal procedures.

856. Factors which are likely to be considered level 4 or 5 factors include that: the losses avoided or benefits gained were significant; the conduct resulted in a serious adverse effect on the orderliness of, or confidence in, the market; the market abuse was committed on multiple times; the individual was in a position of trust or in a prominent position in the market; the market abuse was committed deliberately or recklessly.

857. Factors which are likely to be considered level 1, 2 or 3 factors include that: little or no profits were made or losses avoided as a result of the market abuse; there was no or limited actual or potential effect on the markets; and the market abuse was committed negligently or inadvertently. 

Step 3: Mitigating and aggravating factors 

858. The Authority may increase or decrease the amount of the financial penalty arrived at after Step 2, but not including any amount to be disgorged as set out in Step 1, to take into account factors which aggravate or mitigate the market abuse.  Any such adjustments will be made by way of a percentage adjustment to the figure determined at Step 2.

859. The following list of factors may have the effect of aggravating or mitigating the market abuse:

(1)        the conduct of the individual in bringing (or failing to bring) quickly, effectively and completely the market abuse to the Authority’s attention (or the attention of other regulatory authorities, where relevant);

(2)        the degree of cooperation the individual showed during the investigation of the market abuse by the Authority, or any other regulatory authority allowed to share information with the Authority;

(3)        whether the individual assists the Authority in action taken against other individuals for market abuse and/or in criminal proceedings;

(4)        whether the individual has arranged his resources in such a way as to allow or avoid disgorgement and/or payment of a financial penalty;

(5)        whether the individual had previously been told about the Authority’s concerns in relation to the issue, either by means of a private warning or in supervisory correspondence;

(6)        the previous disciplinary record and general compliance history of the individual;

(7)        action taken against the individual by other domestic or international regulatory authorities that is relevant to the market abuse in question;

(8)        whether Authority guidance or other published materials had already raised relevant concerns, and the nature and accessibility of such materials; and

(9)        whether the individual agreed to undertake training subsequent to the market abuse.

Step 4: Adjustment for deterrence

860. If the Authority considers the figure arrived at after Step 3 is insufficient to deter the individual who committed the market abuse, or others, from committing further or similar abuse then the Authority  may increase the penalty.  Circumstances where the Authority may do this include:

(1)        where the Authority considers the absolute value of the penalty too small in relation to the market abuse to meet its objective of credible deterrence;

(2)        where previous Authority action in respect of similar market abuse has failed to improve industry standards; and

(3)        where the penalty may not act as a deterrent in light of the size of the individual’s income or net assets.

Step 5: Settlement discount 

861. The Authority and the individual on whom a penalty is to be imposed may seek to agree the amount of any financial penalty and other terms.  In recognition of the benefits of such agreements, DEPP 6.7 provides that the amount of the financial penalty which might otherwise have been payable will be reduced to reflect the stage at which the Authority and the individual concerned reached an agreement.  The settlement discount does not apply to the disgorgement of any benefit calculated at Step 1.

Authority’s determination of the penalties to be imposed

862. In outline, for each Trader, the Authority determined the amount of the penalty on the basis that it had not identified any financial benefit to each Trader, the seriousness was level 4, there were no aggravating or mitigating factors, it did not increase the amount for deterrence and no settlement agreement was reached.  Different amounts were assessed as a result of the Authority’s determination of the Step 2 figure, which meant that for Mr Lopez and Mr Sheth this was determined at £100,000 at Step 2, whereas for Mr Urra this was determined to be 30% of £1,318,367, ie £395,510, which was rounded down to £395,500 in accordance with the Authority’s usual practice.

863. We address the position of each Trader, and their submissions on the amount of the penalty below.

Assessment of the financial penalty

864. The Tribunal has the power to determine at its discretion what (if any) is the appropriate action for the Authority to take, including a determination as to whether or not to impose a financial penalty and, if so, the amount of such penalty.  The Tribunal is not bound by the Authority's policy when making an assessment of a financial penalty on a reference, but it pays the policy due regard when carrying out its overriding objective of doing justice between the parties.  In so doing the Tribunal looks at all the circumstances of the case.

865. We have followed the Authority’s five-step framework in determining the amount of the penalties.  We have, however, concluded that we should, in the interests of fairness and justice, depart from the minimum amount of the penalty of £100,000 at Step 2 in respect of Mr Sheth, as explained below.

Mr Urra

866. The Authority applied its five-step framework as follows:

(1)        The Authority has not identified any financial benefit that Mr Urra derived directly from the market abuse.  Step 1 is therefore £0.

(2)        The figure at Step 2 is dependent on whether or not the market abuse was referable to the individual’s employment, and will be the greater of a figure based on “relevant income”, a profit multiple, and, for market abuse cases which the Authority assesses to be seriousness level 4 or 5, £100,000.   

(a)        The relevant period for calculating Mr Urra’s relevant income is the 12-month period ending on 29 July 2016, and his relevant income for that period was £1,318,367.

(b)        Mr Urra intended to mislead other market participants, he knew that the misleading orders would give false or misleading signals to other market participants as to the supply of or demand for Futures, and he knew that placing misleading orders constituted market abuse as a result of his considerable experience in the market and the training that he had undertaken.  This was repeated on multiple occasions in the Relevant Period, both alone and in concert with Mr Lopez and Mr Sheth.  The Authority considered that he deliberately committed market abuse.

(c)        The Authority considered the following level 4 or 5 factors to be relevant: the market abuse was committed on multiple occasions during the Relevant Period; Mr Urra breached a position of trust by conspiring with his direct reports in an abusive strategy when he should have been setting an example for them through his conduct, and the market abuse was committed deliberately or recklessly.

(d)        The Authority considered the following level 1, 2 or 3 factors to be relevant: Limited profits were made or losses avoided, either directly or indirectly.  However, Mr Urra aimed to execute his genuine orders more efficiently and manage better the risk on his book as a result of the placement of his misleading orders.  This would have improved the performance of the Desk which was a factor taken into account in determining the bonus he was to receive.

(e)        The Authority also took into account that: Mr Urra is an experienced industry professional, held a senior position within MHI, held the Certified Roles of working in a Client Dealing Function, being a Material Risk Taker and being the Manager of Certified Persons (namely Mr Lopez and Mr Sheth) and before that had held a number of roles as an Approved Person; and the Authority usually expects to assess deliberate market abuse as seriousness level 4 or 5.

(f)         The Authority considered the seriousness to be level 4, which means the Step 2 figure is the higher of (i) 30% of Mr Urra’s relevant income of £1,318,367, ie £395,510, (ii) a profit multiple of 3 applied to his financial benefit of £0, ie £0, and (iii) £100,000.  Step 2 is therefore £395,510.

(3)        The Authority concluded there are no aggravating or mitigating factors such as to justify an adjustment to the Step 2 figure.  Step 3 is therefore £395,510.

(4)        The Authority considered this represents a sufficient deterrent to Mr Urra and others, and so has not increased the penalty at Step 4.

(5)        No settlement agreement has been reached, and Step 5 is therefore £395,510.

867. The Authority therefore decided to impose a total financial penalty of £395,500 (ie the amount at Step 5 rounded down to the nearest £100) on Mr Urra.

868. Mr George submitted that if the Tribunal found that Mr Urra did commit the market abuse alleged, the penalty of £395,500 which the Authority had decided to impose should be substantially reduced.  This was on the basis that:

(1)        Mr Urra had endured significant hardship as a result of the matters giving rise to the reference, having lost a successful and lucrative career in London, been forced to relocate his family to Spain and has found it difficult to secure alternative employment.  The penalty is grossly disproportionate.

(2)        The Authority’s calculation of “relevant income” had been based on the amounts received by Mr Urra in the 12 months, whether or not they were actually earned in that period.  As a result, the Authority had included £573,567 awarded in respect of cash bonuses, awards and deferred shares received in the period 30 July 2015 to 29 July 2016 but earned two years prior.

(3)        The seriousness level should be no higher than level 3, as little or no profits were made or losses avoided; and there was no, or limited, actual or potential effect on the orderliness of, or confidence in, markets as a result of the market abuse.

(4)        The Authority found no mitigating factors, but Mr Urra relies on his cooperation during the investigation and he has an exemplary disciplinary record.

869. The Tribunal adopts the Authority’s five-step penalty framework.

Step 1

870. We have not identified any financial benefit that Mr Urra derived directly from the market abuse, so Step 1 is £0.

Step 2

871. We consider the amount of Mr Urra’s “relevant income” for the period from 30 July 2015 to 29 July 2016. 

872. Mr George submitted that it is clear that income “earned by the individual in the 12 months preceding the final market abuse” under DEPP 6.5C.2.G(5) is limited to remuneration paid in respect of work undertaken in that 12 month period, and does not encompass prior earned sums which happen to vest in that period.  In any event, the Tribunal should depart from the Authority’s penalty policy and (to the extent that the Authority’s position reflects the true construction of that policy) the Tribunal should depart from it as the Authority’s position is arbitrary and/or does not reflect the overriding purpose of the policy.  The remuneration related to work undertaken before this 12 month period should be removed from the calculation of “relevant income”, reducing it from £1,318,367 to £744,800.

873. The Authority submitted that this approach is wrong in principle.  If correct, it would result in the artificial suppression of a person’s income for the purpose of calculating a penalty and fails to take into account money “earned” during the relevant period which fell due for payment afterwards.  It is also inconsistent with the purpose of the financial penalty regime to punish and deter market abuse.

874. The purpose of imposing a financial penalty is to promote high standards of market conduct by deterring persons who have committed breaches from committing further breaches and helping to deter other persons from committing similar breaches, as well as demonstrating generally the benefits of compliant behaviour.  The focus is thus deterrence rather than, as submitted by the Authority, punishment.

875. DEPP deals with the calculation of “relevant income” as follows:

(1)        DEPP 6.5C.1(1) sets out the Authority’s view that where an individual has been put into a position where he can commit market abuse because of his employment, the fine imposed should reflect this “by reference to the gross amount of all benefits derived from that employment”.

(2)        DEPP 6.5C.2G(4) states that an individual’s “relevant income” will be “the gross amount of all benefits received by the individual from the employment in connection with which the market abuse occurred”, and “benefits” includes, but is not limited to, salary, bonus, pension contributions, share options and share schemes.

(3)        DEPP 6.5C.2G(5) states that where the market abuse lasted less than 12 months, or was a one-off event, the relevant income “will be that earned by the individual in the 12 months preceding the final market abuse”.

876. DEPP thus defines “benefits” widely, and the inclusive definition clearly encompasses the cash bonuses, awards and deferred shares which vested in Mr Urra. 

877. We recognise that DEPP uses the concepts of “derived from”, “received by” and “earned by” at different times; and that these are not the same as each other.  However, we are wary of seeking to construe this guidance as though it were a statutory provision.  DEPP is seeking to capture the benefits from the employment, and provide a means of calculating “relevant income” where the market abuse lasted less than 12 months.

878. We agree with Mr George that the reference in DEPP 6.5C.2G(5) to income “earned by” Mr Urra in the period from 30 July 2015 to 29 July 2016 is distinct from the concept of whether amounts were “received by” Mr Urra in that 12 month period.  We are inclined to prefer an approach which does not treat as “relevant income” amounts which had been earned in prior years but which were paid out or vested during the relevant period.  We have considered whether such an approach does, as submitted by the Authority, artificially suppress Mr Urra’s income for this purpose.  We are not persuaded that it does.  In particular, the Authority’s submissions appear to be concerned to avoid a situation where bonuses received by an individual that are referable to work in earlier years and bonuses earned in the 12 month period but payable in future years are both taken out of account when determining “relevant income” for that period.  On the facts before us, no such concern arises.  Mr Urra has not been awarded any bonus by MHI for the 12 month period in question, and there is no deferred payment or vesting of such an amount which risks being taken out of account.  On this basis, we consider that there is no reason to depart from our preferred approach of identifying the amount earned for the work undertaken in the 12 month period, and not taking into account cash bonuses, awards and deferred shares received in that period but earned two years prior. 

879. The Tribunal determines that Mr Urra’s “relevant income” is £744,800.

880. We agree with the Authority’s identification of relevant level 4 or 5 factors that are met.    

881. The Authority also identified the relevant level 1, 2 or 3 factors, with which we agree.  Mr George submitted that an additional relevant factor was that there was no, or limited, actual or potential effect on the orderliness of, or confidence in, markets as a result of the market abuse.  We accept that the Authority has not established any such actual or potential effect.

882. The Authority also took into account that Mr Urra is an experienced industry professional, held a senior position within MHI and the Authority usually expects to assess deliberate market abuse as seriousness level 4 or 5.

883. Mr George submitted that the presence of these level 1, 2 or 3 factors indicates that the seriousness level is no higher than level 3.  The Tribunal does not agree that the presence of level 1, 2 or 3 factors means that the seriousness level should be no higher than level 3.  This would be to disregard the presence of any level 4 or 5 factors.

884. Taking all of these identified factors into account, the Tribunal has no doubt that the significance of the level 4 or 5 factors, alongside the additional factors taken into account by the Authority, outweigh the level 1, 2 or 3 factors.  The Tribunal considers the seriousness of the market abuse to be level 4.  This means the Step 2 figure is 30% of Mr Urra’s relevant income of £744,800, ie £223,440.

Step 3

885. The Authority did not identify any aggravating or mitigating factors.

886. Mr George submitted that Mr Urra relied upon his cooperation during the investigation of the market abuse; and he has an exemplary disciplinary record.

887. The approach in DEPP is that there is a discretion as to whether to increase or decrease the amount of the financial penalty arrived at after Step 2 to take into account factors which aggravate or mitigate the market abuse.  We have found that Mr Urra has not been the subject of any regulatory investigation or disciplinary sanction until the matters which are the subject of this reference.  We accept that Mr Urra attended an interview with the Authority, in which he exercised his right to give a “no comment” interview.  We are not persuaded that either of these matters merit any reduction in the amount of penalty.

888. Step 3 is therefore £223,440. 

Step 4

889. We consider that the figure of £223,440 is sufficient to deter Mr Urra, and others, from committing further or similar abuse.  We do not increase the penalty at Step 4.

Step 5

890. No settlement agreement has been reached, and Step 5 is therefore £223,440.

891. The Tribunal determines that it will be appropriate for the Authority to impose a financial penalty of £223,400 (rounding down to the nearest £100) on Mr Urra.

Mr Lopez

892. The Authority applied its five-step framework as follows:

(1)        The Authority has not identified any financial benefit that Mr Lopez derived directly from the market abuse.  Step 1 is therefore £0.

(2)        The figure at Step 2 is dependent on whether or not the market abuse was referable to the individual’s employment, and will be the greater of a figure based on “relevant income”, a profit multiple, and, for market abuse cases which the Authority assesses to be seriousness level 4 or 5, £100,000.   

(a)        The relevant period for calculating Mr Lopez’s relevant income is the 12-month period ending on 29 July 2016, and his relevant income for that period was £143,089, calculated on a pro rata basis.

(b)        Mr Lopez intended to mislead other market participants, he knew that the misleading orders would give false or misleading signals to other market participants as to the supply of or demand for Futures, and he knew that placing misleading orders constituted market abuse as a result of his experience in the market and the training that he had undertaken.  This was repeated on multiple occasions in the Relevant Period.  The Authority considered that he deliberately committed market abuse.

(c)        The Authority considered the following level 4 or 5 factors to be relevant: the market abuse was committed on multiple occasions during the Relevant Period; and the market abuse was committed deliberately or recklessly.

(d)        The Authority considered the following level 1, 2 or 3 factors to be relevant: Limited profits were made or losses avoided, either directly or indirectly.  However, Mr Lopez aimed to execute his genuine orders more efficiently and manage better the risk on his book as a result of the placement of his misleading orders.  This would have improved the performance of the Desk which was a factor taken into account in determining the bonus he was to receive.

(e)        The Authority also took into account that: Mr Lopez is an experienced industry professional, holding the Certified Role of working in a Client Dealing Function, and before that had been an Approved Person, first holding a Controlled Function in 2006; and the Authority usually expects to assess deliberate market abuse as seriousness level 4 or 5.

(f)         The Authority considered the seriousness to be level 4, which means the Step 2 figure is the higher of (i) 30% of Mr Lopez’s relevant income of £143,089, ie £42,926, (ii) a profit multiple of 3 applied to his financial benefit of £0, ie £0, and (iii) £100,000.  Step 2 is therefore £100,000

(3)        The Authority concluded there are no aggravating or mitigating factors such as to justify an adjustment to the Step 2 figure.  Step 3 is therefore £100,000.

(4)        The Authority considered this represents a sufficient deterrent to Mr Lopez and others, and so has not increased the penalty at Step 4.

(5)        No settlement agreement has been reached, and Step 5 is therefore £100,000.

893. Mr Jaffey invited the Tribunal to find that the allegations of market abuse were not substantiated such that the reference should be remitted to the Authority with directions that no penalty be imposed.  However, Mr Jaffey did not make any submissions as to the level of the penalty that should be imposed if the Tribunal was to find that the allegations were made out.

894. We have considered the Authority’s approach to the determination of the penalty to be imposed on Mr Lopez, including in the light of our conclusions as to the approach to the determination of the penalty to be imposed on Mr Sheth.  We adopt it ourselves.    

895. The Tribunal determines that it will be appropriate for the Authority to impose a financial penalty of £100,000 on Mr Lopez.

Mr Sheth

896. The Authority applied its five-step framework as follows:

(1)        The Authority has not identified any financial benefit that Mr Sheth derived directly from the market abuse.  Step 1 is therefore £0.

(2)        The figure at Step 2 is dependent on whether or not the market abuse was referable to the individual’s employment, and will be the greater of a figure based on “relevant income”, a profit multiple, and, for market abuse cases which the Authority assesses to be seriousness level 4 or 5, £100,000.  The Authority usually expects to assess market abuse committed deliberately as seriousness level 4 or 5. 

(a)        The relevant period for calculating Mr Sheth’s relevant income is the 12-month period ending on 29 July 2016, and his relevant income for that period was £82,478.

(b)        Mr Sheth intended to mislead other market participants, he knew that the misleading orders would give false or misleading signals to other market participants as to the supply of or demand for Futures, and he knew that placing misleading orders constituted market abuse.  This was repeated on multiple occasions in the Relevant Period.  The Authority considered that he deliberately committed market abuse.

(c)        The Authority considered the following level 4 or 5 factors to be relevant: the market abuse was committed on multiple occasions during the Relevant Period; and the market abuse was committed deliberately or recklessly.

(d)        The Authority considered the following level 1, 2 or 3 factors to be relevant: Limited profits were made or losses avoided, either directly or indirectly.  However, Mr Sheth aimed to execute his genuine orders more efficiently and manage better the risk on his book as a result of the placement of his misleading orders.  This would have improved the performance of the Desk which was a factor taken into account in determining the bonus he was to receive.

(e)        The Authority also took into account that: Mr Sheth is an experienced industry professional, holding the Certified Role of working in a Client Dealing Function, and before that had been an Approved Person; and the Authority usually expects to assess deliberate market abuse as seriousness level 4 or 5.

(f)         The Authority considered the seriousness to be level 4, which means the Step 2 figure is the higher of (i) 30% of Mr Sheth’s relevant income of £82,478, ie £24,743, (ii) a profit multiple of 3 applied to his financial benefit of £0, ie £0, and (iii) £100,000.  Step 2 is therefore £100,000

(3)        The Authority concluded there are no aggravating or mitigating factors such as to justify an adjustment to the Step 2 figure.  Step 3 is therefore £100,000.

(4)        The Authority considered this represents a sufficient deterrent to Mr Sheth and others, and so has not increased the penalty at Step 4.

(5)        No settlement agreement has been reached, and Step 5 is therefore £100,000.

897. Mr Bailin made submissions on the amount of the penalty.  He did not make any challenge to Steps 1, 3, 4 and 5 in the Authority’s calculation of the penalty.  The application of Step 2, and the imposition of the minimum penalty of £100,000 had, he submitted rendered the proposed penalty disproportionate and unjust and unfair.  The purpose of the financial penalty is to deter Mr Sheth and the wider public from committing market abuse; the primary purpose is deterrence, not penal.  Mr Bailin submitted, as explained further below, that the Tribunal should either depart from the minimum penalty or assess the conduct as seriousness level 3.

898. The Tribunal adopts the Authority’s five-step penalty framework.

Step 1

899. We have not identified any financial benefit that Mr Sheth derived directly from the market abuse, so Step 1 is £0.

Step 2

900. The figure at Step 2 will be the greater of a figure based on “relevant income”, a profit multiple, and, for market abuse cases which the Authority assesses to be seriousness level 4 or 5, £100,000. 

901. The relevant period for calculating Mr Sheth’s relevant income is the 12-month period from 30 July 2015 to 29 July 2016, and his relevant income for that period was £82,478.

902. Mr Bailin referred to DEPP 6.5C.2.2, which provides that the minimum penalty “for market abuse cases which the [Authority] assesses to be seriousness level 4 or 5, £100,000.  The [Authority] usually expects to assess market abuse cases committed deliberately as seriousness level 4 or 5”.  Mr Bailin submitted:

(1)        DEPP also acknowledges that level 4 will not be appropriate in every case of deliberate market abuse, and those factors must be weighed against other factors.  DEPP 6.5C contains a list of level 1, 2 and 3 factors, the presence of which in Mr Sheth’s case tend to indicate that his case should have been assessed as level 3 seriousness, namely:

(a)        little, or no, profits were made or losses avoided as a result of the market abuse, either directly or indirectly; and

(b)        there was no, or limited, actual or potential effect on the orderliness of, or confidence in, markets as a result of the market abuse. 

(2)        The consequences for Mr Sheth in the Step 2 figure being assessed at level 4 has meant that the minimum penalty of £100,000 has been engaged; yet had his penalty been calculated based on level 4 at 30% seriousness, the penalty would have been £24,843 (being 30% of £82,478).  The penalty of £100,000 is disproportionate compared to the penalties imposed on Mr Urra and Mr Lopez by reference to their relevant income.  Mr Sheth was the junior trader on the Desk and this is not just and fair.

(3)        There are good and appropriate reasons to follow DEPP 6.5C in principle but to depart from it so as to ensure the penalty is fair and just, namely through the application of a percentage of income based approach.  To not do so produces disproportionate consequences for Mr Sheth.

(4)        Alternatively, the Tribunal could find that the disproportionate consequences stem from the assessment of Mr Sheth’s conduct at level 4, the same level as that assessed for the other Traders.  This takes no account of his more junior status, that he did not cause or encourage others to commit market abuse, he did not occupy a prominent position in the market, he was not an experienced industry professional and he did not hold a senior position at MHI.  If the conduct had been found to be level 3 seriousness, the penalty would have been calculated based on a percentage of Mr Sheth’s relevant income. 

903. We have found that Mr Sheth’s conduct was deliberate and dishonest.  Whilst we accept that the identified level 1, 2 and 3 factors are present, the Tribunal does not consider that the seriousness could be less than level 4 where conduct is deliberate and dishonest, notwithstanding that Mr Sheth was the more junior trader. 

904. If the Tribunal were to apply the approach in DEPP without any adjustments, the Step 2 figure would then be, as calculated by the Authority, the higher of (i) 30% of Mr Sheth’s relevant income of £82,478, ie £24,743, (ii) a profit multiple of 3 applied to his financial benefit of £0, ie £0, and (iii) £100,000. 

905. The Tribunal recognises that the minimum penalty of £100,000 assists with the promotion of the Authority’s policy to promote high standards of regulatory and/or market conduct by deterring persons who have committed breaches from committing further breaches, helping to deter other persons from committing similar breaches and demonstrating generally the benefits of compliant behaviour.  The Tribunal accepts and agrees that a penalty of £24,743 may not be seen as being of sufficient magnitude for this purpose.

906. Addressing Mr Bailin’s submissions by reference to the penalties assessed on Mr Urra and Mr Lopez, their relevant income is £744,800 and £143,089 respectively, and we have determined that it will be appropriate for the Authority to impose financial penalties of £223,440 and £100,000 respectively.  We do accept that, by reference to relevant income, imposing a penalty of £100,000 on Mr Sheth would be disproportionate to those imposed on Mr Urra and Mr Lopez, and that this is not fair or just.  We do not, for this purpose, put any weight on Mr Sheth’s junior status, his relative experience, or their respective roles at MHI; we have found that all three Traders committed market abuse and that their conduct was dishonest.  We have decided that it would be appropriate to calculate the Step 2 figure by reference to the proportion which Mr Sheth’s relevant income bears to that of Mr Lopez and to apply that percentage to the amount determined for Mr Lopez at Step 2.  This results in a Step 2 figure for Mr Sheth of £57,641 (£82,478/£143,089 x £100,000).  This remains imperfect, including by reference to the relevant income of Mr Urra; but we consider that it pays due regard to the Authority’s approach that the figure at Step 2 should not necessarily be limited to 30% of relevant income where the market abuse has been assessed as level 4 seriousness.

907. Step 2 is therefore £57,641.

Step 3

908. The Tribunal considers that there are no aggravating or mitigating factors such as to justify an adjustment to the Step 2 figure.  Step 3 is therefore £57,641.

Step 4

909. We consider that the figure of £57,641 is sufficient to deter Mr Sheth, and others, from committing further or similar abuse.  We do not increase the penalty at Step 4.

Step 5

910. No settlement agreement has been reached, and Step 5 is therefore £57,641.

911. The Tribunal determines that it will be appropriate for the Authority to impose a financial penalty of £57,600 (rounding down to the nearest £100) on Mr Sheth.

Disposition

912. The non-disciplinary references are dismissed.  Our decision is unanimous.

Directions

913. In relation to Mr Urra’s disciplinary reference, we determine that the appropriate action for the Authority to take is to impose on him a financial penalty of £223,400.  That penalty is to be imposed pursuant to s123(1) FSMA 2000 for engaging in market abuse as defined by s118(5) and in contravention of s118(1) between 1 June and 2 July 2016 and engaging in market manipulation as defined by Article 12(1)(a) of the Market Abuse Regulation and in contravention of Article 15 between 3 July and 29 July 2016. 

914. In relation to Mr Lopez’s disciplinary reference, we determine that the appropriate action for the Authority to take is to impose on him a financial penalty of £100,000.  That penalty is to be imposed pursuant to s123(1) FSMA 2000 for engaging in market abuse as defined by s118(5) and in contravention of s118(1) between 1 June and 2 July 2016 and engaging in market manipulation as defined by Article 12(1)(a) of the Market Abuse Regulation and in contravention of Article 15 between 3 July and 29 July 2016.

915. In relation to Mr Sheth’s disciplinary reference, we determine that the appropriate action for the Authority to take is to impose on him a financial penalty of £57,600.  That penalty is to be imposed pursuant to s123(1) FSMA 2000 for engaging in market abuse as defined by s118(5) and in contravention of s118(1) between 1 June and 2 July 2016 and engaging in market manipulation as defined by Article 12(1)(a) of the Market Abuse Regulation and in contravention of Article 15 between 3 July and 29 July 2016.

916. In accordance with s133(6) FSMA 2000 we have dismissed the non-disciplinary references.  It is therefore open to the Authority to make a prohibition order against each of Mr Urra, Mr Lopez and Mr Sheth prohibiting them from performing any function in relation to any regulated activity carried on by an authorised person, exempt person or exempt professional firm.

917. We remit the references to the Authority with the direction that effect be given to our determinations.

 

 

 

JEANETTE ZAMAN

 

UPPER TRIBUNAL JUDGE

 

Release date: 01 July 2025

 

 


 

 

Appendix 1

 

Statement of agreed background facts

1.       This Statement of Agreed Facts has been agreed by all parties pursuant to the Tribunal’s request at the conclusion of the fourth day of trial. It covers four topics plus a glossary, which is also agreed. Those four topics are: (i) the cash BTP market; (ii) Requests for a quote or “RFQs” and orders; (iii) hedging and trading BTP futures on EUREX; and (iv) MHI and the EGB Desk.

The Cash BTP Market

2.       “BTP” stands for “Buoni del Tesoro Poliennali” (literally multi-year treasury bonds) which are long term bonds issued by the Italian Government. Alongside bonds issued by Spain, Portugal and Greece, they are classified as “Peripheral” European Government Bonds or “EGBs” as distinct from “Core” EGBs, including German, French, Dutch, Belgian and Finnish government debt.

3.       A bond is a debt instrument typically issued by governments and companies by which the borrower agrees to repay the sum which it has borrowed to the holder of the bond at a specified future date called the “maturity date”. The face value of the bond, referred to as its “nominal” or “par” value, is the amount to be repaid at the maturity date. Each bond will also have a fixed interest rate attached to it, expressed as a percentage of its face value. This is called the “coupon”.

4.       Bonds, including BTPs, are issued in two ways: (i) via auctions to financial institutions permitted to participate in them; and (ii) syndicated issues to consortiums of investors. In each case, only certain entities are entitled to acquire the bonds from the issuer. These entities are known as “Primary Dealers”. Primary Dealers, amongst other things, underwrite syndicated primary issues of bonds and also sell bonds which they have acquired in “primary” auctions onto the secondary market.

5.       The bonds may then trade on the secondary market. Bonds trading on the secondary market are referred to as “cash bonds” (and cash BTPs in the case of BTPs) and trades in cash bonds as “cash trades”. The price of a particular bond may, over the course of its term, fluctuate according to a variety of factors, including supply and demand, bond credit rating, interest rates, performance of the Italian economy (in the case of BTPs), market sentiment and the proximity of the maturity date. This may result in it being sold at a premium or discount to its par value.

6.       The “yield” on a bond is calculated by reference to its price and its coupon. The coupon is fixed so if the price of the bond increases, the yield (ie the return on the purchase price) declines and vice versa.

7.       There were relatively few Primary Dealers in BTPs during the Relevant Period (20 or fewer). Due to the significant obligations incurred by Primary Dealers, including in relation to the provision of liquidity to secondary markets and supporting scheduled “secondary” auctions (ie the sale of bonds onto secondary markets by the Primary Dealer), and the costs associated with being a Primary Dealer, Primary Dealers tend to be large banks and financial institutions. However, smaller banks such as MHI traded on the secondary market and acted as “market makers”. Accordingly, whilst Primary Dealers engage in market making activities, not all market makers are Primary Dealers.

8.       The process of clients seeking quotes and placing orders is dealt with further below. A market maker typically trades from the fixed income desk of an investment bank and seeks to service their clients and provide liquidity to the market by the provision of quotes to clients and subsequently transacting if the client accepts the quote. They will win business by quoting the most competitive prices.

9.       Providing competitive prices involves traders continuously assessing what prices they are prepared to offer by reference to their own book and market conditions.

10.    Market making in EGBs is very competitive.

11.    US legislation known as the “Volcker Rule” prohibits banks from engaging in proprietary trading (ie, short-term trading for their own profit) but allows an exception for “market-making-related activities”.

RFQs and cash trades

12.    Market makers provide prices to clients to buy or sell cash EGBs in response to requests for quotes. Such a request for a quote is referred to as an “RFQ”. An RFQ may be made by a client in three ways. Generally, the RFQ will come through via an electronic platform such as Bond Vision or Tradeweb or Bloomberg (an “Electronic RFQ”). However, less commonly, it may be made by telephone to the bank’s sales desk, known as a “voice” order or sent via Bloomberg messenger. The sales desk will then pass it to the market maker to provide the quote. RFQs may be placed directly by the underlying client or by a broker on their behalf. RFQs are directed by clients at specified market makers chosen by the client, not the entire market.

13.    Electronic RFQs from the Relevant Period have been retained in searchable form by MHI and are available in the trial bundle.

14.    RFQs must be responded to quickly given the pace at which trading takes place on the EGB market. The process of providing a quote involves the market maker considering the appropriate price for the cash bond in question, which could have a maturity date of anything from a few months up to 30 years. That will involve consideration of their book position, the cost of hedging (including via the BTP futures market), the risk of the trade and any commercial rationale for offering a competitive price to a particular client. A client may ask for quotes from several market makers and then choose the most competitive.

15.    Once a market maker provides a quote for a client via the RFQ process, the client generally has a brief window of a few seconds to decide whether to accept the quote. If they do, the cash trade is executed and entered on the market maker’s systems. The customer may reject the quote (CustRejected), execute the transaction elsewhere (Traded Away), or fail to respond (CustomerTimeOut). 

Hedging and trading BTP futures on EUREX

16.    Changes in market interest rates typically affect the price of the bond. In essence, when the market interest rate rises, the price of a bond falls and when the market interest rate falls, the price of a bond increases. This is the “directional risk” of a bond which represents the sensitivity of the price (and therefore the yield) of a bond to changes in the interest rate. It is also referred to as the “delta risk” or “PV01 risk” [1].

17.    A bond trader is exposed to the PV01 risk in relation to any bonds held. The only way for a bond trader to eliminate that risk entirely is to execute a perfectly offsetting position - ie if one has bought €5m of 30-year BTP cash bonds, then to sell €5m of identical 30-year BTP cash bonds. However, the cash market is less liquid and balance sheet intensive. Traders will commonly seek, therefore, to hedge their risk with BTP futures.

18.    A BTP future is a contract by which the parties agree to, respectively, buy and sell cash BTP bonds (falling within certain qualifying criteria) at a future date at a fixed price. Holding a futures position can offer protection against directional risk.

19.    Cash bonds, because of the underlying capital sum involved, are balance sheet and capital intensive for banks. Futures, however, require limited capital until the futures contract expires and the cash bond(s) are delivered. Prior to delivery, banks post an initial margin, which is typically a fraction of the value of the underlying bond(s) (around 3%), on their balance sheet.

20.    There are several types of BTP future depending on the notional maturity date of the underlying cash BTP. This case concerns a particular type of BTP future called a “Long-Term Euro-BTP Future” (“BTP Future”) where the underlying BTP cash bonds have a maturity of 8.5-11 years. BTP Futures are sold in standard sizes called “lots”. One lot of BTP Futures is the economic equivalent of €100,000 of underlying cash BTPs with a remaining term of between 8.5 and 11 years and a 6% coupon. The minimum movement in the market is a single basis point, ie 0.01%. The value of a single point in respect of one lot is therefore €10.

21.    The exact number of lots of BTP Futures needed to hedge a cash BTP position will depend, inter alia, on both the notional value of the underlying cash BTPs being hedged, but also their maturity (ie a shorter maturity will require fewer lots and a longer maturity will require more lots). BTP Futures are traded electronically using the code “IK” on the EUREX market based in Frankfurt. The exception to this is that orders for 250 or more lots of IK can be privately negotiated bilaterally between counterparties off the exchange and when agreed will then be reported subsequently as “block trades” on the exchange.

22.    To give an example of hedging a cash BTP position using BTP futures: if a trader is “long” the market in respect of cash BTPs (ie the trader has just bought cash BTPs through trading), the trader is exposed to directional risk that the price of bonds fall. The trader can hedge this by selling BTP Futures. Equally, if a trader is “short” the market because the trader has sold cash BTPs, the trader can seek to offset the risk by buying BTP Futures.

23.    Trading in IK BTP futures on EUREX involves offers to both buy (known as the “bid” side) and sell (known as the “offer” or “ask” side) BTP futures. The best price available to buy or sell BTP futures is known as the “best bid” and “best offer” respectively and referred to as the “touch”. The difference between these two prices is known as the “bid-offer spread” or the “bid-ask spread”. When a buy or sell side trade is accepted by another market participant, it is described as being “filled” whether fully or partly.

24.    There is a functionality available for trading which allows a trader on EUREX to display only a part of the order that they wish to trade. This is called “iceberging”. Iceberging is a feature of a trader’s systems, rather than a feature of the EUREX Exchange. Where an order has been iceberged, a trader can choose the number of lots to display. Each time that the displayed part of the iceberged order executes, the EUREX Exchange automatically sends a message to the trader’s systems that the order has been executed, and then the trader’s system automatically places an order for the next tranche. This process will repeat until the order has traded in full or the order is cancelled. A trader is able to cancel an order, or amend its price, at any time before it has fully filled. The use of iceberging is widespread and well known.

MHI and the EGB Desk

25.    MHI is a UK subsidiary of the Japanese bank, Mizuho Financial Group. MHI’s EGB desk was based in London and comprised of five traders, the Applicants and two others. These two others, Mr James Hill and Mr Mehdi Barouti generally traded Core EGBs, which (as set out above) did not include BTPs.

26.    Mr Urra, as head of the EGB Desk reported to Mr Christian Heiberg, the head of Fixed Income Sales and Trading at MHI.

27.    MHI was not a Primary Dealer in respect of EGBs. It was also not a top-tier market maker in respect of cash EGBs but a “third-tier” marker maker. The principal purpose of its EGB Desk was to service client demand. Given its position in the market, MHI had less visibility over client flows in the market. It would receive fewer RFQs and also RFQs of smaller sizes than larger market makers. 


 

 

Glossary

 

Term

Description

Ask Side

A reference to sell orders on the market. Also referred to as the Offer Side.

Best Bid

The highest price offered by a market participant to buy a bond or future at a given time.

Best Offer

The lowest price offered by a market participant to sell a bond or future at a given time. Also referred to as the Best Ask.

Bid

A buy order entered on the exchange.

Bid-Offer Spread

The difference between the Best Bid and the Best Offer. Also referred to as the Bid-Ask Spread.

Bid Side

A reference to buy orders on the market.

Block Trade

A type of trade defined by EUREX as involving (insofar as concerns BTP Futures) an order size of 250 lots or more. Orders of this size could (although this was not mandatory) be traded directly between parties outside of the EUREX Exchange. If that occurred, the fact that a block trade had occurred was reported on the exchange’s trade feed at specified reporting times after completion.

BTP (or Cash BTP)

Buoni del Tesoro Poliennali, the relevant type of Italian Government Bond.

BTP Future

An interest-rate futures contract based, in this case, on a

notional BTP with a remaining term of between 8.5 and 11 years and a 6% coupon. There are other types of BTP future based on different notional maturity dates.

Core EGBs

EGBs issued by governments considered low risk of default, in particular Germany, France, the Netherlands, Belgium and Finland.

Coupon

The periodic interest payments due on a bond, expressed as a percentage of its face value.

Depth

Depth relates to the amount of liquidity available. The deeper a market, the more liquidity is available at various price points.

EGB

European Government Bond.

EUREX Exchange

The Frankfurt-based derivatives exchange on which the Applicants dealt in BTP Futures.

Future

A form of derivative contract based on an underlying financial asset, in this case, BTPs. The contract involves an agreement to buy or sell the underlying asset at a future date for a fixed price.

Hedging

A risk management strategy employed to offset losses in investments by taking an opposite position in a related asset.

High Frequency Traders or HFTs

Traders using algorithms, seeking to profit from high volume, low margin trading, often trading in small or very small amounts.

Hit

Selling at a bid price.

Iceberg Order

An order where only part of the total number of lots making up the order is visible to the market at any one time. Once

the visible part trades another part of the “iceberged” amount becomes visible. This occurs until the order trades in full.

Lift

Buying at an offer price.

Liquidity

Liquidity refers to the ease of buying and selling in volume. The more liquid the market, the more activity there will be from buyers and sellers at the relevant price point.

Lot

One futures contract in relation to €100,000 notional of underlying BTP Bonds.

Notional

The amount to be repaid by the issuer upon maturity of the bond (ie the face or par value).

Mandate

The mandate of the EGB Desk.

Maturity Date

The date on which the principal sum of a bond has to be repaid.

Offer

Sell orders entered on the exchange.

Offer Side

Another term to describe the Ask Side.

Primary Dealer

Generally large banks, Primary Dealers are the original purchasers of Government Bonds, including BTPs, which are then made available on the secondary market. Primary Dealers usually have the largest balance sheets and inventory of bonds and have the greatest access to liquidity, market intelligence and fees deriving from the syndicated issue of bonds.

PV01

The change in the price value of a bond based on a one basis point change in yield. The longer the maturity of a bond, the more sensitive its price will be to changes in yield. This is also described as “Delta” or “DV01”.

RFQ

“Request for Quote”: An inquiry from a client, either electronically or via call to the sales team, for the price at which a Market Maker is prepared to buy or sell a particular product.

Spread or the “Bid-Ask Spread” or “Bid-Offer Spread”

The difference between the Best Bid and the Best Offer.

Tick

A one point variation in the price of a future. This is the minimum price variation permitted by EUREX. The price moves up and down in “ticks”. Each tick represents a value of €10 per lot.

Touch

A term used to describe the Best Bid or Best Offer.

Volatility

The volatility of a market refers to the historical or expected variability in price of a market or security over a defined period of time. Generally, the less deep and liquid a market is, the more volatile it is.

Volcker Rule

A US banking rule prohibiting proprietary trading by banks.

Yield

The return on a bond, calculated by reference to its coupon and current market price. Price and yield move inversely to one another (ie the higher the yield of a bond, the lower its price, and vice versa).

 


 

 

Appendix 2

Example data for Trading Instances

 

This document sets out an example of the trading data available in relation to trading instance F7, being the earliest dated of the Specified Instances.


 

Example of a Specified Instance Narrative (Source: B3/4)

 

F7

 

1.          This is an instance of Mr Urra undertaking the abusive trading strategy alone on 7 June 2016.

 

2.          At 15:31:07, Mr Urra placed a sell order of 40 lots as an Iceberg Order, iceberged with a maximum show of 9 lots at a time, at what was the Best Bid (crossing the spread) (the Genuine Order).

 

3.          Almost instantaneously, the first and second tranches of the Genuine Order (a total of 18 lots) traded, and the third tranche of 9 lots became visible.

 

4.          Approximately 11 seconds later (the remaining 22 lots of the Genuine Order still not having traded, and sitting at the Best Offer), at 15:31:18, Mr Urra placed a buy order of 444 lots, 1 tick below the Best Bid (the Misleading Order) with all 444 lots visible to other market participants.

 

5.          Almost instantaneously after the Misleading Order had been placed (and within the following second), the third tranche of the Genuine Order (9 lots) traded, and the fourth (9 lots) and fifth (4 lots) tranches became visible and immediately traded (the Genuine Order thereby fully trading).

 

6.          Approximately one second after the Genuine Order finished trading, at 15:31:20, Mr Urra cancelled the Misleading Order.

 

7.          Through this conduct, Mr Urra gave or was likely to give a false or misleading impression to other market participants that he wanted to buy BTP Futures through the Misleading Order.  In fact, Mr Urra wanted to sell 40 lots of BTP Futures (through the Genuine Order).

 

8.          The features of this example that demonstrate that Mr Urra did not intend to trade his Misleading Order and thus gave or was likely to give a false or misleading impression to the market include the following:

 

8.1        Mr Urra placed the Misleading Order after the remaining 22 lots of the Genuine Order remained on the market for approximately 11 seconds at the Best Offer without trading.

 

8.2        He placed the Misleading Order close enough to the Best Bid to be visible to other market participants and to increase the pressure on the order book, without being at Best Bid, so that other market participants would be less likely to trade the order.

 

8.3        He cancelled the Misleading Order 2 seconds after he placed it, which was 1 second after the Genuine Order traded.

 

8.4        He fully executed the Genuine Order of 40 lots and did not execute any of the Misleading Order of 444 lots.

 

8.5        Mr Urra did not place the Misleading Order as an Iceberg Order, with the result that all 444 lots were visible to the market. It is implausible that Mr Urra would consider an order of this size to be incapable of impacting the market at the same time as he considered it necessary to enter his considerably smaller Genuine Order as an Iceberg Order.

 


 

Example of the Annex I extract (Source: B3/2)

 

 

 

 

A table with a number of people

AI-generated content may be incorrect.

 

 

 


 

Example of the Replay graph (Source: B3/2)

 

 

 

 

A screen shot of a graph

AI-generated content may be incorrect.

 

 

 

 

 


 

Example of the Spread graph(s) (Source: B3/2)

 

 

 

 

 

 

 

A screen shot of a graph

 

 

 

 

 

 


 

Cash Bond Trades (Source: H1/1)

 

 

 

 

 

 

 


 

RFQs (Source: H1/2)

 

 

 

 

 

 

 

 


 

Intraday Price Graph (Source: D/4)

 

 

 

A graph showing a line graph

AI-generated content may be incorrect.

 

 



[1] Present value of a basis point.  In evidence, it has also been referred to as DV01 (delta value of a basis point).

About BAILII - FAQ - Copyright Policy - Disclaimers - Privacy Policy amended on 25/11/2010