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

England and Wales High Court (Chancery Division) Decisions


You are here: BAILII >> Databases >> England and Wales High Court (Chancery Division) Decisions >> Britned Development Ltd v ABB [2018] EWHC 2616 (Ch) (09 October 2018)
URL: http://www.bailii.org/ew/cases/EWHC/Ch/2018/2616.html
Cite as: [2018] EWHC 2616 (Ch)

[New search] [View without highlighting] [Printable PDF version] [Help]


Neutral Citation Number: [2018] EWHC 2616 (Ch)

Claim No. HC-2015-000268

IN THE HIGH COURT OF JUSTICE

BUSINESS AND PROPERTY COURTS OF ENGLAND AND WALES

BUSINESS LIST (ChD)


Royal Courts of Justice

Rolls Building

Fetter Lane

London EC4A 1NL


Date: 9 October 2018

 

Before :

 

THE HONOURABLE MR JUSTICE MARCUS SMITH

- - - - - - - - - - - - - - - - - - - - -

Between :

 

BRITNED DEVELOPMENT LIMITED
Claimant

- and -


(1) ABB AB
(2) ABB LTD
Defendants

 

 

- - - - - - - - - - - - - - - - - - - - -

- - - - - - - - - - - - - - - - - - - - -

 

Mr Robert O’Donoghue, QC and Mr Hugo Leith (instructed by Squire Patton Boggs (UK) LLP ) for the Claimant

Mr Mark Hoskins, QC , Ms Sarah Ford, QC and Ms Jennifer MacLeod (instructed by Freshfields Bruckhaus Deringer LLP ) for the Defendants

 

Hearing dates: 7-9, 12-16, 19-23 February, 5-6 March 2018

- - - - - - - - - - - - - - - - - - - - -

Judgment Approved


 

CONTENTS

A.

INTRODUCTION

Para. 1

(1)

The Cartel

Para. 1

(2)

BritNed’s claim and ABB’s response to that claim

Para. 4

(3)

Approach and the structure of this Judgment

Para. 9

B.

LEGAL PRINCIPLES AND APPROACH

Para. 10

(1)

Elements of the tort

Para. 10

(2)

A preliminary pleading point: the Overcharge Claim and the definition of “overcharge”

Para. 13

(3)

A presumption of overcharge and the principle of effectiveness

Para. 19

(4)

ABB’s prior bad conduct

Para. 24

C.

THE EVIDENCE

Para. 26

(1)

Introduction

Para. 26

(2)

Factual witness evidence

Para. 28

(a)

General points

Para. 28

(b)

BritNed’s factual witnesses

Para. 32

(i)

Mr Mathew Rose

Para. 33

(ii)

Mr Michael Jackson

Para. 36

(c)

ABB’s factual witnesses

Para. 39

(i)

Mr Hans-Åke Jönsson

Para. 40

(ii)

Mr Peter Leupp

Para. 47

(iii)

Mr Stefan Ekman

Para. 52

(iv)

Mr Magnus Larsson-Hoffstein

Para. 56

(v)

Mr Hans-Magnus Röstlund

Para. 61

(3)

The documentary evidence

Para. 64

(4)

The Decision of the European Commission

Para. 67

(a)

The law regarding the bindingness of decisions

Para. 67

(b)

My approach in the case of the Decision

Para. 68

(5)

The expert evidence

Para. 72

(6)

Overview

Para. 76

D.

THE NATURE AND OPERATION OF THE CARTEL

Para. 80

(1)

General nature and operation

Para. 81

(2)

ABB and the Cartel

Para. 90

E.

THE CHARACTERISTICS OF SUBMARINE CABLES

Para. 97

(1)

Introduction

Para. 97

(2)

The difference between submarine and underground cable projects

Para. 98

(3)

Alternating current or direct current

Para. 103

(4)

Cable width

Para. 108

F.

THE HISTORY OF THE BRITNED TENDER

Para. 111

(1)

Genesis

Para. 111

(2)

The lots

Para. 114

(3)

Transmission capacity

Para. 115

(4)

The OJEU Notice (August 2005)

Para. 117

(5)

BritNed’s negotiating team

Para. 118

(6)

Expressions of interest (September to October 2005)

Para. 120

(7)

ABB’s tender team and its approach to tendering

Para. 122

(a)

The ABB tender team

Para. 122

(b)

ABB’s methodology during the “technical” process

Para. 126

(i)

Pre-tender communications

Para. 128

(ii)

Initial design on tender

Para. 129

(iii)

Costings

Para. 130

(iv)

Risk review

Para. 133

(v)

Submission of a tender

Para. 134

(vi)

Use of Product Pricing Models

Para. 135

(c)

The process post-tender

Para. 136

(8)

Pre-qualification to tender

Para. 137

(9)

Tender evaluation “Phase One”

Para. 138

(10)

ABB’s discussions with Nexans

Para. 141

(11)

The ILEX Study (May 2006)

Para. 142

(12)

Prysmian did not submit a tender (June 2006)

Para. 143

(13)

Areva did not submit a tender (June 2006)

Para. 146

(14)

Nexans’ tender (June/July 2006)

Para. 147

(15)

ABB’s tender (June 2006)

Para. 148

(16)

BritNed’s position at this point

Para. 149

(17)

Evaluation by BritNed

Para. 150

(18)

Tender evaluation “Phase Two” (July 2006 onwards)

Para. 152

(19)

Best and final offers (March 2007)

Para. 158

(20)

Last and final offers (March 2007)

Para. 160

(21)

Final cable price negotiations (April 2007)

Para. 161

(22)

Letter of intent and contract

Para. 167

(23)

The price offered by ABB over time

Para. 169

G.

COMPETITIVE PRESSURES ARISING IN THE NEGOTIATION PROCESS

Para. 171

(1)

Introduction

Para. 171

(2)

The limited response from bidders that BritNed had to contend with

Para. 173

(3)

The various commercial pressures that BritNed sought to deploy during the negotiations

Para. 176

(a)

The commercial pressures identified

Para. 176

(b)

Keeping Siemens in

Para. 177

(c)

Full bipole versus simple bipole

Para. 182

(d)

Comparison with NorNed

Para. 184

(e)

Risk of the project not going ahead

Para. 188

(f)

ABB’s lack of its own cable laying vessel

Para. 190

(g)

Perception that ABB needed the work

Para. 197

(4)

The advantage ABB derived from its participation in the Cartel

Para. 201

(a)

Introduction

Para. 201

(b)

BritNed’s position

Para. 204

(c)

Approach to analysing ABB’s knowledge advantage

Para. 211

(d)

Who knew about the Cartel and who did not know about the Cartel?

Para. 220

(e)

What these persons knew regarding the BritNed bid?

Para. 226

(i)

The fact that projects were allocated by the Cartel and that BritNed was allocated to ABB

Para. 227

(ii)

Likely absence of competition from Japan/Korea

Para. 230

(iii)

Level of competition from European suppliers as regards cable projects

Para. 234

(iv)

Level of competition as regards the Converter element and Lot 3

Para. 243

(v)

Conclusion

Para. 247

(f)

Whether, and if so how, any “useful” knowledge regarding competition could have been deployed within ABB

Para. 249

(i)

Introduction

Para. 249

(ii)

Direct influence on the level of the bid

Para. 252

 

ABB’s process and the distinction between direct costs and common costs

Para. 252

 

Influencing the level of direct costs

Para. 255

 

Influencing the level of common costs

Para. 266

(iii)

Indirect influence on the level of the bid

Para. 273

 

Communication of market capacities

Para. 275

 

Affecting the commercial approach of ABB

Para. 279

(g)

Conclusions

Para. 284

H.

THE EXPERTS’ ANALYSIS

Para. 285

(1)

Introduction

Para. 285

(2)

The approach of Dr Jenkins

Para. 288

(a)

A summary of Dr Jenkins’ approach

Para. 288

(b)

Statistical tools and devices

Para. 293

(i)

What is a coefficient?

Para. 293

(ii)

Correlation

Para. 294

(iii)

Measuring correlation

Para. 295

(iv)

Regression analysis

Para. 299

(v)

Testing how good a model is: statistical significance

Para. 303

(vi)

The Practical Guide on Quantifying Harm

Para. 309

(vii)

One-sided and two-sided hypothesis testing

Para. 310

(c)

Dr Jenkins’ dataset

Para. 312

(d)

Information regarding each project forming part of Dr Jenkins’ dataset

Para. 314

(e)

Dr Jenkins’ process of normalisation

Para. 316

(f)

Dr Jenkins’ regression analysis

Para. 318

(3)

The approach of Mr Biro

Para. 321

(a)

A summary of Mr Biro’s approach

Para. 321

(b)

Mr Biro’s first approach: a price comparison analysis controlling for ABB’s actual costs of supply

Para. 323

(i)

Overview

Para. 323

(ii)

Comparable projects

Para. 325

(iii)

Multiple margins and margin consolidation

Para. 326

(iv)

Calculation of gross margin

Para. 327

(c)

An econometric analysis of the relationship between prices and ABB’s actual costs of supply

Para. 334

(d)

A price comparison analysis which does not directly control for ABB’s actual costs of supply, but instead uses proxy measures based on the technical characteristics of the projects

Para. 339

(5)

A critical assessment of the two approaches

Para. 343

(a)

Introduction

Para. 343

(b)

The reliability of the ABB data

Para. 349

(i)

Dr Jenkins’ criticisms

Para. 349

(ii)

Reliability of direct costs

Para. 356

(iii)

Reliability of common costs

Para. 361

(iv)

Baked-in inefficiencies

Para. 365

(v)

Cartel savings

Para. 369

(vi)

Lost bids

Para. 372

(vii)

Successful Cartel submarine bids

Para. 374

(c)

Reliability of Dr Jenkins’ model

Para. 376

(i)

The relevant uncertainties

Para. 376

(ii)

The inclusion of Cartel projects

Para. 382

(iii)

The inclusion of underground cable projects

Para. 388

(iv)

Exclusion of the time trend variable

Para. 398

(v)

Exclusion of the order backlog variable

Para. 402

(d)

Which approach is preferable?

Para. 414

(i)

The reliability of Mr Biro’s model

Para. 414

(ii)

The reliability of Dr Jenkins’ model

Para. 417

(iii)

Appropriateness

Para. 419

I.

MY ASSESSMENT OF THE OVERCHARGE

Para. 422

(1)

The “gist” damage: accrual of a cause of action

Para. 422

(2)

Quantification of the overcharge

Para. 430

(a)

Introduction

Para. 427

(b)

Was there a directly influenced overcharge in ABB’s common costs?

Para. 438

(c)

Baked-in inefficiencies

Para. 445

(d)

Cartel savings

Para. 454

(e)

Was there an indirect influence on ABB to be uncompetitive?

Para. 459

(3)

Conclusion

Para. 464

J.

THE LOST PROFIT CLAIM

Para. 466

(1)

Introduction

Para. 466

(2)

The law

Para. 468

(3)

The facts

Para. 470

(a)

Introduction

Para. 470

(b)

BritNed’s default position

Para. 472

(c)

Non-price factors affecting capacity choice

Para. 475

(i)

Unproven technology

Para. 476

(ii)

Bigger is not necessarily better

Para. 477

(iii)

A limit of 1,320MW

Para. 480

(iv)

A flexible capacity

Para. 481

(v)

Conclusion

Para. 482

(d)

The need to make a choice between Base Cases

Para. 483

(e)

The choice actually made by BritNed

Para. 487

(4)

The counterfactual assessment

Para. 497

(a)

Dr Jenkins’ approach

Para. 497

(b)

At what price would ABB have tendered for Base Case 3?

Para. 502

(5)

Conclusion

Para. 507

K.

THE REGULATORY CAP ISSUE

Para. 509

(1)

The pleaded case

Para. 509

(2)

The relevant provisions

Para. 512

(a)

Regulation 1228/2003

Para. 512

(b)

Directive 2003/54/EC

Para. 516

(c)

Domestic provisions

Para. 519

(d)

BritNed’s application for an exemption

Para. 520

(3)

The arguments

Para. 524

(a)

ABB’s contentions

Para. 524

(b)

BritNed’s response

Para. 527

(4)

The proper construction of the IRR Cap

Para. 529

(5)

The law

Para. 531

(a)

A brief statement of the relevant principles

Para. 531

(b)

Has BritNed suffered any loss at all?

Para. 534

(c)

British Transport Commission v. Gourley

Para. 538

(d)

Conclusion

Para. 541

(6)

Assessing the future effect of the IRR Cap

Para. 542

L.

THE COMPOUND INTEREST CLAIM

Para. 543

M.

DISPOSITION

Para. 550

TABLES AND FIGURES

Table 1

The price offered by ABB over the course of the negotiations. Prices are exclusive of VAT

Para. 169

Table 2

ABB bid prices as evidenced by PPMs

Para. 258(4)

Table 3

The dataset used by Dr Jenkins

Para. 312

Table 4

Control variables used by Dr Jenkins

Para. 318

Table 5

Outcome of Dr Jenkins’ regression analysis

Para. 319

Table 6

Outcome of Mr Biro’s margin analysis

Para. 331

Table 7

Variables used by Mr Biro

Para. 335

Table 8

Outcome of Mr Biro’s regression analysis

Para. 337

Table 9

Effect on Dr Jenkins’ model(s) of certain adjustments by Mr Biro

Para. 378

Figure 1

95% two-sided confidence interval for the output of Dr Jenkins’ model

Para. 418(3)

Table 10

Confidence intervals for the output of Dr Jenkins’ model

Para. 418(3)

Table 11

Selected margin data from Table 6

Para. 443(2)

Table 12

Effect of interest on gross margin

Para. 452(3)

Table 13

Overcharge due to baked-in inefficiencies

Para. 453(4)

Table 14

Further margin and direct cost analysis

Para. 462

Table 15

Per MW prices

Para. 491

Figure 2

Illustration of ABB’s IRR regulatory cap argument

Para. 525(2)

ANNEXES

Annex 1

Terms and abbreviations used in the Judgment

Annex 1

 

Mr Justice Marcus Smith:

 

A.      INTRODUCTION

(1)     The Cartel

1.           By a decision of the European Commission dated 2 April 2014 in Case AT.39610 – Power Cables (the “Decision” [1] ), the European Commission found the existence of a cartel (the “Cartel”) in relation to (extra) high voltage submarine and underground power cable projects.

2.           The Cartel was global, with participants from Europe, Japan and Korea. Amongst the undertakings participating in the Cartel were the Defendants, collectively referred to herein as “ABB”.

3.           The Cartel operated between 1999 and 2009.

(2)     BritNed’s claim and ABB’s response to that claim

4.           The Claimant – “BritNed” – is jointly owned by National Grid and TenneT, the operators of the UK and Dutch electricity grids. BritNed owns and operates the BritNed “Interconnector”, a 1,000-megawatt (“MW”) capacity electricity submarine cable system connecting the Dutch and UK electricity grids. It was constructed between 2009-2010.

5.           BritNed was a customer of ABB. ABB supplied the cable element of the BritNed Interconnector (the “Cable”). The other, significant, element of the BritNed Interconnector was the converter element (the “Converter”), which ABB was not asked to supply by BritNed, but which ABB tendered for.

6.           BritNed contends that, by reason of the Cartel and its operation, it has suffered loss and damage. It brings a claim in tort against ABB, for breach of statutory duty, specifically a restriction of competition by object contrary to Article 101 of the Treaty on the Functioning of the European Union (“TFEU”) and Article 53 of the Agreement on the European Economic Area (“EEA Agreement”).

7.           BritNed claims three heads of loss:

(1)         Overcharge. BritNed contends that, as a result of the Cartel, the price it paid for the cable element of the BritNed Interconnector was higher than it otherwise would have been.

(2)         Lost profit. BritNed contends that, absent the Cartel, it would have acquired a cable of a higher capacity – 1,320MW rather than 1,000MW – which would have generated additional revenues and higher profits than the 1,000MW cable actually purchased.

(3)         Interest. BritNed contends that, as a result of the overcharge (paragraph 7(1) above), it incurred higher capital costs in commissioning the Interconnector than would otherwise have been the case under competitive conditions. BritNed claims compound interest on this basis.

I shall refer to these three heads of loss as, respectively, the “Overcharge Claim”, the “Lost Profit Claim” and the “Compound Interest Claim”.

8.           ABB cannot, and does not, deny the existence of the Cartel, nor its participation in the Cartel. Nevertheless, all three of BritNed’s heads of claim are disputed:

(1)         Response to the Overcharge Claim. ABB does not accept that the operation of the Cartel resulted in any overcharge to BritNed in relation to the BritNed Interconnector. It contends that the Cartel had no effect on the prices charged to BritNed for the Cable element of the BritNed Interconnector.

(2)         Response to the Lost Profit Claim. ABB denies that, absent the Cartel, BritNed would have acquired different capacity cabling.

(3)         Response to the Compound Interest Claim. It follows, from ABB’s denial of BritNed’s Overcharge and Lost Profit Claims, that any claim to interest is also denied. However, if (contrary to ABB’s primary position) there was an overcharge, then ABB disputes that compound interest is recoverable by BritNed.

(4)         A regulatory cap on earnings. Because of the regulatory regime to which BritNed was subject, and which is described in greater detail in this Judgment, ABB contends that if (contrary to ABB’s primary position) BritNed does have a claim for damages, those damages should be assessed in light of a regulatory cap imposed on BritNed’s earnings. This point is really an aspect of quantification of BritNed’s loss, but it makes sense to treat it separately. I shall refer to the point as the “Regulatory Cap Issue”.

(3)     Approach and the structure of this Judgment

9.           The points in dispute between the parties are thus both extensive and fundamental. This Judgment approaches matters in the following way:

(1)         Although the cause of action on which BritNed’s claim is based – breach of statutory duty – might be said to be straightforward, the interrelationship between a breach of duty and the quantification of loss arising out of that breach is not. Section B considers various questions in relation to the tort, including the relationship between liability and quantum and (in a point that arose during the course of the trial) precisely the nature of the overcharge being alleged by BritNed.

(2)         Section C considers the evidence that was presented to me during the course of the trial and on the basis of which I must make the necessary findings of fact. The evidential material before me comprised four discrete strands: (i) factual witness evidence; (ii) contemporary documentary evidence; (iii) the Decision of the European Commission; and (iv) the expert evidence. Section C describes and assesses this material.

(3)         The next Sections contain the bulk of the analysis to enable me to determine BritNed’s various claims. That analysis has various strands:

(a)            Section D considers the nature and operation of the Cartel in general terms and ABB’s role in the Cartel in particular.

(b)           Section E considers the characteristics of submarine cables: the Cable element of the BritNed Interconnector was (unsurprisingly, given the geography) submarine. The distinctions and similarities between submarine and underground cables are points of considerable significance when the views of the experts come to be considered.

(c)            Section F considers the history of the tendering process commenced and run by BritNed for the supply and provision of the BritNed Interconnector.

(d)           Section G considers the competitive pressures that arose – or (as importantly) that did not arise, because of the Cartel – during the process of negotiation between BritNed and ABB. This Section also considers the effect of the Cartel within ABB and how ABB’s participation in the Cartel affected its negotiations with BritNed.

(e)            Section H considers the experts’ assessment of the overcharge. This Section sets out the approach of each expert, and then seeks critically to assess each approach in terms of its viability in assisting me in assessing what, if any, was the overcharge in this case.

(4)         Section I, drawing on the findings made in Sections D to H, sets out my assessment of, and conclusions in relation to, the overcharge and determines the Overcharge Claim. The next Sections consider and determine, respectively, the Lost Profit Claim (Section J), the Regulatory Cap Issue (Section K) and the Compound Interest Claim (Section L).

Finally, Section M sets out in brief the conclusions that I have reached and how this dispute is to be disposed of.

B.      LEGAL PRINCIPLES AND APPROACH

(1)     Elements of the tort

10.         In English law, competition law infringements are vindicated as statutory torts. To establish a claim, two things must be shown: (i) an infringement of competition law; and (ii) actionable harm or damage, caused by that infringement. [2] As has been stated in the context of the tort of negligence – but the point holds good for breach of statutory duty – “[i]t is a truism that a fundamental requirement for a claim in negligence is that the plaintiff has suffered some past “damage”. A breach of duty by the defendant is not enough. The cause of action will not accrue until actionable damage occurs. The damage is said to form the gist of the action. Recovery is not limited to this threshold “gist damage”, but without it there is no cause of action.” [3] Proving actionable damage inevitably involves demonstrating a causal link between the infringement and the damage, generally using the “but for” test of causation. [4]

11.         It is often said that in terms of what a claimant has to prove in order to recover a loss depends upon the distinction between past facts on the one hand and future events or hypothetical actions on the other. In Mallett v. McMonagle , [5] Lord Diplock said this:

“In determining what did happen in the past a court decides on the balance of probabilities. Anything that is more probable than not it treats as certain. But in assessing damages which depend upon its view as to what will happen in the future or would have happened in the future if something had not happened in the past, the court must make an estimate as to what are the chances that a particular thing will or would have happened and reflect those chances, whether they are more or less than even, in the amount of damages which it awards.”

12.         This is a helpful summary of the law. In order to apply this statement in the present case, it is necessary to unpack a little further the implications of Lord Diplock’s statement:

(1)         It is essential to be clear as to what the elements of the claimant’s cause of action are. These elements will – inevitably – be in the past. If they are not in the past, then the claimant’s cause of action will not have accrued, and (at best) the claimant will be entitled to some form of quia timet relief.

(2)         Damage is by no means always a pre-requisite for a complete cause of action. It is not necessary to show loss or damage to bring an action for breach of contract, but it is necessary for an action in the tort of negligence or for breach of statutory duty. [6]

(3)         Where loss or damage is a necessary element of the cause of action, it must be borne in mind that in some cases the law treats the loss of a chance of a favourable outcome as compensable damage in itself. [7]

(4)         Since the elements of the cause of action are in the past, they must be proved on the balance of probabilities. As regards such questions, the court adopts an “all-or-nothing” approach: [8]

“When the question is whether a certain thing is or is not true – whether a certain event did or did not happen – then the court must decide one way or the other. There is no question of chance or probability. Either it did or it did not happen. But the standard of civil proof is a balance of probabilities. If the evidence shows a balance in favour of it having happened then it is proved that it did in fact happen.”

(5)         If and when the cause of action is established, then inevitably the inquiry is driven to the hypothetical. The measure of loss, in the case of contract, is the amount of damages that will place the claimant in the same situation as if the contract had been performed. [9] The measure of loss, in tort cases, is the amount of damages that will place the claimant in the situation he or she would have been in, had the tort not been committed. [10] Both of these inquiries involve an assessment of what would have happened in a hypothetical or counter-factual case – a case where the contract was performed or where the tort was not committed, so that the claimant’s damages can be quantified.

(6)         During this quantification exercise, English law moves away from the balance of probabilities. An assessment or quantification of damages involves the taking into account of all manner of risks and possibilities. [11] Of course, “loss of a chance” analysis may be appropriate when quantifying a claimant’s loss, but that is by no means the only tool or even the most useful tool that is available to the court. Fundamentally, the process is evidence driven, and it is difficult to be very prescriptive. As Popplewell J noted in Asda Stores Ltd v. Mastercard Inc , [12] “the court takes a pragmatic approach”.

(7)         The Asda decision helpfully sets out the approach that courts take to questions of quantification. It was suggested by BritNed that this articulation of the law did not apply in the present case, on grounds that Asda was an “effects” case, and this case is not. [13] It was suggested that – because of the information asymmetry that existed between BritNed and ABB, some other approach should be taken. I do not accept this contention. I consider that Asda is doing no more than articulate principles relevant to the quantification of loss generally, albeit with an emphasis on the quantification of loss in competition cases. Indeed, it will be noted that Popplewell J’s articulation of the relevant principles emphasises that a lack of information should not prevent a quantification. In short, I consider Popplewell J’s articulation of the principles a helpful one for the purposes of this case.

(8)         The following articulation of principles draws on Popplewell J’s articulation at [306] of Asda . I have not repeated the citation of the authorities:

(a)            Only as much certainty and particularity is insisted on in proof of damage as is reasonable, having regard to the circumstances and to the nature of the acts by which the damage is done. [14]

(b)           The fact that it is not possible for a claimant to prove the exact sum of its loss is not a bar to recovery. In many cases, the assessment of damages will involve an element of estimation and assumption. Restoration by way of compensatory damages is often accomplished by “sound imagination” and a “broad axe” or a “broad brush”. The court will not allow an unreasonable insistence on precision to defeat the justice of compensating a claimant for infringement of its rights. [15]

(c)            Indeed, I would add that the exercise of “sound imagination” will involve the court in trying to understand the overall context in which the alleged harm was suffered. The broad brush must be used to paint a canvass that is a consistent and rational portrayal of circumstances in which the claimant and the defendant operated, so that the central question (what would have happened, had the tort not been committed) is answered in its context. This is not unlike the “theory of harm” that is articulated by regulators when seeking to ascertain whether there has been a competition law infringement.

(d)           Popplewell J found value in the following passages from a Commission Staff Working Document Practical Guide on Quantifying Harm in Actions for Damages (the “Practical Guide on Quantifying Harm”), as do I: [16]

“16.   It is impossible to know with certainty how a market would have exactly evolved in the absence of the infringement of Article 101 or 103 TFEU. Prices, sales volumes, and profit margins depend on a range of factors and complex, often strategic, interactions between market participants that are not easily estimated. Estimation of the hypothetical non-infringement scenario will thus by definition rely on a number of assumptions. In practice, the unavailability of data will often add to this intrinsic limitation.

17.     For these reasons, quantification of harm in competition cases is, by its very nature, subject to considerable limits as to the degree of certainty and precision that can be expected. There cannot be a single “true” value of the harm suffered that could be determined, but only best estimates relying on assumptions and approximations…”

Quantification of loss is not a question of mathematical calculation (although mathematical calculations will, no doubt, have their place), but turns on developing a robust understanding of what would have happened in the counterfactual case.

(9)         In [307] of Asda , Popplewell J said this: [17]

“…where the court is compelled to use a broad brush in the absence of precision in the evidence of the harm suffered by a claimant, it should err on the side of under-compensation so as (a) to reflect the uncertainty as to the loss actually suffered and (b) to give the defendant the benefit of any doubts in the calculation”.

The claimant’s compensation cannot simply be “plucked from the air”. It must be grounded in the evidence before the court. The court must, when quantifying loss, be astute to identify those points where the evidence falls short, and where the court becomes reliant upon estimates or assumption. Such estimates or assumptions will need to take account of the fact that the probabilities in the counter-factual world may not mean that these estimates or assumptions will inevitably hold good. [18] I do not take this dictum to mean that every calculation made in the course of assessment of damages must be reduced to avoid the risk of over-compensation.

(2)     A preliminary pleading point: the Overcharge Claim and the definition of “overcharge”

13.         In this case, the infringement of competition law has been established by the Decision. But the question of whether that infringement has caused actionable harm or damage to BritNed is in dispute and for me to determine.

14.         So far as the Overcharge Claim is concerned, the Particulars of Claim plead that BritNed paid a price for the Interconnector “that was unlawfully inflated above the price which would have prevailed had there been no Cartel”. [19]

15.         There was a dispute between the parties as to how this unlawful inflation of price – the overcharge – was to be assessed. Two alternatives were contended for:

(1)         The overcharge was the difference between the price actually agreed and the price that would have been agreed between ABB and BritNed had there been no Cartel; alternatively

(2)         The overcharge was the difference between the price actually agreed and the price that would have resulted had there been no Cartel whether the party contracting with BritNed would have been ABB or some other supplier.

16.         ABB contended that the recoverable overcharge was the first of these two alternatives. ABB’s position is stated in its written closing submissions: [20]

“296. BritNed’s pleaded claim in respect of overcharge is as follows:

“As a result of the activities carried out by ABB through their participation in the Cartel from 1 April 2000 to 17 October 2008, BritNed: (a) paid a price under the Agreement that was unlawfully inflated above the price which would have prevailed had there been no Cartel.”

297.   The Agreement in question is the contract between BritNed and ABB for the BritNed Works signed on 21 May 2007.

298.   Both Mr Biro and Dr Jenkins, in considering the competitive price absent the Cartel, had used ABB’s data from before and after the Cartel: they have therefore proceeded on the basis that the competitive price is the price that ABB would have charged to BritNed absent the Cartel (the “ABB counterfactual price”).

299.   Although its position is not clear, during the course of the trial BritNed may have posited a further counterfactual, based on the premise that a third party other than ABB would have won the project at a price lower than the ABB counterfactual price.

300.   ABB submits that it is not open to BritNed to advance this argument for the first time during the course of the trial. It has not been considered by the experts and ABB has not had the opportunity to address it, by way of disclosure or factual or witness evidence.”

BritNed contended for the second of the two alternatives. [21]

17.         Considering the terms of BritNed’s pleading – which, as it seems to me, is certainly the crucial and perhaps the only question that arises – I have no doubt that BritNed’s pleaded case on overcharge defines the overcharge as the second of the two alternatives described in paragraph 15 above, that is the difference between the price actually agreed and the price that would have resulted had there been no Cartel , whoever the party contracting with BritNed would have been in the counter-factual world. As to this:

(1)         Paragraph 5D of the Particulars of Claim pleads some of the characteristics of the Cartel, including the fact that the BritNed project was allocated to ABB by the other cartelists, and that the other cartelists would either refrain from bidding or else submit uncompetitive bids.

(2)         The counterfactual scenario which must, therefore, be considered, is one where ABB was not “allocated” the BritNed project. That obviously implies competitive tenders from others, which (i) might render ABB more competitive, but which (ii) might result in a competitor putting forward a more competitive price than ABB and thereby winning the contract.

(3)         I do not consider that paragraph 7(a) of the Particulars of Claim confines BritNed’s claim to the more competitive price that ABB might have offered. Paragraph 7(a) pleads that the price under the Agreement was unlawfully inflated, and that this price was inflated above the price which would have prevailed had there been no Cartel. This counter-factual price is not limited to the price that would have been offered by ABB, and I see no reason for implying such a restriction into the Particulars of Claim.

(4)         It is true that both parties have focussed on ABB’s costs and how – in a competitive market – ABB’s price might have changed. That I consider to be a reflection of the evidence available to the parties, rather than a consequence of BritNed’s pleading. ABB has provided, on disclosure, a great deal of evidence regarding the other projects it was involved in and the costs associated with these projects. This has been considered – as I described – by the experts. There has been no corresponding disclosure from ABB’s competitors, and none could reasonably have been expected by either party. [22] Inevitably, the experts and the parties have done what they can on the evidence available to them; but that does not mean that the counterfactual inquiry is limited to a consideration of what price ABB would have offered. Such an approach is tantamount to treating the Cartel as if it still operated, at least to the extent of preventing competitive bids from suppliers other than ABB.

18.         Accordingly, the overcharge that I am seeking to assess is the difference between (i) the price agreed between ABB and BritNed and (ii) the price that would have been agreed – whether with ABB or by another provider – had the Cartel not operated. That said, for the reasons given in paragraph 17(4) above, the sort of price that a third-party provider would offer is extremely difficult to determine, given the (lack of) evidence. Inevitably, that has a bearing on my approach to the assessment of the overcharge.

(3)     A presumption of overcharge and the principle of effectiveness

19.         Directive 2014/104/EU (the “Damages Directive”), which has been implemented into English law by an amendment to the Competition Act 1998, requires Member States to establish a presumption of harm in cartel damages cases. [23] The rationale for this is explained in Recital (47) of the Damages Directive:

“To remedy the information asymmetry and some of the difficulties associated with quantifying harm in competition law cases, and to ensure the effectiveness of claims for damages, it is appropriate to presume that cartel infringements result in harm, in particular via an effect on prices. Depending on the facts of the case, cartels result in a rise in prices, or prevent a lowering of prices which would otherwise have occurred but for the cartel. This presumption should not cover the concrete amount of harm. Infringers should be allowed to rebut the presumption. It is appropriate to limit this rebuttable presumption to cartels, given their secret nature, which increases the information asymmetry and makes it more difficult for claimants to obtain the evidence necessary to prove the harm.”

20.         Paragraph 13 of Schedule 8A to the Competition Act 1998 now provides:

“For the purposes of competition proceedings, it is to be presumed, unless the contrary is proved, that a cartel causes loss or damage.”

21.         BritNed accepted that neither the Damages Directive nor the amendment to the Competition Act 1998 applied in this case. Paragraph 13 of Schedule 8A applies to cases arising after March 2017 and so does not apply to this case.

22.         Nevertheless, BritNed contended that the principle of effectiveness requires a presumption of harm. BritNed invited me (rebuttably) to presume that the Cartel had caused loss to BritNed.

23.         I reject this contention:

(1)         Clearly, when the amendments to the Competition Act 1998 pertain, they will be applied. They do not apply now.

(2)         If the principle of effectiveness required the creation of a presumption of loss or damage, then it is difficult to see why the Damages Directive requires such a presumption to be established and why an amendment to the Competition Act 1998 has been effected.

(3)         I do not consider that a presumption of harm particularly assists in the assessment of damages in cartel cases, and I certainly do not consider that it is appropriate for me to pre-empt legislation specifically introducing into future cases this presumption. That is especially so, given the approach to the assessment of damages taken by English law, described in paragraph 12 above and expanded upon in paragraphs 419 ff below.

(4)         In its written closing submissions, BritNed suggested that the presumption of harm had a “strong and sound basis in economics” [24] and that “the facts as established at trial provide a compelling basis for the court to apply a presumption that the Cartel did have some effect on the BritNed price”. [25] This strikes me as a somewhat arid and rather circular point. I will obviously have regard to all of the facts and evidence, including economic theory regarding the operation of cartels, in order to determine whether there has been an overcharge and, if so, how much. If the economic analysis and the facts are as compelling as BritNed contend (and this, of course, is the substance of my analysis in Sections D to I below) then BritNed will establish an overcharge without the need to rely on a presumption. If, on the other hand, the economic analysis and the facts are less cogent, then I fail to see why (absent legislation compelling me) I should buttress an otherwise weak case with a presumption that there has been such loss and damage.

(5)         Obviously, I take the point about informational gaps, and the potential asymmetry in information that will exist between a cartel member and an outsider. This issue, however, is fully factored into the approach English courts take to the quantification of loss and damage. I fail to see how a bare presumption of harm – particularly one, which does not involve a presumed quantification of harm – takes matters any further at all.

(4)     ABB’s prior bad conduct

24.         In opening, BritNed stated that ABB was a “dirty company”: [26]

“My Lord, in 22 years of practice, this is not a word I have used before any tribunal, but I’m afraid to say that ABB is a dirty company. For over two decades, its power business has involved itself in at least three pernicious global cartels, covering power cables, gas-insulated switchgear and power transformers, and a fourth cartel actually in pre-insulated pipes, which is something slightly different. These cartels were organised and implemented at the highest levels within ABB.”

The suggestion that the court should have regard to the fact that the Cartel was preceded by other cartels involving ABB’s power division was reiterated in BritNed’s written closing submissions. [27]

25.         ABB contended that this was a jury point to be disregarded. [28] I agree:

(1)         Whilst it may be that ABB’s participation in other illegal cartels might amount to “similar fact” evidence as to ABB’s propensity to participate in cartels, the fact is that ABB’s participation in the Cartel is admitted.

(2)         I do not consider that ABB’s propensity to participate in cartels can say anything about the nature or extent of the economic benefit that ABB derived from this. I accept that, as an organisation, ABB would not have participated in cartels generally unless it saw some benefit from this. But the same point can be made in relation to ABB’s participation in the Cartel itself. I accept that a rational organisation, and rational people, are not going to engage in illegal cartelist behaviour unless they are satisfied that it benefits them. The fact that ABB participated in other cartels adds nothing.

(3)         Obviously, participation in cartels is unlawful, and the evidence of employees of ABB who knew of the Cartel and who therefore participated in unlawful conduct will have to be treated with caution. I obviously will consider this point when assessing the weight to be attached to the evidence of the witnesses who were called before me. But, again, the fact that ABB – or even these witnesses – participated in other cartels does not assist me in the matters I must address.

C.      THE EVIDENCE

(1)     Introduction

26.         The evidence before me comprised the following types:

(1)         Factual witness evidence (i.e. by witnesses of fact called to give evidence before me).

(2)         Documentary evidence.

(3)         The Decision.

(4)         Expert evidence.

27.         Each type of evidence gave rise to particular difficulties in terms of evaluation and the drawing of conclusions. Sections C(1) to C(5) describe and assess these various different types of evidence. Section C(6) provides some overview.

(2)     Factual witness evidence

(a)     General points

28.         The general problems presented by witnesses of fact are well-known and have been clearly articulated by Leggatt J in Gestmin SGPS S.A. v. Credit Suisse (UK) Limited , [2013] EWHC 3560 (Comm) at [15] to [22]. Even if I had heard evidence from a number of persons regarding the operation of the Cartel, reconstructing its operation with particular regard to the BritNed Interconnector would have been extremely difficult.

29.         As it was, I only heard evidence from one person with actual knowledge of the Cartel. BritNed’s factual witnesses, unsurprisingly, had no contemporaneous knowledge of the Cartel. Of ABB’s five factual witnesses, only one knew of the Cartel and ABB’s participation in it. The other four, it was accepted by BritNed and I find, had no such knowledge. If and to the extent that their conduct caused BritNed to suffer an overcharge, this was unconscious on their part.

30.         Thus, not only did the relevant events take place a number of years ago, but also the witnesses whose evidence I received could, with one exception, at best and at most, describe their innocent actions in the context of a cartelised bid. Inevitably, I have had to tread extremely carefully in assessing such evidence, and I say that out of no disrespect to these witnesses.

31.         The evidence of the exceptional witness – Mr Hans-Åke Jönsson – who did know of the Cartel presented different and even greater problems, which I consider in greater detail below. In substance, however, those problems are twofold:

(1)         First, Mr Jönsson’s understanding of the Cartel was itself limited.

(2)         Secondly, Mr Jönsson was, of course, testifying as to his own misconduct and I consider that this inevitably had a distortive effect on his evidence.

(b)     BritNed’s factual witnesses

32.         BritNed called two factual witnesses, Mr Mathew Rose and Mr Michael Jackson.

(i)      Mr Mathew Rose

33.         At the times material to these proceedings, Mr Rose worked for National Grid plc. He worked for National Grid plc from 1999 to 2013, in a variety of roles. Between July 2005 and October 2007, Mr Rose was the managing director of BritNed, where he was responsible for running the BritNed project alongside his colleague, Mr Dick Bos. Mr Rose and Mr Bos were joint Project Managers, and they led the BritNed project team. That does not mean to say that Mr Rose was always, or even mostly, involved in the front-line negotiations regarding the BritNed project – often such negotiations or dealings would be handled by other members of the team. But I accept that Mr Rose was centrally involved in the BritNed project between July 2005 and October 2007.

34.         Mr Rose made three written statements in these proceedings:

(1)         A first statement, dated 8 February 2017 (“Rose 1”).

(2)         A second statement, dated 28 April 2017 (“Rose 2”).

(3)         A third statement, dated 7 November 2017 (“Rose 3”).

Mr Rose gave evidence on Day 4 (12 February 2018).

35.         Mr Rose was an impressive witness. Highly articulate, he radiated a quiet confidence in his abilities, and had a good recollection of the events he had been involved in and the way the BritNed project had developed. He was, during the time of his involvement in the BritNed project, unaware of the Cartel. His subsequent knowledge of the existence of the Cartel inevitably coloured how he saw past events: unsurprisingly, he saw them in a different light. For instance, the reason for the non-participation of certain parties in the BritNed tender, which might have puzzled him at the time, became clearer over time. I regard it as inevitable that Mr Rose’s recollection would be coloured by events he learned of subsequently. I consider he did his best to separate what would have been his thinking in 2005-2007 and what his interpretation of those events was now, in light of his appreciation of the existence of the Cartel. He was an honest, straightforward and, as I have said, impressive witness.

(ii)     Mr Michael Jackson

36.         Mr Michael Jackson provides (through his company, Mike F Jackson Consulting Limited) consultancy services to a range of clients in the power sector. Between August 2006 and August 2008, he was engaged, as an independent contractor, as commercial manager for the BritNed project. His primary role was to lead BritNed’s negotiation team.

37.         Mr Jackson made one witness statement in these proceedings, dated 28 April 2017 (“Jackson 1”). He gave evidence on Day 4 (12 February 2018).

38.         Mr Jackson was a bluff and forthright witness. He gave his evidence honestly, but I consider he had greater difficulty than Mr Rose in separating his subsequent knowledge about the Cartel from his attempt to recollect his thinking at the time. As I have already stressed in the case of Mr Rose, I regard this colouring of recollection as inevitable, and this is in no sense a criticism of Mr Jackson. Nevertheless, it is important to bear this factor in mind: it certainly coloured Mr Jackson’s evidence regarding the “discount” provided by ABB during the course of the final stages of the negotiation of the contract between BritNed and ABB. It will be necessary to return to this discount later on in this Judgment: it constitutes one of the clearest indicators that ABB was under some competitive pressure, which of course sits uneasily with the Cartel effects contended for by BritNed. Mr Jackson’s evidence on this point was a little skewed by the fact that he could not reconcile the granting of this discount by ABB with his after-the-event views of the effect of the Cartel.

(c)     ABB’s factual witnesses

39.         ABB called five witnesses in the following order:

(1)         Mr Hans-Åke Jönsson.

(2)         Mr Peter Leupp.

(3)         Mr Stefan Ekman.

(4)         Mr Magnus Larsson-Hoffstein.

(5)         Mr Hans Magnus Röstlund.

(i)      Mr Hans-Åke Jönsson

40.         Mr Jönsson is now retired. He worked for ABB between 1982 and 2009. For a short period after 2009, he worked as a consultant. He knew of the Cartel. His departure from ABB in 2009 was as a consequence of his involvement in the Cartel. The terms of his departure from ABB were set out in an agreement (that remains confidential, but which I have seen) made in 2010.

41.         From 1994 to 2000, Mr Jönsson was the general manager of the ABB Reactive Power Compensation Business in Västerås, Sweden. From 2001 to 2009, he was the general manager and vice-president of the ABB High Voltage Cables business. During the period – specifically, between 2004 and 2008 – he was additionally the manager of ABB’s cables factory in Karlskrona.

42.         Mr Jönsson made four witness statements in these proceedings:

(1)         A first statement, dated 5 February 2017 (“Jönsson 1”).

(2)         A second statement, dated 28 April 2017 (“Jönsson 2”).

(3)         A third statement, dated 30 May 2017 (“Jönsson 3”).

(4)         A fourth statement, dated 9 November 2017 (“Jönsson 4”).

43.         Mr Jönsson gave evidence on Day 5 (13 February 2018), Day 6 (14 February 2018) and Day 7 (15 February 2018).

44.         Mr Jönsson was a self-confessed cartelist. I consider that he had an obvious interest in minimising his – and therefore ABB’s – involvement in the Cartel. I do not consider that Mr Jönsson was a dishonest witness, but I do consider that he did, throughout his evidence, seek to minimise the effect of the Cartel and ABB’s (and his) role in it, so far as he possibly could. Mr Jönsson was an extremely clever man, and his approach was to accept, pretty much without question, unequivocal findings in the Decision. However, the moment there was wriggle-room – for instance, where a more benevolent interpretation of the facts could be taken or where there was an alternative (non-Cartel-related) explanation for events – Mr Jönsson showed a distinct pre-disposition to opt for this, more advantageous, course. That does not mean to say that Mr Jönsson was wrong in what he said; still less that he was being dishonest. But the trend was very much one-way, and that is a factor I consider that I must bear in mind when evaluating Mr Jönsson’s evidence.

45.         There was, in short, a lack of frankness in Mr Jönsson’s evidence as to how, in practical terms, the Cartel operated, specifically in relation to the BritNed Interconnector.

46.         Accordingly, I am minded to treat Mr Jönsson’s evidence with a relatively high degree of caution. That is particularly the case because – as is frequently the situation with cartels – the documentation regarding the detailed operation of the Cartel was probably always quite sparse, and most of such documentation as did exist has either failed to survive or else is kept under wraps by the European Commission itself (and is not available to me).

(ii)     Mr Peter Leupp

47.         At the time of the BritNed project, Mr Leupp was the head of division, Power Systems, within ABB Limited, based in Switzerland. ABB’s cables business was, at this time, within the Power Systems Division.

48.         Mr Leupp was involved in the process by which ABB sought to tender for the BritNed project. He claimed not to be aware of the existence of the Cartel, and BritNed did not seek to contend otherwise. I accept this evidence.

49.         Mr Leupp was in a high-level position within ABB. [29] He would have relied on others to brief him on the details of the BritNed project, including the competitive dynamics and the state of negotiations. Those briefing Mr Leupp would have included Mr Jönsson, although Mr Jönsson did not report directly to Mr Leupp. Mr Jönsson reported to a Mr Per Haugland (who did not give evidence before me), who reported to Mr Leupp. [30]

50.         Mr Leupp made one witness statement dated 27 September 2016 (“Leupp 1”). He gave evidence on Day 7 (15 February 2018).

51.         Mr Leupp was a precise, clear and articulate witness. I regard his evidence as reliable.

(iii)    Mr Stefan Ekman

52.         Mr Ekman is now the senior advisor (finance) at NKT HV Cables AB. Prior to its sale to NKT in March 2017, Mr Ekman was the manager business controlling for ABB’s High Voltage Cables business from 2010. Prior to that, he was the chief financial officer in the same business, which was located in Karlskrona, Sweden.

53.         Mr Ekman had not been involved in the Cartel, [31] as BritNed accepted. He also played no role in the negotiation of the BritNed contract. His evidence was submitted in response to certain points made by Dr Jenkins – BritNed’s expert economist – regarding ABB’s order backlog. The significance of this point – and the significance of Mr Ekman’s evidence – is considered further below.

54.         Mr Ekman made one witness statement dated 9 November 2017 (“Ekman 1”). He gave evidence on Day 8 (16 February 2018).

55.         Given the nature of his factual evidence – which essentially drew him into a dispute regarding the modelling of the effect of the Cartel and the extent of the overcharge between the expert economists – Mr Ekman was (I find quite understandably) somewhat of a defensive and cautious witness, determined to speak only to matters within his knowledge. Sometimes that caution was a little excessive: there were times when Mr Ekman declined to be drawn even on documents referenced in his own statement. But the fact is that because of the somewhat recondite area on which he was giving evidence, his defensiveness and caution were perfectly understandable, and I consider that he was doing his very best to assist the court and that his evidence is reliable.

(iv)    Mr Magnus Larsson-Hoffstein

56.         Mr Larsson-Hoffstein is a project manager at NKT HV Cables (Sweden) AB, a position he has held since 1 February 2017. Prior to that date, between late 2007 and 31 January 2017, he was employed in the same position by ABB AB. He originally joined ABB AB in 2000.

57.         Mr Larsson-Hoffstein was centrally involved in the tender process for the BritNed project. He had a team of people under him, working on the tender, and there were several echelons of people above him in the organisation (including Mr Jönsson) to whom he reported (directly or indirectly) and who also (directly or indirectly) had input into the tender process.

58.         Mr Larsson-Hoffstein was not involved in nor aware of the Cartel, [32] and this evidence was not challenged by BritNed. [33] At the trial, Mr Larsson-Hoffstein made a single witness statement dated 6 February 2017 (“Larsson-Hoffstein 1”). He gave evidence on Day 8 (16 February 2018).

59.         Mr Larsson-Hoffstein’s English was good, but not as good as the other witnesses called by ABB. Nevertheless, I am satisfied that he understood all of the questions put to him, and his answers were clear. He was a transparently honest witness, who provided me with clear insight into the way in which ABB put together tenders and conducted its negotiations.

60.         Subsequent to the trial, and in order to deal with a discrete point regarding the cost of copper purchased by ABB for use in the cable for the BritNed Interconnector, Mr Larsson-Hoffstein made a further witness statement dated 29 March 2018 (“Larsson-Hoffstein 2”).

(v)     Mr Hans Magnus Röstlund

61.         Mr Röstlund began working for ABB in 2000. Throughout his career he was involved in sales. He was not involved in, nor aware of, the Cartel (a fact which BritNed did not challenge). [34] He commented on power cables projects from a technical perspective. He was not directly involved in the BritNed project but provided evidence regarding power cables projects more generally.

62.         Mr Röstlund made two witness statements in the proceedings, the first dated 12 July 2016 (“Röstlund 1”) and the second dated 6 February 2017 (“Röstlund 2”). He gave evidence on Day 9 (19 February 2018).

63.         Mr Röstlund was a straightforward and extremely competent witness. He knew what he was talking about and was careful to make clear the limits of his knowledge (in particular on specific technical aspects). He was an impressive witness.

(3)     The documentary evidence

64.         In the ordinary course, when assessing factual evidence, a Judge has well in mind the approach of Lord Goff in Grace Shipping Inc. v. CF Sharp and Co (Malaya) Pte Ltd 1987] 1 Lloyd's Rep 207 at 215:

“In such a case [where witnesses were seeking to recall events and telephone conversations of five years earlier] memories may very well be unreliable; and it is of crucial importance for the judge to have regard to the contemporary documents and to the overall probabilities…”

65.         Whilst I obviously have had regard to the contemporary documents, it is (as I have said) in the nature of cartels that material documents are sparse. Either notes or records are not kept or else they are destroyed. In the case of the Cartel, and ABB’s recording of its activities, the practice appears to have been for ABB to keep as few records as possible. Mr Jönsson’s evidence was as follows: [35]

Q (Mr O’Donoghue, QC)

…secrecy was a major feature of this cartel, wasn’t it?

A (Mr Jönsson )

Yes.

Q (Mr O’Donoghue, QC)

…In your activities, which phone did you use for Cartel discussions?

A (Mr Jönsson )

My phone, my office phone.

Q (Mr O’Donoghue, QC)

Insofar as you sent emails, which email did you use for Cartel activities?

A (Mr Jönsson )

I did very – there is maybe one or whatever – I never sent emails related to Cartel activities. There is – I can see one that I know of, but otherwise I don’t send any emails.

Q (Mr O’Donoghue, QC)

So you were very careful not to send emails, were you?

A (Mr Jönsson )

Yes.

Q (Mr O’Donoghue, QC)

To avoid detection?

A (Mr Jönsson )

Yes.

 

Q (Mr O’Donoghue, QC)

Did you create, at any stage, documents which were subsequently destroyed?

A (Mr Jönsson )

No. I can say the following: when I was in meetings, sometimes I had – there are the handwritten notes which have been connected to several of the meetings which we did back in 2009, but then otherwise I would be remembering, I was keeping it in my head. There has not been any document destroyed

Q (Mr O’Donoghue, QC)

So you took steps to avoid creating documents at all stages?

A (Mr Jönsson )

Correct.

Q (Mr O’Donoghue, QC)

To avoid detection?

A (Mr Jönsson )

Correct.

Meetings between cartelists took place under the cover of legitimate excuses to meet. One of the practices within the Cartel was to use the international cable-makers conferences as a cover to discuss the Cartel. [36]

66.         There are thus significant gaps in the contemporary documentary record. In the first place, there is before the court only ABB’s disclosure; [37] and that, as Mr Jönsson made clear, was an intentionally incomplete record.

(4)     The Decision of the European Commission

(a)     The law regarding the bindingness of decisions

67.         The relevant law may be stated in the following propositions:

(1)         The Decision is a decision of the European Commission. [38] As such, it is binding in its entirety upon those to whom it is addressed. [39] ABB is an addressee of the Decision. But neither BritNed nor (to state the obvious) this court is an addressee.

(2)         It is well-known that decisions of the European Commission comprise an “operative” part coming after a series of recitals. [40] The Decision is no exception. Commencing with the basis upon which the Commission’s jurisdiction is founded, under the word “Whereas”, there follow – numbered (1) to (1078) – a series of recitals, before (on page 185) the words “HAS ADOPTED THIS DECISION” appear. There then follows, comprising a single Article, the operative part of the Decision. This provides:

Article 1

The following undertakings infringed Article 101 of the Treaty and Article 53 of the EEA Agreement by participating, in a single and continuous infringement in the (extra) high voltage underground and/or submarine power cables sector:”

The undertakings are then listed, and they include ABB. The operative part continues:

“This Decision shall be enforceable pursuant to Article 299 of the Treaty and Article 110 of the EEA Agreement.

Done at Brussels…”

(3)         In the provisional non-confidential version of the Decision published in November 2017, the operative part did no more than identify the participating undertakings. There was no statement of the periods of their participation, nor of the fines imposed. A final non-confidential version was published on 4 July 2018, where this information was provided.

(4)         The distinction between recitals and the operative part is important and well-recognised in European Union law. In Case C-164/02, Kingdom of the Netherlands v. Commission of the European Communities EU:C:2004:54, [2004] ECR I-1179, the European Court of Justice held: [41]

“…it need merely be stated that, regardless of the grounds on which such a decision is based, only the operative part thereof is capable of producing legal effects…By contrast, the assessments made in the recitals to a decision are not in themselves capable of forming the subject of an application for annulment. They can be subject to judicial review by the Community judicature only to the extent that, as grounds of an act adversely affecting a person’s interests, they constitute the essential basis for the operative part of that act.”

(5)         The following points fall to be made:

(a)            There is a central ambiguity in the term “decision”:

(i)            It can refer to the instrument by which a decision is or decisions are made. On this reading, the Decision is a decision. This is the “wide” meaning of the term “decision”.

(ii)           Alternatively, it may refer simply to the operative part of the decision (in the wide sense). This, “narrow”, understanding of a decision looks only at what is capable of producing legal effects.

(b)           It is quite clear from the European Union jurisprudence cited above that what matters is “decision” in the “narrow” sense understood by the European Court of Justice in Case C-164/02, Kingdom of the Netherlands v. EC Commission and that is how (from hereon) I shall use the term in this Judgment. I shall refer to the instrument containing a decision as just that – the “instrument”.

(6)         It follows that an instrument may contain three different types of provision:

(a)            A decision. A decision in an instrument, as has been described, [42] is binding on its addressees. A decision is also – to the extent it constitutes a final infringement decision within the meaning of section 58A of the Competition Act 1998 – binding on this court. [43]

(b)           A recital constituting part of the essential basis for a decision. As Case C-164/02, Kingdom of the Netherlands v. EC Commission makes clear, whilst generally speaking recitals are not acts capable of review by the courts, an exception is made in the case of those recitals constituting the essential basis for the operative part of that act. Here, the relevant act is a decision and to the extent that a recital constitutes part of the essential basis for a decision then – where the decision is binding on this court – so too is such a recital. This, as it seems to me, is an inevitable consequence of section 58A of the Competition Act 1998 and Article 16(1) of Regulation 1/2003. [44] Clearly, what constitutes a recital constituting a part of the essential basis for a decision depends largely on the nature of the decision itself.

(c)            A recital not constituting part of the essential basis for a decision. Such recitals are not binding on this court. I do not consider that such a conclusion is inconsistent with the duty of sincere cooperation arising out of Article 4(3) of the Treaty on European Union, and I can identify no other rule – whether of European Union law or English law – that compels such a conclusion. In Crehan v. Inntrepreneur Pub Co. (CPC) [2006] UKHL 38, Lord Hoffmann said this at [69]:

“The correct position is that, when there is no question of a conflict of decisions in the sense which I have discussed, the decision of the Commission is simply evidence properly admissible before the English court which, given the expertise of the Commission, may well be regarded by the court as highly persuasive. As a matter of law, however, it is only part of the evidence which the court will take into account. If, upon an assessment of all the evidence, the judge comes to the conclusion that the view of the Commission was wrong, I do not see how, consistently with his judicial oath, he can say that as a matter of deference he proposes nevertheless to follow the Commission. Only a rule of law, in the nature of an issue estoppel which obliges him to do so, could produce such a result…”

(b)     My approach in the case of the Decision

68.         Applying this approach to the Decision is relatively straightforward. ABB did not seek to challenge the operative parts of the Decision nor, indeed, the recitals in the Decision that might be said to constitute part of the essential basis for a decision. [45] Indeed, Mr Jönsson agreed with the general description of the Cartel as set out in recitals (528) to (535) of the Decision. [46]

69.         The problem is that most of the Decision does not deal with the BritNed Interconnector, although some reference is made to it. The Decision – entirely unsurprisingly – operates at an altogether higher level of generality, since it is concerned with the Cartel generally and not with any single project that was the object of the Cartel. In this case, of course, I am concerned only with the BritNed Interconnector: obviously, the findings I make will be made against the backdrop of the fact that the Interconnector was sold in a cartelised market, but it is difficult to see what specific findings this impels me to.

70.         There are two other factors that are relevant to the weight that I attach to the Decision:

(1)         Large parts of the Decision were redacted. Redactions were in two forms:

(a)            Some redactions simply highlighted confidential passages. Thus, I was able to read the passage in question, but was alerted to its confidential nature.

(b)           Some redactions – and there were many of these – replaced the passage with “[…]”, so that it was impossible to read or understand the redacted passage.

(2)         A number of the documents on which the Commission relied were not available to me. It is obviously difficult to say without actually seeing them, but some appeared to me to be potentially quite important.

71.         In these circumstances, I am inclined to be cautious in terms of the weight I attach to non-binding recitals in the Decision. The statements by the European Commission may well be justifiable: but on these points I am the determiner of fact, and I am only prepared to accept a non-binding statement by the European Commission where it seems to me that it is a finding I can properly make on the evidence viewed as a whole.

(5)     The expert evidence

72.         I heard evidence – over six days – from two expert economists. BritNed called Dr Helen Jenkins of Oxera; ABB called Mr Zoltan Biro of Frontier Economics.

73.         The experts submitted their reports in these proceedings in accordance with my order of 20 July 2017. That order required the parties to produce a list of issues to be addressed by the experts in an agreed form. That list was then supplemented by a statement by each expert, setting out in broad terms the methodological approach each expert proposed to take in relation to each issue. The list also identified on which issue or issues each expert was to take the lead. Save on the issue of overcharge, where the approaches of the two experts were so different that each expert took the lead in articulating his/her approach, this meant that the issues between the experts were addressed sequentially.

74.         In all, the following reports were submitted by the experts:

(1)         A report of Dr Jenkins dated 15 September 2017 (“Jenkins 1”).

(2)         A report of Mr Biro dated 15 September 2017 (“Biro 1”).

(3)         A report of Dr Jenkins dated 10 November 2017 (“Jenkins 2”).

(4)         A report of Mr Biro dated 10 November 2017 (“Biro 2”).

(5)         A joint statement of Dr Jenkins and Mr Biro dated 12 December 2017 (the “Joint Statement”).

Additionally, in response to two questions I addressed to the experts at the end of July 2018, the experts produced a joint response (the “24 August 2018 Response”), which helpfully set out their answers to these questions.

75.         The experts gave evidence in two stages. Before any factual evidence was called, I indicated that it would be helpful if each expert could provide a neutral explanation, under oath, of their working methodology. This came to be referred to as the “teach-in” and it took place on Day 3 (9 February 2018), after the parties’ oral openings, which took place on Days 1 and 2 (7 and 8 February 2018). The experts were then cross-examined, for just over two days each, on Days 9 to 13 (19 to 23 February 2018). Both experts were extremely impressive witnesses. I consider that they did their very best to assist the court in what was, on any view, a complex and difficult exercise.

(6)     Overview

76.         Each of the different types of evidence before me has its strengths and weaknesses. The factual evidence, the documentary evidence and the Decision inform the substance of Sections D, E F and G below. The expert evidence, although it features incidentally in these Sections, is on the whole treated separately, in Section H. The various strands are then brought together, and the various claims and defences determined, in Section I and the subsequent Sections of the Judgment.

77.         In considering the various strands of evidence before me, I have been concerned to build a picture of the relationship between BritNed and ABB and the manner and extent to which that relationship was affected by the Cartel. Without building as comprehensive a picture as possible, it will be difficult to discern the true nature and true effects of the Cartel. 

78.         In opening, BritNed said that “[t]he narrative side of the case, in terms of the contours and fabric of the Cartel, are in my submission critically important when it comes to understanding the factual evidence.” [47] I agree with this statement, and consider the general nature and operation of the Cartel in Section D below. But that is not the only factual aspect that is of critical importance. Of equal importance are the characteristics of submarine cable projects (Section E), the history of the BritNed tender (Section F), and the competitive pressures that arose or – because of the Cartel – did not arise during the negotiating process (Section G). As I have noted, the expert assessment stands somewhat separate but also draws on this material. For that reason, I consider it separately (Section H).

79.         The material before me inevitably has gaps. These I seek to bridge through careful deployment of the broad brush. I have not considered it necessary or appropriate to draw adverse inferences in relation to the absence of material before me, although I was invited to do so by BritNed. [48] It seems to me that, at least in this case, it would be a mistake to do so. I am not concerned, as I have noted, in determining the existence of the Cartel nor in deciding ABB’s participation in it. These are givens. I am concerned with the economic effects of the Cartel, translated into a claim for damages. To draw adverse inferences would be a distortion of the process of assessment that I must undertake, in much the same way as a presumption of overcharge for the reasons given above. [49]

D.      THE NATURE AND OPERATION OF THE CARTEL

80.         I shall consider the nature and operation of the Cartel first in general terms (Section D(1)), before turning to ABB’s specific role and involvement in the Cartel (Section D(2)).

(1)     General nature and operation

81.         The Cartel had its genesis in a perceived excess of capacity amongst cable suppliers. [50] Although there were calls to reduce existing excess capacity, [51] the Cartel sought to deal with this problem by maintaining price levels and allocating bids. [52]

82.         Essentially, the Cartel operated on a territorial basis, using a “home territory” principle. [53] Thus, Japanese and Korean producers would not compete for power cable projects in the European home territory and Europeans would not compete for power cable projects in the Japanese and Korean home territories. [54]

83.         Within these territories, there was further territorial allocation. The Baltic and North Sea area was allocated to ABB and (to some extent) Nexans. The Mediterranean area was divided between Prysmian and Nexans. [55] But there were ad hoc exceptions to this territorial approach [56] and friction was generated when multiple parties sought the same contract. [57]

84.         The Decision says this about the Cartel:

“(67) Adhering to the rules of the Cartel, from February 1999 onwards, the parties allocated projects according to their geographic region or customer. In addition, they exchanged information on prices and other commercially sensitive information in order to ensure that the designated power cable supplier or “allottee” would make the lowest price while the other companies would submit a higher offer or refrain from bidding or submit an offer that was unattractive to the customer. The parties installed reporting obligations to allow monitoring of the agreed allocations. Finally, the parties also implemented practices to reinforce the [Cartel] such as the collective refusal to supply accessories or technical assistance to certain competitors in order to ensure the agreed allocations.”

(68)   To ensure the implementation of the Cartel arrangements, the parties held periodical meetings and had contacts by email, telephone or fax.

(69)   Within the Cartel, and in line with [information pre-dating the infringement period], the European producers Nexans and Pirelli/Prysmian were normally referred to as “R” (“Regular”) members, the Japanese producers Sumitomo, Hitachi and JPS, Furukawa and VISCAS (and later also EXSYM) as “A” (“Associated”) members and the Korean companies LS Cable and Taihan as “K”. In addition, the parties used the term “R associates” for ABB, Sagem/Safran, Brugg and nkt, while “A associates” was also used to refer to LS Cable, Taihan and Mitsubishi, Showa and EXSYM during a certain period.

(70)   Most of the parties participated in two main types of meetings:

(a)     the so-called “A/R meetings”, between representatives of the European and Japanese producers, and

(b)     regional meetings, such as the recurrent “R meetings” (also referred to as “seminars”) in which only the local producers participated.

(71)   In addition to the A/R meetings and R meetings, meetings including the Korean companies (“A/K/R” meetings), bilateral and multilateral meetings between selected parties and meetings at the occasion of industry conferences (such as the International Cablemakers Federation (“ICF”) sessions) were also frequent.

(72)   Given the long period over which the Cartel has been operating, certain aspects and details of the Cartel arrangements, such as the geographic areas considered as “home territories”, the voltage levels covered by the arrangements or how projects were allocated within certain territories, have evolved over time. However, the evidence gathered by the Commission shows that the main features of the Cartel arrangements described below have in essence been maintained over time.

(73)   The Cartel had two main configurations:

(a)     On the one hand, the European, Japanese and Korean producers had as their objective the allocation of territories and customers. This configuration is referred to as the “A/R Cartel configuration”…Pursuant to this configuration, Japanese and Korean producers refrained from competing for projects in the European home territory while the European producers would stay out of Japan and Korea. The parties also allocated projects in most of the rest of the world and made use of a 60/40 quota arrangement for a certain period of time.

(b)     On the other hand, the “European Cartel configuration” involved the allocation of territories and customers by the European producers for projects inside the European home territory or allocated to the European producers…

(74)   These configurations were not separate but formed a composite whole.”

85.         The Cartel had, within its allocations, “compensation” mechanisms to ensure “fairness”. Thus, if one member of the Cartel forwent a particular opportunity to bid (either by not bidding at all or by putting in an uncompetitive bid), that member would in due course receive “compensation” (generally in the form of being the favoured bidder in another project). [58] For example:

(1)         The Eirgrid SM power cable project in Ireland was allocated to ABB in exchange for Nexans obtaining the Fennoskan II project. [59]

(2)         The NorNed project was allocated to ABB in exchange for foregoing the North Sea Interconnector project. [60]

86.         This, of course, involved keeping track of allocations and monitoring who got what. [61]

87.         In order to allocate projects to particular cartelists, it was, of course, necessary to exchange information regarding bids, so that the cartelists who were not to succeed could (if they were going to bid) ensure that their bids were appropriately unattractive. [62]

88.         The Cartel involved a great many meetings, although not necessarily all of the cartelists attended all of the meetings. The Decision says this: [63]

“In order to achieve their overall aim, the parties established a network of multilateral and bilateral meetings and contacts and participated in one or more of the following cartel activities:

(a)     All producers implicitly or explicitly entered into an agreement or concerted practice through which the European home territory was protected from competition by Japanese and Korean power cable suppliers and vice versa

(b)     In addition, the European Cartel members participated in the European cartel configuration; an agreement or concerted practice through which they allocated territories and customers within the EEA…

(c)     All producers participated in the allocation of projects in the export territories…

(d)     Several parties agreed on the prices to be offered for [submarine] and [underground] power cable projects by either the establishment of a floor price or the coordination of price levels…

(e)     Several parties participated in the submission of cover bids in order to ensure the agreed allocation of [submarine] and [underground] power cable projects. To this end, the parties exchanged prices and other sensitive commercial terms and conditions, required for the preparation of the cover bids. These agreements concerned…projects in the EEA…

(f)      Several parties participated in the exchange of other sensitive commercial and strategic information such as their available capacity or interest in participating in specific tenders. These agreements concerned both projects in the EEA as well as in the export territories…

(g)     Some parties participated in the implementation of practices to reinforce the Cartel such as the collective refusal to supply accessories or technical assistance to certain competitors…

(h)     Several parties were involved in the monitoring of the implementation of the allocation and price agreements through the exchange of position sheets, market information and the establishment what reporting obligations. These arrangements concerned both projects in the EEA as in the export territories…”

89.         Although the Cartel had clear objectives, there were internal conflicts, rivalries and cheating (in the sense that a cartelist bid competitively for a project not allocated to it by the Cartel). [64]

(2)     ABB and the Cartel

90.         ABB was not in the Cartel from the beginning. According to the Decision, ABB started participating in the Cartel between April and June 2000. Other members of the Cartel considered ABB’s participation important. [65]

91.         In his witness statement, Mr Jönsson emphasised Recital (453) of the Decision, which noted that “ABB’s position as a non-core player has prevented it from obtaining detailed information on the general application of the [Cartel]”. [66] Recital (453) refers to the information that the various leniency applicants were able to provide to the Commission. But the Decision also notes at Recital (563): [67]

“…Because of its absence from the A/R meetings, ABB was not able to set out the parameters of the Cartel. The level of participation of ABB is therefore lower than that of the core players. However, its deep involvement in many of the Cartel activities as set out in Recital (493) [68] and its participation in many contacts and meetings do not qualify ABB as a fringe player .”

92.         The Decision finds that ABB was aware of the allocation of projects within Europe: [69]

“ABB’s awareness of the allocation of projects within Europe is also evidenced in an internal email of 10 April 2000…This email contains the phrase: “I suspect that when [non-addressee] let Viking go to Pirelli and NorNed to us, the NSI [North Sea Interconnector, linking Norway and England] became their compensation”.

93.         Mr Jönsson was told about the Cartel in January 2001, when he became business unit manager for cables in the high voltage cable business of ABB. [70] The Decision says this: [71]

“In March 2001, Mr Jönsson (ABB) was introduced by his predecessor, Mr Carlstedt, to Mr Romand (Nexans) and [company representative B1] (Pirelli). During this meeting, held in a hotel in Zurich, Mr Jönsson was made aware of the cartel arrangements and of the ways in which the illicit cooperation was being carried out. ABB has declared that it was clear for Mr Jönsson that his role was to continue the cooperation between the companies that had taken place prior to his assignment to the cable business…” [72]

94.         Mr Jönsson attended a further meeting in July 2001, in Zurich again, to discuss allocation of certain projects. [73]

95.         Entirely unsurprisingly, BritNed sought to play up ABB’s role and involvement in the Cartel, whilst ABB sought to play it down. It is unnecessary for me to reach many findings as regards ABB’s role and involvement in the Cartel in general terms, as opposed to the effect of the Cartel on the tender process and price of the BritNed Interconnector. It is sufficient for me to find, as I do, ABB was (as an organisation) appreciative of the general nature and operation of the Cartel, as I have described it in Section D(1) above.

96.         It is unnecessary for me to consider the extent to which ABB was involved in the highest level of Cartel meetings, nor who exactly within ABB was aware of the Cartel. The general operation of the Cartel is only relevant insofar as it sheds light on the manner in which the Cartel affected the BritNed tender. That is a matter that I turn to in Section G below.

E.      THE CHARACTERISTICS OF SUBMARINE CABLES

(1)     Introduction

97.         There are a number of technical points that need to be understood about the nature of submarine cable projects. These are as follow:

(1)         The differences between submarine and underground cable projects.

(2)         Alternating current or direct current.

(3)         Cable width.

(2)     The difference between submarine and underground cable projects

98.         The very significant differences between underground and submarine cable projects were described in detail by Mr Röstlund in his statement. [74] His evidence was not – for the most part [75] – challenged in cross-examination.

99.         The differences between underground and submarine cable projects are important because – as will be described – Dr Jenkins used both underground and submarine cable projects for the purposes of her overcharge analysis and modelling. Dr Jenkins recognised that there were significant differences between underground and submarine cable projects and sought to compensate for this in her analysis. I consider the extent to which she was successful in this regard below. [76]

100.       It is, therefore, necessary to understand the differences that Dr Jenkins was seeking to represent in her model. Mr Röstlund identified a number of differences between underground and submarine cable projects:

(1)         The cable structure is different according to whether the cable is for underground or submarine use. Mr Röstlund noted that “the design and structure of submarine and underground cables differ substantially and the resulting impact on the mix and volumes of materials used in their manufacturing process leads to substantial differences in their production costs, which are reflected in the prices charged to customers.” [77]

(2)         Submarine cables are manufactured on a bespoke basis, whereas underground cables tend to be bought “off the shelf”. Mr Röstlund stated that “the variety of challenges posed by the submarine environment and the range of different project types means that submarine cables are designed and manufactured on a bespoke basis, with the cables tailored specifically to the requirements of each individual project. In contrast, underground cables are a more commoditised product supplied on a relatively standardised basis, manufactured to pre-defined type-tested designs. This difference means that the prices, costs and margins associated with underground projects are very different to those relating to submarine projects.” [78]

(3)         Submarine cables are more complex to manufacture than underground cables. Mr Röstlund’s evidence on this point was that “submarine cables are produced in a very long sections of tens of kilometres, as it is important that submarine cables have as few joints as possible. In contrast, underground cables are typically produced in sections of a few hundred metres. The production of a very long (tens of kilometres) single length of submarine cable is fundamentally different to the production of short (few hundred metres) sections of underground cable. The production of a single length of submarine cable requires specialised manufacturing capabilities, which substantially increases the costs of manufacturing submarine cables and limits the range of suppliers who are capable of competing for this business.” [79]

(4)         The installation requirements for submarine cables are different to the requirements in relation to underground cables. Mr Röstlund said that “the installation of submarine cables is fundamentally different from that of underground cables, requiring entirely different equipment (e.g. laying vessels) and capabilities. The complexity of the submarine environment creates particular risks and challenges. Submarine cable installation is necessarily highly bespoke, with requirements depending on the depth, seabed and weather conditions, as well as on the particular customer specification and permits which impose regulatory installation conditions. In contrast, underground cable installation used more standard and less costly equipment (i.e. cable laying vessels are obviously not required, but normally just standard trucks, mechanical diggers, etc.) and is much more straightforward than submarine cable installation. These differences in the nature of installation are similarly reflected in differences in the prices, costs and margins associated with submarine cables projects, compared to those associated with underground cables projects.” [80]

(5)         The supply chain is different in relation to underground and submarine cable projects. Mr Röstlund’s evidence was that “submarine cables projects at higher voltage levels are almost always supplied as so-called “turnkey” projects, in which the cable manufacturer takes responsibility for the end-to-end delivery of the whole project, including all design, manufacturing, installation activities and accessories. This is not the case for underground cables projects. While in certain cases underground cables may be supplied on a turnkey basis, manufacturers will frequently supply only the cables (sometimes referred to as “naked” cable sales) direct to the customer or to an EPC (engineering, procurement and construction) contractor who takes responsibility for the overall project delivery. This distinction in the nature of the supply chain means that the costs and risks faced by the manufacturer when supplying submarine cables are fundamentally different to those faced when supplying underground cables.” [81]

As Mr Röstlund explained in relation to each of the five factors set out above, each has a significant effect on the price that will be charged in relation to submarine cable projects in contradistinction to underground cable projects, with the former generally being materially higher in price than the latter.

Mr Röstlund did identify a sixth factor, which I set out below, but which I propose to treat much more cautiously when considering the differences between underground and submarine cable projects. Mr Röstlund’s sixth factor was as follows:

(6)         The competitive environment is different in relation to underground cable projects than for submarine cable projects. Mr Röstlund said this: [82]

“…the number of rival firms that can manufacture and install submarine cables projects is limited, due to the expertise and capabilities required. As underground cables are much more straightforward to manufacture, the number of firms actively competing to supply underground cables is significantly larger. This is because more firms have the capacity to produce underground cables than to produce and deliver submarine cables projects. This significant difference in the competitive environments relating to the manufacture and supply of submarine and underground cables is reflected in the differences in their respective prices and margins.”

When considering the manner in which the differences between underground and submarine cable projects might be represented in an analysis like that of Dr Jenkins, I consider that this factor needs to be handled very carefully, first because the essential object of Dr Jenkins’ exercise was to assess the level of the Cartel-induced overcharge, and secondly because both the underground and the submarine cable markets were part of the Cartel.

101.       In cross-examination of Mr Röstlund, it was suggested that there was a degree of interchangeability or equivalence between underground and submarine cable projects because both types of project were allocated by the Cartel, and the participants in the Cartel would be determined to obtain a “fair” allocation. This, it was put, suggested that the differences between the two types of project were less significant than Mr Röstlund was saying. [83] Mr Röstlund did not accept this, and I do not consider that simply because it is possible to allocate different types of project across the Cartel that fact renders the technical or pricing differences articulated by Mr Röstlund any the less relevant. It seems to me that the members of the Cartel would focus on the value of a particular project to them : that value would turn on a number of subjectivities – their factory loading, their ability to do certain types of work, their margins etc.

102.       In short, I accept the evidence of Mr Röstlund so far as the differences between submarine and underground cable projects are concerned. The significance of these differences – for the purposes of the experts’ evaluation – is a matter considered later on in the Judgment.

(3)     Alternating current or direct current

103.       The BritNed cable was a high voltage direct current (“HVDC”) cable. [84] Mr Rose described the rationale for a direct current cable as follows: [85]

“Given the length of cable required to traverse the North Sea (approximately 245km of submarine cable and, once the cable had “arrived” on land, 9km of land cable) BritNed decided to utilise a high voltage direct current (HVDC) cable. The rationale behind the decision to use a direct current (“DC”) cable, as opposed to an alternating current (“AC”) cable, was primarily because DC cables have far lower transmission losses when compared to AC cables. Transmission losses refer to the amount of energy lost (i.e., which in effect “leaks out” mainly in the form of heat) when being transferred through the interconnector. The greater the length of the cable, the more significant the impact of transmission losses becomes. In simple terms, transmission losses result in lost revenues which, for obvious reasons, BritNed was keen to minimise.”

104.       Mr Röstlund’s evidence was that:

(1)         Submarine AC and DC cables shared many of the features described in paragraph 100 above. [86] Both were very different from underground cables. [87] But that there were design implications in the AC/DC choice, notably in terms of the need for converter stations when using DC cables. [88]

(2)         The constant electrical flow of DC lines means that over long distances, the amounts of power losses through a DC cable are lower than for an AC cable. For long distances – anything over 100 kilometres with a power rating of above a few hundred MWs – DC would be chosen. [89]

(3)         Although there are design differences between AC and DC cables, resulting in differences in cost, [90] the real additional cost is the need for converters at either end of the cable, to convert the current from DC to AC. [91]

105.       The BritNed project was – given the distance (260 kilometres) and power rating (1,000MW [92] ) – plainly always going to use a DC cable. [93] The price for using a DC cable, however, was the requirement for converter stations in both the UK and the Netherlands: [94]

“The Interconnector is connected at its landfall in both the UK and the Netherlands to a converter station. Given that the electricity grids in both the UK and the Netherlands supply electricity in AC rather than DC format, the Project required the construction of converter stations to convert the AC electricity into DC format for transmission through the interconnector and to convert it back again into AC format for onward supply to customers through the respective electricity grids in the UK and the Netherlands.”

106.       The BritNed interconnector thus comprised two, major, elements:

(1)         Converter station construction; and

(2)         Cable manufacture and installation. [95]

107.       The distinction between AC and DC cables highlights two important, and different, measures when considering the costs of a project: capital expenses (“CapEx”) and operating expenses (“OpEx”). The terms are self-evident: CapEx represents the expenditure needed to acquire an asset, here the cost of the building the Interconnector. OpEx represents the costs of the day-to-day running of that asset. Very often, there may be a trade-off between the two. Clearly – since no-one appears to have suggested an AC cable for the Interconnector – the OpEx costs over the life of the project will have outweighed the additional CapEx costs of paying for the converters implied by a design based on a DC cable.

(4)     Cable width

108.       Generally speaking, the higher the power rating of the cable, the larger the dimensions of the conductor and therefore the larger the amount of material required for the conductor. [96] Thus, a 1,000MW capacity will imply a thicker cable than a 500MW capacity.

109.       The relationship between cable thickness and power rating is not invariably proportionate. An alternative to increasing the cross-section of the cable conductor, in order to reach a higher power rating, is to increase the voltage. [97]

110.       However, the diameter of the core of a cable, which is made of copper, affects the level of transmission losses. The thicker the cable, the lower the level of transmission losses – but the greater the cost of the cable. [98] There is, thus, once-again, a potential trade-off between CapEx and OpEx in the choice of cable width. This was a factor that BritNed had in mind. [99]

F.      THE HISTORY OF THE BRITNED TENDER

(1)     Genesis

111.       As has been described, BritNed was a joint venture between National Grid and TenneT. [100] There is a shareholders’ agreement, between National Grid, TenneT and BritNed setting out the terms of their relationship. [101]

112.       The purpose of the BritNed Interconnector was to enable providers of electricity in both the UK and the Netherlands to meet demand for electricity in these different jurisdictions. A UK provider could thus meet a Netherlands demand and vice versa .

113.       BritNed regarded the Interconnector project as commercially quite risky, for reasons explained by Mr Rose: [102]

“12.   At the time the Project was first conceived in and around 2000/2001, not only was there no interconnector between the Netherlands and the UK electricity markets, but it would have been the first interconnector to have been built between the UK and mainland Europe in approximately 20 years. Its development began against the backdrop of the collapse, in 2003, of the North Sea Interconnector (“NSI” project, a proposed submarine interconnector project between Norway and the UK (which was not dissimilar from the Project), despite the significant development expenditure of the two key stakeholders, Statnett and NG.

13.     The capacity of the cable (namely, the “volume” of electricity capable of being transported between the UK and the Netherlands) was to be marketed through short-term auctions on the basis of capacity contracts of different durations. It was not expected that the duration of the capacity contracts would, or indeed could, exceed one year, given that there was little appetite in the market at that time for long-run contracts of that nature (i.e., customers would not commit beyond a year) and because of the regulatory pressure, particularly from the European Commission, to encourage open access to transmission.

14.     It should be borne in mind that the interconnector is a separate business from the regulated business of National Grid and TenneT. It is a commercial merchant link and therefore its primary goal from the joint venture participants’ perspectives was to make a profit. At the same time, the system for the sale of the capacity (and the fact that interconnectors are “price takers” (i.e., the price of capacity is determined by the market rather than the operator)) meant that the joint venture participants did not have the security of having long term fixed price contracts in place with customers. This inevitably entailed TenneT and National Grid taking on merchant risk, without having the security of long term contracts to underpin the investment that was necessary to construct the interconnector.

15.     As a result of the above, the Project was an inherently risky investment and meant that the business case had to meet certain risk-adjusted financial “hurdles” in order to be deemed viable and proceed. Indeed, it was determined, in conjunction with the joint venture participants, that, ultimately, the business case for the Project had to show that it had, at the very least, in internal rate of return (“IRR”) of between [X] [103] and a substantial Net Present Value (“NPV”).”

(2)     The lots

114.       The tender process around the Project was based upon seeking tender prices for three lots: [104]

(1)         Lot 1. An engineering, procurement and construction (“EPC”) contract for the design, engineering, procurement, manufacture, testing, installation and commissioning of the HVDC converter stations (“Lot 1”).

(2)         Lot 2. An EPC contract for the design, engineering, procurement, manufacture, testing, installation and commissioning of the HVDC cable system (“Lot 2”).

(3)         Lot 3. An EPC contract for the design, engineering, procurement, manufacture, testing, installation and commissioning of the HVDC converter stations and the HVDC cable system (“Lot 3”). Lot 3 thus combined Lots 1 and 2.

(3)     Transmission capacity

115.       A BritNed procurement and contracting strategy paper recommended that suppliers be asked to tender for three transmission capacity options or “Base Cases”:

(1)         700MW (subsequently changed to 650MW); [105]

(2)         1,000MW; and

(3)         1,320MW.

This was in order to maintain BritNed’s flexibility in terms of the Interconnector it was seeking. [106]

116.       BritNed was conscious that asking tenderers to price three bids, rather than one, would involve tenderers incurring additional costs, and might result in “half-hearted” bids for one or more of the options by one or more of the tenderers. [107] However, the advantage of flexibility was considered to outweigh this risk, [108] and BritNed planned to strive to delete capacity options, when it became clear that such an option ceased to be viable for technical reasons or on the basis of price or both. [109]

(4)     The OJEU Notice (August 2005)

117.       BritNed caused to be published in the Official Journal of the European Union a Period Indicative Notice in relation to the Project in August 2005 (the “OJEU Notice”). [110]

(5)     BritNed’s negotiating team

118.       Mr Rose was in overall charge of the BritNed interconnector procurement. It was he who had to ensure that BritNed’s procurement strategy and procurement team were properly in place. [111] He was responsible for bringing in Mr Jackson to lead the negotiations for BritNed. [112]

119.       Mr Rose said this about BritNed’s negotiating team: [113]

“So, the lead negotiator on our behalf was Mike Jackson, supported by Louise Negus and Marco Kuijpers. They were the three primary members of the procurement team. Also, they drew upon the expertise of many people around them, but those were the primary people.”

(6)     Expressions of interest (September to October 2005)

120.       Following the publication of the OJEU Notice, the expressions of interest and tenders subsequently received by BritNed came from the following European entities:

(1)         Siemens AG (“Siemens”) in respect of Lot 1.

(2)         Areva T&D UK Ltd (“Areva”) in respect of Lot 1.

(3)         Nexans Norway AS (“Nexans”) in respect of Lot 2.

(4)         Prysmian Cavi and System Energia S.r.l (“Prysmian”, owned by Pirelli) in respect of Lot 2.

(5)         ABB in respect of Lots 1, 2 and 3.

(6)         A consortium of Prysmian and Siemens in respect of Lot 3. [114]

121.       No expressions of interest were received from any Asian based cable manufacturers. [115] Nexans, [116] Prysmian [117] and (obviously) ABB were all cartelists.

(7)     ABB’s tender team and its approach to tendering

(a)     The ABB tender team

122.       ABB’s tender team was described by Mr Jönsson: [118]

“17.   At the time of the tender process for the BritNed project, I was head of the factory at Karlskrona. Magnus Larsson-Hoffstein was appointed as the tender manager for the BritNed project and had primary responsibility for preparing the response to the tender for the supply of the power cables system of the BritNed project (Lot 2).

18.     For much of the time, I left the running of the tender largely to Magnus Larsson-Hoffstein and his team. Magnus Larsson-Hoffstein was experienced in pricing large submarine power cables projects, as he had been involved in the pricing of Gjøa and Estlink, two other large submarine cables projects. During the BritNed tender, he worked closely with Bo Pääjärvi, of the ABB HVDC (high voltage direct current) power converters team based in Ludvika, Sweden on the project. I would review some of the information that Magnus Larsson-Hoffstein and Bo Pääjärvi (and Åke Nilsson, ABB’s in-house counsel) prepared in producing ABB’s tender response. In particular, I reviewed and considered the risks, insurance and technical specifications of the power cables system element of the BritNed project.

19.     The power converters element of the project was priced separately by ABB’s HVDC converters team. I had no involvement in that pricing, or indeed in relation to any other aspect of the converters bid. In my few discussions with the customer during the tender process, I did, however, seek to emphasise the advantages of buying the whole package (i.e. the converters and the cables) from ABB, including by emphasising the efficiency benefits of having a joined-up team.

20.     The BritNed negotiations were primarily led on the ABB side by its power converters team and, particularly in the early stages, largely took place between Bo Pääjärvi of ABB and Dick Bos, project manager for the BritNed project at TenneT. I therefore attended a few meetings with the customer during the course of negotiations. I became more involved in the discussions with BritNed in the period shortly before the power cables portion of the contract was awarded to ABB…”. [119]

123.       Thus, ABB’s tender team was bifurcated between the Cables element and the Converter element. Mr Pääjärvi was in charge of the Converter side of the process, and I heard relatively little evidence about this. Mr Jönsson was in charge of the Cable side, but:

(1)         He reported to Mr Leupp, via Mr Haugland. [120] Mr Leupp did not know of the Cartel. [121]

(2)         He delegated to Mr Larsson-Hoffstein. Mr Larsson-Hoffstein did not know of the Cartel. [122]

(3)         Mr Pääjärvi’s converter bid team was closely involved. [123]

124.       Mr Larsson-Hoffstein’s influence was most significant during the earlier phases of the tender process, when Mr Larsson-Hoffstein and his team were assessing what, technically, was required by the tender, and how much it would cost. After the submission of a tender, the process became a rather more brutal process of what I termed during the trial as “horse-trading”, where ABB would seek to defend its price, and BritNed would seek to obtain price concessions. Mr Larsson-Hoffstein’s role was much less significant at this stage: if he, personally, made a concession, he would do it on instruction from his superiors. [124] But, as will be seen, it was not he, but Mr Leupp, who made a critical concession in the course of this tender.

125.       It is, therefore, appropriate to consider separately ABB’s processes during the “technical” and “non-technical” stages of the tender process.

(b)     ABB’s methodology during the “technical” process

126.       This process was described in abstract terms by Mr Röstlund in his second statement, Mr Röstlund not himself having been involved in the BritNed project. His evidence was not challenged. Mr Larsson-Hoffstein gave evidence of how he conducted the process in the specific case of BritNed. [125]

127.       In broad terms, without particularly referencing the BritNed project, the process was as described below.

(i)      Pre-tender communications.

128.       Even before a formal invitation to tender, there might be informal communications between the customer and one or more cable suppliers. These might involve high level estimates of cost. [126] The invitation to tender, however, would outline the key technical specifications that the customer sought. Invitations to tender would generally be issued to multiple suppliers. [127] During the tender process, there would be considerable interaction between each cable supplier and the customer, which might well cause the specification in the invitation to tender to develop. [128]

(ii)     Initial design on tender.

129.       On receipt of an invitation to tender, a sales team would be formed to assess the implications of the project. There would be input from various different departments within the ABB power cables business unit. [129] An initial design for the project would be produced, including a description of the requirements to manufacture and install the cable according to the customer’s requirements. [130]

(iii)    Costings

130.       Based upon this design, costings would be produced. This would involve assessing the costs of supplying the services and products required, including internal costs and loading for risk and contingencies plus a margin representing ABB’s hoped-for profit. [131] Mr Röstlund described the role of the tender manager as follows: [132]

“…having received costings from its other departments, as well as any estimates of prices for third party suppliers or supplies from other ABB business units, the tender manager will assess the need to factor in any risks and contingencies as cost line items, and will also build in a margin based on the indicative margin targeted by the ABB Division in charge of power cables in a given year…No additional margin is applied to the costings by different departments involved in the supply of a cable (e.g. cable installation and manufacture) before this overall margin is applied. However, if together with the cables system, the turnkey project includes elements that are supplied by another ABB business unit – for example when the installation of the onshore cable is carried out by the local ABB entity – a separate price will be provided for that element, reflecting an internal margin applied by the relevant ABB business unit in charge of supplying this.”

Mr Röstlund expanded upon this later on in his second statement: [133]

“128. In addition to the underlying costs of producing and installing the power cable, and the risk involved, the price will also incorporate the margin that ABB seeks to earn on the project.

129.   As explained in paragraph 80(e) above, in preparing an offer for a submarine cable project, the tender manager assesses the costs, risks and contingencies associated with the project. A margin will then be added on top. This is referred to as the “costs plus” approach. The costs, risks and contingencies used to calculate the project’s gross margin are those that are directly attributable to the supply of the project. These costs include mainly:

(a)     any research and development related to the design and tests of the cable;

(b)     materials, labour and machinery used in the production of the cable;

(c)     transportation and installation of the cable;

(d)     risks and contingencies associated with the design, production and installation of the cable.

130.   The costs used in the calculation of the net margin of a project are those listed in the paragraph above plus an allocation of the general and commercial overheads from the ABB Power Cable Business Unit.”

131.       Mr Larsson-Hoffstein confirmed that he was responsible for putting together the initial pricing proposal for the Cables element of the BritNed project. [134] He adopted the approach described by Mr Röstlund. [135] As regards the compilation of the costings, Mr Larsson-Hoffstein stated that “[t]his did not involve starting with the pricing of any other project, but instead involved looking at the cost of each element of the project (e.g. conductors, insulation, and raw material prices) and putting together a costing on that basis”. [136]

132.       Mr Larsson-Hoffstein was asked to describe the process in a little greater detail during the course of his cross-examination: [137]

Q (Marcus Smith J)

So, Mr. Larsson-Hoffstein, your responsibility in this process at the start was compiling the tender price for the cabling side?

A (Mr Larsson-Hoffstein)

Tender price and the tender document.

Q (Marcus Smith J)

And the tender document, yes, of course, I appreciate it is more than just a figure, a lot of work goes into it.

I appreciate that your superiors had some input into that process. First of all, can you enlighten me as to the number of people working below you who would have helped you put together the tender documents for the cabling side or was it just you?

A (Mr Larsson-Hoffstein)

No, it was a team. The price depends quite a lot [on] the technical side, so we had engineers, we had people from procurement, we had people from manufacturing and the installation. So the price was built up.

One part was the cable design, which was included by the engineering solution under the lead engineer.

Another big part of the price is the installation, where we were dependent on sub-suppliers. The installation department together with procurement could influence the price on the subcontractors. So that was the whole team working here.

Q (Marcus Smith J)

So when in your witness statement you say you adopted a costs plus approach, that is actually rather minimising the amount of work that goes into putting together a tender.

Let me explain why I say that. You have to first of all work out how you are going to meet the tender specification?

A (Mr Larsson-Hoffstein)

[Nods.]

Q (Marcus Smith J)

So you have to do all the technical work there?

A (Mr Larsson-Hoffstein)

Yes.

Q (Marcus Smith J)

And only when you have worked out what your technical solution is – how you are going to do it – can you actually start working on costs?

A (Mr Larsson-Hoffstein)

Yes, in principle.

Q (Marcus Smith J)

And, presumably, that involves speaking to subcontractors to work out their own price for a bit of the contract?

A (Mr Larsson-Hoffstein)

Yes.

Q (Marcus Smith J)

So, would you, or your team, have made inquiry of – for instance – the provider of a cable-laying ship?

A (Mr Larsson-Hoffstein)

Yes. Yes, not the ship, as such, but installation works including…

Q (Marcus Smith J)

Including the ship?

A (Mr Larsson-Hoffstein)

Yes, yes.

Q (Marcus Smith J)

And how specific would your team’s inquiries be? Let’s take the installation, including the ship, as an example. As I understand it, ABB didn’t have its own cable-laying ship?

A (Mr Larsson-Hoffstein)

That’s correct.

Q (Marcus Smith J)

So, presumably, you would have looked into who could have provided that facility?

A (Mr Larsson-Hoffstein)

Yes.

Q (Marcus Smith J)

And suppose you encountered a situation– and I am just talking hypothetically here – where, within the anticipated time for the completion of the project, there simply was no cable-laying ship available? What would you do then?

A (Mr Larsson-Hoffstein)

In fact, that was the situation for BritNed. The two main bids that we had from Van Oord and from Oceanteam was based – from Van Oord it was a big modification of a ship or a barge, as I remember, and the other bid from Oceanteam was based on a ship that was not built yet.

Q (Marcus Smith J)

I am moving ahead of myself, a little bit, but that sort of risk, would that be factored into the margin that you would add to the costs plus approach that you took?

A (Mr Larsson-Hoffstein)

Yes. It always is taking a subcontractor. Even if we try to push as much risks as possible to the subcontractor, the limits – their limits – of their liability is a percentage of the subcontract, which is much lower than ABB’s total liability, which is a percentage of a total price. So it is always an intermediate risk of the main contractor that needs to be priced in.

Q (Marcus Smith J)

Now, I think it follows, but correct me if I am wrong, that your cost-plus approach means that you only look at the direct costs relating to the project in question?

A (Mr Larsson-Hoffstein)

Direct cost – there are also overhead costs, and so on; that is then add-on in the price calculations, administrative and sales overhead and so on and so forth.

Q (Marcus Smith J)

Right. But is that added later on?

 

A (Mr Larsson-Hoffstein)

It is – first, we take the direct costs, and then it is added on, yes.

Q (Marcus Smith J)

Yes. So taking it in stages. You start with the direct cost?

A (Mr Larsson-Hoffstein)

Yes.

Q (Marcus Smith J)

You then add a margin which is related to the specific risks arising out of the project that you have identified?

A (Mr Larsson-Hoffstein)

Yes, yes.

Q (Marcus Smith J)

And then you add a further margin, which is what I would call “common costs” like overheads…

A (Mr Larsson-Hoffstein)

Yes.

Q (Marcus Smith J)

…and your profit?

A (Mr Larsson-Hoffstein)

Yes, on top of that is the profit, yes.

Q (Marcus Smith J)

And when you have reached the end of those various stages, you have, obviously with the documents, what we call your tender price?

A (Mr Larsson-Hoffstein)

Yes.

Q (Marcus Smith J)

Now, in the process leading up to the finalisation of your tender price, what is your practice in terms of discussing problems or questions with your superiors in the process? What do you do? You don’t, presumably, hand them one tender price with all the documents saying “This is it”?

A (Mr Larsson-Hoffstein)

In general, we make a presentation of the overall technical solution and the related risks and the price and the answer. So it is not only the price; it is also presentation of the whole solution.

Q (Marcus Smith J)

So it is an iterative process?

A (Mr Larsson-Hoffstein)

Yes, yes.

Q (Marcus Smith J)

And would I be right in thinking that where there are certain issues – now, they might be technical issues or they might be issues about availability, like the cabling ship – but where there are particular issues which will have a bearing on the tender price, you would presumably take care to identify those to your superiors?

A (Mr Larsson-Hoffstein)

Yes.

Q (Marcus Smith J)

And at that stage you would get guidance or instruction back saying “Well, this is how we think or we say you should treat that particular risk”?

A (Mr Larsson-Hoffstein)

Well, we had some meetings, the management team and the tender team, where we present different solutions and there was some common conclusion on the way to go from those meetings.

Q (Marcus Smith J)

So, let’s take an example of an alternative set of technical solutions to a particular issue, one costing X and one costing XX. At that stage you would have a debate within the team, including your superiors, to say “Well, on what technical solution will we base our tender price”?

A (Mr Larsson-Hoffstein)

Yes.

Q (Marcus Smith J)

And there would be a debate and there would be a decision?

A (Mr Larsson-Hoffstein)

Yes.

Q (Marcus Smith J)

And once the decision has been made, that is what you would bake into your tender price?

A (Mr Larsson-Hoffstein)

That’s right.

Q (Marcus Smith J)

Is there a paper trail which can show the evolution of your tender price over time? In other words, do you present – I’m sure there is a huge volume of papers, I’m interested in something which might be less voluminous, something which shows how the various decisions that resulted in the final tender price that was submitted to the customer, how that was arrived at? In other words, can one see the various delta points that existed, whereby decisions were made as to certain pricing choices?

A (Mr Larsson-Hoffstein)

I think it is difficult to have the full tracking. Some decisions were done verbally. Many were followed up by email. But at certain points, when we submitted different bids to the customer, we updated the cost calculations. So those points where we made negotiated bids and updated the bids to the customer can be tracked back to cost calculations.

Q (Marcus Smith J)

Yes. I’m interested at the moment in the process before the customer ever gets involved. So it is at the stage before you have got a tender price that you submit to the customer.

What I wondered was whether you had, as it were, draft costings with the margins appended, where you would say, “Well, here is where we are at on this date?”. You would then discuss that document and make whatever changes arose out of that debate to the next draft.

Is that how you worked or am I completely wrong?

A (Mr Larsson-Hoffstein)

In principle, but I’m not sure if we have a document with every decision and if it is fully traceable.

(iv)    Risk review

133.       Larger cable projects would go through an internal ABB risk assessment exercise, known as a “risk review process”. The aim of the process, unsurprisingly, was to ensure that the relevant risks and contingencies had appropriately been considered and that the price being put forward generated the minimum level of expected margin. [138] It was possible for there to be multiple risk reviews. [139] The BritNed project was risk reviewed. Mr Larsson-Hoffstein said this in relation to the risk review of his costings for the BritNed project: [140]

“I ran my proposed price past my managers within ABB at this time, including Mr Jönsson, but no alterations were made to those figures. Similarly, no adjustments were made to my figures as a result of the risk review process for the project.”

(v)     Submission of a tender

134.       As was the case with BritNed, there might be two discrete components to the tender. In the case of BritNed, there were the Cable element and the Converter element. These discrete elements would have to be combined, to form ABB’s tender.

(vi)    Use of Product Pricing Models

135.       ABB’s cables unit used Product Pricing Models (or “PPMs”) throughout the tendering phase to help determine the sales price of a project. Such PPMs would be used on an iterative basis, frequently updated during the tender process so as to enable the reassessment of costs and risks. [141]

(c)     The process post-tender

136.       This was much less formalised and, as I have noted, was really a process of negotiation. What occurred in the course of the BritNed tender is described below.

(8)     Pre-qualification to tender (March 2006)

137.       Tenderers who had pre-qualified for one or all of the Lots were informed that they had done so by letter dated 6 March 2006. [142]

(9)     Tender evaluation “Phase One”

138.       On 7 April 2006, BritNed produced a document entitled Tender Evaluation Procedure: Phase One (19 June – 3 July) (the “Phase One Procedure Paper”). [143] The paper set out, amongst other things, the various Base Cases for which suppliers would need to submit prices. [144] The paper also confirmed that, following evaluation of the bids, BritNed anticipated selecting one capacity option from the three Base Cases. Mr Rose describes the position as follows: [145]

“The Phase One Procedure Paper also confirmed that, following the evaluation of bids provided by the suppliers, the Project Team “anticipated” selecting one capacity option from the three Base Cases after it had made a recommendation to the JV Board on 30 June 2006 and the decision had been notified to tenderers on 3 July 2006…Once the capacity option had been chosen, fixed prices would be required at the selected capacity during Phase Two, which would commence after the notification to tenderers had been dispatched on 3 July 2006. However, the Phase One Procedure Paper also confirmed that there was a possibility that two capacity options could “be taken forward to Phase Two” should they both be deemed viable during the evaluation process for Phase One.”

139.       A pre-tender meeting and site visit took place at Maasvlakte on 19 April 2006, which all tenderers attended. [146] At this meeting, it was made clear that there was no guarantee that the Project would go ahead: this would depend upon the business case, which was itself obviously in part dependent on the level of tenders received. [147] BritNed also refused to be drawn on which of the three Base Cases it favoured. [148]

140.       After this meeting, various of the suppliers indicated that they would not be able to provide prices in respect of all of the Base Cases requested by BritNed. [149] This resulted in an amendment to the Phase One tender process. [150]

(10)   ABB’s discussions with Nexans

141.       During the course of the Phase One tender process, ABB discussed the bid with Nexans. [151] Mr Jönsson’s evidence was as follows: [152]

“26.   I do not have firm recollections of the pre-qualification stage of the BritNed tender. However, I recall having discussions with Nexans during Phase 1 of the tender process…regarding allocation of the BritNed project. I have been shown a copy of [Decision/Recital (395)] which states that I “met with Mr Romand (Nexans) on 9 or 10 May 2006 at Zurich airport in order to discuss the allocation of the BritNed project. The project was allocated to ABB and the two companies agreed that Nexans would ensure that it would not meet the customer’s delivery requirements in its offer. In exchange, ABB would subcontract a portion of the work to Nexans. ABB has stated that this agreement was made even though the details of the project were not yet clear”. This accords with my recollection.

27.     This initial agreement regarding allocation was my only discussion with another cartelist regarding BritNed. At no stage did I discuss pricing of the BritNed project with Mr Romand or the other cartelists.

28.     We did not include Nexans as a subcontractor in the June 2006 ABB bid…and quickly dropped any prospect of joint tendering with Nexans as it became clear that ABB had sufficient cable factory capacity at the time. We therefore proceeded alone and only engaged with Nexans (along with other companies with cable laying expertise) in relation to a potential cable-laying role. Mr Romand was annoyed with me for ignoring the initial discussion about allocation. I have been shown [Decision/Recital (424)] which notes that, at a meeting between ABB and Nexans on 4 June 2007, “Mr Romand (Nexans) reproached Mr Jönsson (ABB) for doing all the work for BritNed itself”. This, again, accords with my recollection.” [153]

(11)   The ILEX Study (May 2006)

142.       In order to further investigate the levels of revenue that the Interconnector would generate, BritNed commissioned a study from ILEX Energy Consulting Limited (the “ILEX Study”). The ILEX Study undertook a forward-looking analysis of the potential revenues that the Interconnector could generate. It was provided to BritNed in May 2006. [154]

(12)   Prysmian did not submit a tender (June 2006)

143.       Prysmian had expressed an interest in tendering and had pre-qualified for both Lot 2 and (with Siemens) Lot 3. [155] On 9 June 2006, Prysmian informed BritNed that it would not be submitting a tender. The email in question stated: [156]

“It is with much regret from our part that we have to inform you that Prysmian will not be submitting a bid for this project due to the acquisition of a major order and consequent impossibility to meet the deliveries request for BritNed.

We do hope that this will not cause BritNed…any disruption and take the opportunity to wish you good luck for the successful implementation of this potential submarine interconnection project.”

144.       Prysmian did not identify to BritNed the “major order” in question, [157] but there was no obligation on them to do so.

145.       So far as BritNed was concerned, this meant that the possible consortium between Prysmian and Siemens in relation to Lot 3 could not proceed. [158] Of course, that was a matter of BritNed’s knowledge: what other tenderers – and in particular, ABB – knew is a different question that is considered elsewhere. Prysmian was, of course, a cartelist. [159]

(13)   Areva did not submit a tender (June 2006)

146.       In a letter dated 13 June 2006, Areva wrote to BritNed stating that having reviewed the tender documents, they were “not able” to make an offer for the Project. [160] The reasons for Areva not tendering were discussed before this letter was sent. One of the reasons – but it was only one – was the perceived dominance of ABB in being able to provide a full Lot 3 solution. [161] Areva was not a cartelist. [162]

(14)   Nexans’ tender (June/July 2006)

147.       Nexans submitted initial and revised tenders for Lot 2 in June and July 2006. [163] These tenders were both priced higher than ABB’s bids [164] and were assessed by BritNed as non-compliant on the basis that Nexans was unable to meet BritNed’s required deadline for the completion of the Project. [165] The issue of meeting BritNed’s deadline was explored further with Nexans, [166] but proved to be impossible to resolve. Given its inability to meet the required deadlines by some considerable margin, BritNed ultimately (in October 2006) took the decision not to proceed with Nexans. [167] Nexans was also a cartelist. [168]

(15)   ABB’s tender (June 2006)

148.       ABB submitted its tender in accordance with the amended Phase One tender process on 16 June 2006. [169]

(16)   BritNed’s position at this point

149.       Mr Rose described BritNed’s position in the following terms: [170]

“38.   At the outset of the procurement process, members of the Project Team produced a document, dated 16 January 2006, entitled “Main outlines of procurement process” (the “Procurement & Contracting Strategy Paper”).

39.     The Procurement & Contracting Strategy Paper covered a number of issues, including the contracting strategy for the Project, which was based, in part, on the expressions of interest that had been received from Prysmian, Nexans, Siemens, Areva and ABB. In particular, the paper outlined the three contracting options available, with the “Single EPC approach” (namely, one tenderer or consortium being awarded Lot 3) being identified as the “preferred option”. The alternative was to award separate contracts for separate and distinct lots of work (award one contract for the complete converter station system (Lot 1) and a separate contract for the cable supply and installation (Lot 2)). This, however, involved some “interface” risk between the respective contractors.

40.     However, in the absence of any expressions of interest or bids whatsoever from Asian suppliers, the failure of Prysmian to submit any bids despite its previous expressions of interest, and after Nexans effectively disqualified itself from the process by stating that it was unable to meet BritNed’s delivery deadlines, BritNed had no choice but to undertake a procurement process in the full knowledge that there was no competition for ABB in respect of both Lot 2 and Lot 3.

41.     In particular, the failure of Prysmian to submit a bid for the cable element of the Project not only meant ABB faced no competition from Prysmian in respect of Lot 2 but it also ensured that ABB faced no competition in respect of Lot 3, as without a consortium partner with cable expertise, Siemens was unable to bid for the combined package. This, in addition to Nexans’ non-compliant tender submission, was a significant disappointment, given that it removed any competition for the cable element of the Project from the very outset and inevitably limited our scope for manoeuvre thereafter, in particular the ability to maintain any competitive pressure on ABB.

42.     Consequently, despite the Project Team’s preferred contracting approach being the single EPC approach, it had to consider the alternative of awarding separate contracts Lot 1 and Lot 2, a contracting strategy that the Project Team had stressed throughout to the tenderers was a viable alternative to the single EPC approach, should it be necessary to pursue it.”

I accept this as an accurate statement of BritNed’s position.

(17)   Evaluation by BritNed

150.       The joint venture board considered the outcome of the Phase One tender process on 30 June 2006. [171] Base Case 1 was rejected as not being competitive on a price per megawatt basis when compared to either Base Case 2 or Base Case 3. [172]

151.       According to Mr Rose, the project was viable, but there were considerable commercial risks. [173] He recommended pursuing Base Case 2. [174] As to Base Case 3, Mr Rose said this in his witness statement: [175]

“However, if the budgetary prices provided by ABB for Base Case 3 had been lower, Base Case 3 would inevitably have been more attractive. Indeed, in accordance with the Capped Annual Revenue Figures in the ILEX Study, Base Case 3 would have meant access to additional revenue of EUR 11m per annum. From a revenue perspective, this option would therefore have presented a distinct advantage when compared to Base Case 2. As it was, the pricing received from ABB in its Initial Tender Returns for Base Case 3 ensured that the level of CAPEX that would have been required to pursue this option posed a substantial risk given the revenue uncertainty and precluded any recommendation being made to the JV Board that this option should either be pursued or explored further.”

(18)   Tender evaluation “Phase Two” (July 2006 onwards)

152.       On 3 July 2006, BritNed notified the tenderers of the decision to proceed with Base Case 2. [176] On 24 July 2006, BritNed produced a document entitled Tender Evaluation Procedure: Phase Two (31 July – Financial Close) (the “Phase Two Procedure Paper”). [177]

153.       The Phase Two Procedure Paper, which (like the Phase One Procedure Paper, was internal to BritNed) recorded some of the developments during Phase One tender evaluation and set out a timetable for Phase Two. It stated: [178]

“As part of the Phase One evaluation process, Prysmian declined to submit a Tender for Lot 2 and therefore the Siemens/Prysmian joint venture withdrew from Lot 3. No submission was received from Areva for Lot 1 and subsequent to 30 th June 2006 a letter has been received confirming their decision not to participate further in the Tendering Process.

As part of the Phase One evaluation process, the capacity options were reduced to one capacity and voltage option, namely 1,000MW – 450kV. A limited number of optional prices have been requested as part of this Phase Two Tender submission and a full and detailed breakdown of the Tender Prices. As part of the Phase Two submission, the Tenderers are required for Lots 2 & 3 to provide firm prices for the cable installation element of the project, which was only submitted as a budget price at Phase One, no other changes have been requested.”

BritNed’s approach to evaluation was that the tenderer who represented the “most economically advantageous solution” would be selected as the preferred tenderer or tenderers. [179] This had been made clear to the tenderers themselves in instructions to tenderers dated 6 April 2006. [180]

154.       Thus, by the commencement of the Phase Two tender evaluation, the only Base Case that suppliers were tendering for was the 1,000MW option. [181]

155.       ABB submitted revised tender returns in respect of Lots 2 and 3. [182] Nexans submitted a tender return in respect of Lot 2. [183] Nexans’ bid was some €41,338,664 more expensive than ABB’s equivalent bid. [184]

156.       On 10 October 2006, BritNed held a technical and procurement meeting to decide upon the strategy to be followed with the remaining tenderers during the negotiation phase of the project, scheduled to commence in October/November 2006. [185] It was at this meeting that the decision was taken not to progress with Nexans, given Nexans’ inability to provide the cable within the required timeframes. [186] BritNed intended to begin detailed negotiations only with ABB and Siemens. Again, this was a decision that was internal to BritNed.

157.       On 10 November 2006 and 13 December 2006, the joint venture board was updated on the state of play. [187] The Interconnector project continued to remain economically viable, but only just. [188] What appears to be a Powerpoint slide presentation entitled “Procurement Update December 2006” noted that under the current market position: [189]

“•       Only ABB can provide cable in required timescales

 •       Both ABB and Siemens have limited but available capacity for converter stations

 •       Both have heavy tendering workload in Converter Stations and limited resource or appetite to engineer multiple options”

(19)   Best and final offers (March 2007)

158.       Best and final offers (“BAFO”) were received from ABB and Siemens in March 2007. [190] According to Mr Rose, “[w]hilst price negotiations continued with both tenderers, the Project Team was largely content with Siemens’ BAFO but extremely dissatisfied with those submitted by ABB”. [191] One aspect of ABB’s BAFO was that its prices for the cable differed as between Lot 3 and Lot 2, with the prices in Lot 3 being lower. [192] Clearly, ABB was prepared to offer better value if it could obtain the entire job, i.e. the Cable element as well as the Converter element.

159.       An internal ABB email dated 10 March 2007, sent to Mr Pääjärvi and copied to Mr Jönsson recorded ABB’s perception:

“1.     BritNed will take a very tough position in the cable meetings, making it clear to ABB that based on the BAFO, the project will not get a go.

2.       In the cable meetings next week, BritNed will request ABB to further reduce the price and to accept a more aggressive risk allocation…

3.       At some point during the next coming weeks, BritNed will again introduce the converters into the discussions. They will try to squeeze ABB on the total package. BritNed will use two major arguments:

·          If the total investment cost is not reduced and if the total risk exposure for the owners are not lowered, the project will not be approved by the Board.

·          If ABB do not lower the price on the Converters, BritNed will award this part to another supplier…

The first meetings next week will be very important for us in order to get a better understanding of what BritNed really expects (price and risks). Based on this, we need to develop a strategy for how to proceed…”

(20)   Last and final offers (March 2007)

160.       ABB submitted its last and final offer (“LAFO”) on 28 March 2007. [193] ABB was prepared to offer a €10m reduction on the cable element of Lot 3, should it be chosen to provide the combined package. [194]

(21)   Final Cable price negotiations (April 2007)

161.       By this stage, BritNed only discussed the prices that had been submitted to it in respect of the Cable element of the project. In other words, there were only discussions with ABB, and not with Siemens. [195] A meeting was held with ABB on 18 April 2007. [196]

162.       At this meeting, ABB was persuaded to provide the Cable in Lot 2 for the price offered for Lot 3 in the LAFO. [197] In other words, the discount which had previously only applied if ABB was awarded the entire contract (Lot 3) was to be applied to the Cable only part comprising Lot 2.

163.       After the work on the project had been awarded, the parties were naturally more open about their thinking during the tender process. In particular, ABB received feedback from TenneT regarding ABB’s loss of the Converter element of the project. ABB recorded the feedback as follows:

“Today, I had lunch with Dick Bos, project manager acquisition BritNed on TenneT side. Afterward, we had a short chat with Marco Kuijpers, the purchaser. They gave me the following feedback on our lost the order.

Dick was very surprised that ABB gave a discount without a firm coupling of the package (cables + converters). The package was the trump card on which we should have won the order. In his perception (Dick has been a professional BtB salesman in aviation) this was our biggest mistake and really opened the door to go with Siemens!”

164.       Mr Rose did not appreciate that this was Mr Bos’ view: but, independently of Mr Bos, Mr Bos’ view reflected his (Mr Rose’s) own, [198] and he regarded it as “an unusual behaviour”. [199] He went on: [200]

“…it is an unusual behaviour if you are trying to direct people to take your main bid, the Lot 3. I don’t know what was going on in the ABB camp at that point. It strikes me even more as slightly odd behaviour, perhaps not very joined up behaviour, but I don’t know…”

165.       Mr Jackson also regarded ABB’s conduct in this regard as incomprehensible: [201]

Q (Mr Hoskins, QC)

When you say:

“We do not understand why ABB gave away the discount €10m for the Lot 2.”

Can you just explain what you meant by that?

A (Mr Jackson)

It is similar to the discussion we were having earlier in relation to Dick Bos’ comments that ABB had in their power, if you like, the price to – the opportunity to use the fact that they were the only supplier of cable to alter the price in Lot 3 by giving the larger discount in the cable in that area and no discount in Lot 2.

Q (Mr Hoskins, QC)

You say they were the only supplier of cable. Why would they make you this surprising offer for Lot 2?

A (Mr Jackson)

Hence the comment. That is why I don’t understand why they did it.

Q (Mr Hoskins, QC)

It may well be in fact that the reason they made you this offer that was surprising to you was because they felt under such competitive pressure they felt compelled to do it?

A (Mr Jackson)

No, I don’t believe that.

Q (Mr Hoskins, QC)

That is a possibility, isn’t it?

A (Mr Jackson)

I don’t believe that.

166.       Knowing what they did, those on BritNed’s side of the negotiations obviously regarded this as an inexplicable error on the part of ABB. But, of course, ABB did not or did not necessarily know what BritNed knew. Instead of a mistake, the reduction in the Lot 2 price might reflect a concern on the part of ABB that it was not guaranteed to win even Lot 2. This was a decision made not by Mr Jönsson, but by Mr Leupp. [202]

(22)   Letter of intent and contract

167.       A letter of intent was agreed with ABB on 27 April 2007 in which the overall price of the cable contract was agreed at €263m (exclusive of VAT). [203]

168.       Even when the tender had been agreed, some of the prices therein remained provisional. [204] That was so, in the case of the Interconnector, as regards metals, contractors all risks insurance and currency. [205] This remained the position in the final contract between ABB and BritNed. [206] Essentially, the contract provided that the cost of the various provisional items was to be adjusted to reflect their actual cost. [207]

(23)   The price offered by ABB over time

169.       It is convenient to record the price offered by ABB for the cable supply over the course of these negotiations. I leave out of account the price offered by ABB in relation to the Converter element (Lot 1). The table below, thus focuses on Lot 2 and the Cable element within Lot 3. It is also necessary, at least for the earlier prices, to have regard to the three “Base Cases” for which BritNed sought prices. [208]


Lot 2 (cables)

Lot 3 (cable element only)

16 June 2006

ABB’s initial tender returns [209]

 

 

Base Case 1 (650 MW)

 

 

Cable supply

€152,485,360

€152,485,360

Cable supply (provisional sum)

€75,190,317

€75,190,317

Cable installation

€60,000,000

€60,000,000

CAR insurance

€33,802,878

€33,802,878

Total

€321,478,555

€321,478,555

Base Case 2 (1,000 MW)

 

 

Cable supply

€135,240,643

€135,240,643

Cable supply (provisional sum)

€49,756,790

€49,756,790

Cable installation

€60,000,000

€60,000,000

CAR insurance

€29,000,000

€29,000,000

Total

€273,997,433

€273,997,433

Base Case 3 (1,320 MW)

 

 

Cable supply

€174,544,067

€174,544,067

Cable supply (provisional sum)

€74,033,904

€74,033,904

Cable installation

€66,029,836

€66,029,836

CAR insurance

€37,057,318

€37,057,318

Total

€351,665,125

€351,665,125

July 2006

ABB’s revised prices [210]

 

 

Base Case 2 (1,000 MW) [211]

 

 

Cable supply

€135,240,643

€135,240,643

Cable supply (provisional sum)

€49,756,790

€49,756,790

Cable installation

€72,931,016

€72,931,016

CAR insurance

€29,500,000

€29,500,000

Total

€287,428,449

€287,428,449

March 2007

ABB’s best and final offer [212]

 

 

Base Case 2 (1,000 MW)

 

 

Cable supply

€135,276,491

€131,830,648

Cable supply (provisional sum)

€36,308,024

€36,308,024

Cable installation

€81,699,344

€80,065,358

CAR insurance

€28,150,000

€27,587,000

Total

€281,433,859

€275,791,030

March 2007

ABB’s last and final offer [213]

 

 

Base Case 2 (1,000 MW)

 

 

Cable supply and installation

€248,000,000

€238,000,000

CAR insurance

€28,150,000

€27,587,000

Total

€276,150,000

€265,587,000

April 2007

Final negotiations with ABB [214]

 

 

Base Case 2 (1,000 MW)

 

 

Cable supply and installation

€233,000,000

€233,000,000

CAR insurance

€28,000,000

€28,000,000

Total

€261,000,000

€261,000,000

21 May 2007

Amount in the concluded agreement [215]

 

 

Base Case 2 (1,000 MW)

 

 

Cable supply

€120,403,580

 

Cable supply (provisional sum)

€36,915,826

 

Cable installation

€77,165,925

 

CAR insurance

€27,587,000

 

Other insurance

€1,000,000

 

Total

€263,072,231

 

Table 1 : The price offered by ABB over the course of the negotiations. Prices are exclusive of VAT.

170.       The total contract price was subsequently varied by a deed of settlement, dated 15 December 2011, and increased to €280,749,582.72. [216]

G.     COMPETITIVE PRESSURES ARISING IN THE NEGOTIATION PROCESS

(1)     Introduction

171.       In order to assess the manner in which the Cartel had or may have had an effect on the finally concluded price reached between BritNed and ABB, it is important to pay regard to the process of negotiation between the parties; the competitive pressures that existed; and the extent to which these were subverted by the Cartel.

172.       Having considered, in Section F, the history of the negotiations between BritNed and ABB, this Section seeks to articulate the competitive pressures operating on ABB and BritNed during that process. In particular, it considers the extent to which BritNed was affected by the Cartel; the extent to which BritNed was able to deal with the effects of the Cartel; and the extent to which ABB was able to exploit these effects. Specifically, the following factors are considered:

(1)         The limited response from bidders that BritNed had to contend with. One of the objectives of the Cartel was to limit the number of suppliers putting forward a bid or a competitive bid (for, as has been seen, cover pricing did take place) for a particular project, so as to ensure that one particular supplier got the work. [217] This occurred in the case of the BritNed tender: ABB was the “preferred” bidder of the Cartel for the BritNed tender, resulting in fewer actual bids and also in non-compliant bids. [218] This obviously had an effect on BritNed’s ability to play one bidder off against another. This is considered in Section G(2) below.

(2)         The various commercial pressures that BritNed sought to deploy during the negotiations. Notwithstanding the thin field that was bidding for the BritNed Interconnector contract, and the lack of competition consequent on this, there were a number of commercial pressures that BritNed could, and did, deploy. It is necessary to understand what these were: they are considered in Section G(3) below.

(3)         The knowledge advantage ABB derived from the Cartel. As I have noted, the limited number of parties interested in tendering for the Interconnector was one consequence of the Cartel. Another was ABB’s appreciation that it had been “allocated” the BritNed project by the Cartel. It is very important to understand precisely what advantage ABB derived from this. This is a question that is considered in Section G(4) below.

I should stress that a more limited field of potential suppliers [219] and a knowledge advantage on the part of that supplier who is the preferred supplier under the Cartel [220] are not the only effects of the Cartel. They are two effects of the Cartel considered in this Section because they go to the negotiation process between BritNed and ABB. Two other effects that I will come to [221] are what I term “baked-in inefficiencies” and “cartel savings”.

(2)     The limited response from bidders that BritNed had to contend with

173.       Clearly, as the party conducting the procurement, BritNed knew exactly who was interested and who was bidding and on what terms. As has been described, and as I accept, interest in the BritNed Interconnector project was thin. That may in part have been due to proper, commercial, factors, such as a potential supplier’s inability to meet the deadlines of the project, but a major factor was the Cartel. I consider and find that the operation of the Cartel explains:

(1)         The absence of any bids from Asian suppliers; and

(2)         The absence of any real interest from Prysmian and Nexans. Certainly, neither company submitted a competitive bid, and I find that that was because of the Cartel. As regards the Cable element, these two companies were ABB’s only European competition.

174.       Of course, BritNed did not know of the Cartel and therefore did not know why the response to its invitation to tender was so weak. That, to my mind, is irrelevant to an assessment of BritNed’s conduct during the course of the tender. What matters is that BritNed recognised that the response to its invitation to tender was thin, and it recognised that it would be less able to play-off one supplier against another. [222]

175.       As shall be seen, BritNed did, to an extent, manage to keep pressure on those suppliers actually tendering, despite the thinness of the response to the invitation to tender, but I recognise that BritNed was significantly hampered in its ability to negotiate by the thinness of the response to its invitiation to tender for the Interconnector.

(3)     The various commercial pressures that BritNed sought to deploy during the negotiations

(a)     The commercial pressures identified

176.       Despite being hampered by a thin field of interested suppliers, BritNed was nevertheless able to generate some commercial pressure on bidders. The significant factors that BritNed deployed were:

(1)         Ensuring that Siemens continued to bid for the Converters element of the project.

(2)         Pressing ABB on its failure initially to offer a “full bipole” solution.

(3)         Pressing comparisons with the costs of the (earlier) NorNed project.

(4)         Stressing the risk that the project would not go ahead if the price was too high.

(5)         Stressing ABB’s lack of its own cable laying vessel.

(6)         Noting that ABB appeared to need the work represented by the BritNed Interconnector.

(b)     Keeping Siemens in

177.       BritNed’s negotiating team sought to ensure that Siemens remained a tenderer throughout the process. There were concerns within BritNed that Siemens was seriously considering withdrawing from the process, which would render Lot 3 the only option for BritNed if the Interconnector project were to proceed. [223] BritNed was successful in ensuring Siemens remained a tenderer for Lot 1.

178.       Siemens’ role – albeit limited – was significant. Siemens’ presence created competition in relation to the Converter element of the project, and so (even if, as to which I make no findings at present, ABB knew it was the only bidder in relation to the Cable element) ABB would be under pressure in terms of its desire to win the entire contract (i.e. Lot 3).

179.       I find that Siemens’ participation in the project (i) increased competition and (ii) increased ABB’s perception of competition. Although BritNed’s witnesses were a little reluctant to accept this latter point, in the end they did. Mr Rose said this: [224]

Q (Mr Hoskins, QC)

Mr Rose, your evidence is that you were searching, as one would expect, and perfectly normally, to find ways to put competitive pressure on ABB. One of the ways of putting competitive pressure on ABB was to hold out the Lot 3 carrot, keep them thinking as long as possible they might win Lot 3. The idea you would identify that as a competitive pressure and then not actually convey that to ABB is not particularly credible, is it?

A (Mr Rose)

I don’t agree. I think it is entirely understood, was entirely understood at the time, that the prize of winning the entire project for ABB was there and was obvious for all to see. There was no need to express something quite so obvious and we didn’t, to my knowledge.

Q (Mr Hoskins, QC)

So the point you are making in paragraph 3 is not that there was no competitive pressure on ABB as a result of the Lot 3 carrot; your evidence is that it was so obvious, you didn’t have to tell ABB that that was the position. Is that a fair summary?

A (Mr Rose)

Yes, I think that is a fair summary. The idea that one company can win the lot, the whole lot, should be a prize considering the nature of the project.

180.       In his witness statement, Mr Jackson appeared not to accept the significance of Siemens’ role: [225]

“35.   …I do not agree that BritNed attempted to use the combined Lot 3 package, for which ABB was the sole bidder, as a means to drive down ABB’s cable price…

36.     …BritNed’s preferred contracting strategy would have been to award Lot 3 to either a single tenderer or a consortium of parties with the requisite expertise. However, BritNed could not be seen to rely on Lot 3, since there was only one bidder for the combined package. This meant that Siemens’ bid for Lot 1 was the only means of maintaining any vague semblance of competition in the bidding process for the BritNed Project. To have allowed ABB to believe that BritNed was committed to awarding Lot 3 would have run counter to BritNed’s interests.”

Reading these paragraphs as a whole – and in particular the emphasised part – it is clear that Mr Jackson did see the presence of Siemens as a means of placing commercial pressure on ABB, and I frankly do not understand his denial of this in paragraph 35 of his statement. Of course, I accept that BritNed’s preference might well have been for a Lot 3 solution. But only if the price was right.

181.       Mr Jönsson considered that “BritNed…used the fact that ABB was competing not just for the cable but also for the converter element of the project (in the combined Lot 3) to its advantage by using the “carrot” of the combined bid to put pressure on ABB’s cables price. This was ultimately reflected in the significant discount offered by ABB and was the result in my view of clever negotiation by BritNed.” [226]

(c)     Full bipole versus simple bipole

182.       One of the differences between the Siemens bid and the ABB bid was that Siemens offered a “full bipole” solution, whereas ABB, initially, did not. [227] The technical detail does not matter: it is sufficient to note that a full bipole solution was BritNed’s preferred solution. ABB only moved to offer such a solution in late April 2007. [228] Whilst this difference between the bids continued in the period up to April 2007, BritNed sought to ascribe a monetary value to it, in order to assess the respective bids. [229] BritNed assessed the “negative valuation” for a simple bipole design for the Project at around €25m to €35m. [230]

183.       The failure, on the part of ABB, to provide a full bipole solution was obviously one that played its part in the negotiations. [231] It clearly increased the pressure on ABB regarding the competition ABB perceived it had with Siemens in relation to Lot 1, and so pushed ABB to a more competitive stance in relation to Lot 3.

(d)     Comparison with NorNed

184.       During the tender process, BritNed sought to derive a sense of value and costs by comparing the Project with the NorNed project. [232] The NorNed project was one which had involved TenneT (as purchaser) and ABB (as supplier).

185.       Quite how comparable the two projects actually were is not, for present purposes, the issue. The question is how the point played in negotiations: [233]

Q (Mr Hoskins, QC)

But you are saying you had no awareness that BritNed was using NorNed as a means to try to get ABB to lower its price? Is that your evidence?

A (Mr Rose)

No, I’m saying that I don’t have any – I don’t – I wasn’t involved in any direct discussions on it, but I’m aware that the analysis was done to draw a comparison between the two projects and that the calculations and discussions around that, I was aware of, yes.

Q (Mr Hoskins, QC)

What was the purpose of doing those calculations?

A (Mr Rose)

The purpose was to improve our understanding of any differences between the two projects, to try and understand if we were getting a competitive market price for BritNed. We had no other reason to assume that TenneT hadn’t paid – at Statnett hadn’t paid – a sensible competitive price for NorNed. So as a reference.

Q (Mr Hoskins, QC)

What was the purpose of telling ABB about the conclusions you had drawn?

A (Mr Rose)

To demonstrate our awareness of the pricing and where the pricing looked materially different or higher to enquire as to why that was the case and therefore to try to understand any differences and ensure we were getting a competitive price.

Q (Mr Hoskins, QC)

To put pressure on ABB to lower the price it was offering?

A (Mr Rose)

Yes, as one of the consequences, yes. If that could be achieved.

186.       It may be that, in the course of negotiations, BritNed offered challenging conclusions as to price, drawn from its NorNed comparisons. [234] That would be expected in the course of commercial negotiations, and perfectly legitimate as an approach to negotiation.

187.       Mr Jönsson described the effect of this use of NorNed as a comparator: [235]

“…I recall that in January 2007, TenneT was complaining about the price of ABB’s tender in relation to Lot 2 by comparing it to the price of the NorNed cable. I was particularly annoyed at BritNed’s attempts to compare the price of the NorNed cable with the BritNed cable price in the run up to our BAFO submission. The NorNed project has some similarities with the BritNed project in that both projects are 450 kV HVDC submarine interconnectors in the North Sea which connect transmission grids of two countries, one of which is the Netherlands. However, there are significant differences between the projects.”

The emails referenced below [236] provide some contemporaneous confirmation that this particular attempt at commercial pressure on the part of BritNed may not have been especially successful. Nevertheless, it was a factor.

(e)     Risk of the project not going ahead

188.       BritNed would only proceed with the Interconnector project if its two shareholders considered that the project was viable according to their criteria. That is an entirely reasonable commercial position to adopt: but it is also, by providing an upper limit as to cost, a negotiating tool. Mr Rose was clearly aware of this: [237]

“The pressure we used across the project, particularly in relation to the cable, because we knew we only had one competitor, one party for the cable, was that the project risked not actually happening at all if a price that came in was too high and damaged the business case.

So, I accept that type of pressure was used, rightly I think, to try to create a feeling of competitive pressure.”

189.       Naturally, this risk of the project not proceeding was conveyed to the tenderers. [238] ABB was aware of the risk. [239]

(f)      ABB’s lack of its own cable laying vessel

190.       At all material times, ABB lacked its own cable laying vessel. Were it to succeed in the bid, the laying of the cable would be work that ABB would have to sub-contract.

191.       In contracts of this size, an element of sub-contracting is inevitable. Mr Rose did not consider this to be a particular concern: [240]

Q (Mr Hoskins, QC)

…did BritNed think it was a concern that ABB did not have its own cable-laying vessel?

A (Mr Rose)

I don’t think it was a particular concern. We knew, it was quite clear in the market, which suppliers did and didn’t have their own vessel; and if a company doesn’t have its own vessel, then it has to subcontract. And given the nature of the contracting arrangements, we would expect them to take the risk of that and to manage that.

Q (Mr Hoskins, QC)

Do you believe it put ABB at a disadvantage against companies who did have their own cable-laying vessels?

A (Mr Rose)

I don’t think we viewed it like that. We viewed it that they could take the sub-contracting route, protect us, as BritNed, commercially from any risk that that might introduce. I do accept that that presents a management issue for that model. But no, we were pretty agnostic about that arrangement.

192.       BritNed unsurprisingly identified ABB’s cable laying capability as an issue. An evaluation of the various expressions of interest that BritNed had received – circulated within BritNed on 1 February 2006 – stated:

“ABB has no ship or strategic relationship for the provision of a ship. However, the cable manufacturing capacity is smaller than cable laying capacity. Has to be taken care of during the tender phase.”

193.       A cable laying capacity was one of the factors specifically listed in a table evaluating the various expressions of interest.

194.       Mr Rose was pressed on whether this was an area of “concern” for BritNed, and eventually agreed that it was. [241] I consider that “concern” actually puts matters too highly. The point would only become one of concern if the tenderer in question could not adequately reassure BritNed that the issue could be solved. In short, it was something that, as Mr Rose said, needed to be bottomed-out during the tender process.

195.       It is, therefore, likely that the matter would have been raised with ABB. Mr Rose was asked whether ABB would have felt pressure because it lacked a cable-laying vessel. He felt unable to speculate as to ABB’s internal position on this. [242]

196.       Mr Jönsson said that ABB considered this a weakness: [243]

“…ABB did not have its own cable laying vessel and I remember that BritNed had expressly said that not having a vessel had weakened our bid. BritNed had made it clear that it wanted to award the contract to a supplier who could complete the entire works as a turnkey project. Without a vessel of its own, ABB felt at the mercy of others, particularly as our position, being in need of sub-suppliers to complete what was known to be a turnkey project, was apparent to the market.”

(g)     Perception that ABB needed the work

197.       The extent to which ABB was hungry for work was a matter which was, essentially, a matter of ABB’s understanding. It appears that in the course of negotiations, BritNed got the impression that ABB was keen for more work, and obviously factored that into negotiations. [244]

198.       The “loading” of a factory – fully loaded, meaning that the factory has no capacity for a particular project – is a complex matter. This is because the process of manufacturing cables is not a single process, but a complex mixture of different processes. Whether a factory can take on an additional cable job will depend upon a complex mixture of timings (when a particular process must be undertaken), processes (which particular processes are necessary for the job) and client demands (which may be capable of variation during negotiations). [245]

199.       So far as the BritNed project was concerned, ABB regarded BritNed as a “priority project” as regards both the Cable element and the Converter element. This was because “there was a lot of spare capacity in both the cables and the converters factories”. [246] In his witness statement, Mr Larsson-Hoffstein said this: [247]

“When subsequently submitting ABB’s prices on 3 April 2006 to BritNed, I stated to TenneT that we were keen to start production of the BritNed cable as early as possible. This was because ABB did not have any other MI cables in the pipeline and we wanted to fill the factory’s capacity.”

200.       I find that ABB regarded the BritNed Interconnector as important work, that it was keen to secure. Of course, I do not leave out of account the fact that the Interconnector project had been “allocated” to ABB pursuant to the Cartel.

(4)     The advantage ABB derived from its participation in the Cartel

(a)     Introduction

201.       The advantage derived by ABB from the Cartel would have been in relation to the extent and strength of the competition it faced. This has two aspects:

(1)         First , simply as a matter of objective fact, ABB faced less competition.

(2)         Secondly , and it is here that the knowledge of officers within ABB becomes relevant, ABB might know that it faced less competition. Clearly, any bidder in a tender would seek to garner intelligence about the competition it faced: ABB’s participation in the Cartel gave it an advantage in this regard.

202.       The first matter to be explored is how BritNed sought to handle and exploit the competitive pressure that an invitation to tender creates (or seeks to create). Unsurprisingly, BritNed:

(1)         Sought to negotiate with multiple bidders;

(2)         Made that clear to the bidders; and

(3)         Sought to keep confidential who the rival bidders and what the rival bids were.

203.       These, somewhat elementary, points are considered in Section G(4)(b) below. Having considered them, I then proceed to the much more difficult question of ABB’s own knowledge of its competitive position. The question is difficult because by no means everyone at ABB involved in the BritNed tender actually knew of the Cartel.

(b)     BritNed’s position

204.       At the pre-tender meeting and site visit at Maasvlakte on 19 April 2006, one of the participant suppliers asked whether BritNed would be negotiating with just one bidder. The answer was that BritNed “would be negotiating with more than one bidder and potentially on more than one lot”. [248]

205.       The Phase One Procedure Paper made clear, to those acting for BritNed, that it was important to keep the relative positions of tenderers confidential. Paragraph 5.3.1 stated that “[d]uring the Phase One Tender Evaluation, it will be extremely important to avoid giving any indications to Tenderers of the relative positions, particularly important in relation to the Phase Two Tender submissions”. The importance of maintaining such confidentiality in a competitive process is self-evident, and Mr Rose confirmed that he and his team understood the importance of this. [249]

206.       The same point was made in the Phase Two Procedure Paper and again – entirely unsurprisingly – Mr Rose confirmed the importance of this requirement. [250]

207.       I am very confident that BritNed would have maintained confidentiality on its side. As I have noted, I was impressed with both Mr Rose and Mr Jackson, and I do not consider that information would inappropriately have “leaked” to tenderers.

208.       Of course, tenderers would seek information about competitors and the Cartel – as I have described – involved exchanges of information between cartelists. The question arises as to what ABB would have known about this , the BritNed tender.

209.       Mr Jackson was extremely sceptical of ABB’s denials that they knew they were the only bidder for Lot 2 or Lot 3. [251] In cross-examination, he said: [252]

Q (Mr Hoskins, QC)

…I just want to sort of test the temperature of what you are saying here. At [Jackson 1/para. 24], you say:

“Consequently, I do not think it is credible, and nor do I believe, that ABB’s representatives, including Peter Leupp, Hans-Åke Jönsson and Magnus Larsson-Hoffstein would have harboured any belief that Nexans and/or Prysmian were competing with them for Lot 2 or for Lot 3.”

Just to make it clear, are you suggesting that they are lying, or is it something more nuanced than that?

A (Mr Jackson)

I think it is slightly more nuanced. I always hesitate to call somebody a liar. But I don’t believe, given the very thin market in the cable world, where everybody knows everybody else – and, for instance, everybody goes to the same insurance people to obtain insurance quotes on such jobs, people go to the same metal providers. It is a very, very thin world and I do not believe that, even notwithstanding the Cartel arrangements not being in place, that ABB did not know they were the only supplier.

210.       I take Mr Jackson’s point regarding informal exchange of information in the market and market gossip. I also take the point that ABB was part of the Cartel. Clearly, these are both factors that would or could affect ABB’s knowledge. Beyond that, however, it seems to me that Mr Jackson’s subjective belief as to ABB’s knowledge does not take me any further as regards that knowledge. [253] It is a matter for me to determine. It is this aspect that I now proceed to consider.

(c)     Approach to analysing ABB’s knowledge advantage

211.       It is not ABB’s understanding of the Cartel in general that matters, but the extent to which ABB’s participation in the Cartel gave it an appreciation of the competitive position in which it stood in relation to the BritNed project .

212.       There are two general points that arise out of this.

213.       First, I appreciate that there is a distinction to be drawn between “ordinary” market gossip regarding the BritNed Interconnector tender – i.e. communications which might have occurred even if there was no Cartel – and information acquired through the operation of the Cartel. In my judgment, to the extent that there were valuable communications between competitors for the BritNed project, these would have been due to the Cartel, and I so find. In the ordinary course, whilst genuine competitors may exchange market “gossip”, they will be acutely aware of the need to preserve their own competitive advantage by maintaining confidence in “valuable” information. [254] The converse is more likely to be true as regards members of the Cartel. I consider that the appropriate approach to take is to presume that if valuable information was exchanged between “competitors” (e.g. as to participation in a bid or the level of pricing), that exchange occurred by reason of the Cartel, and would not have occurred but for the Cartel.

214.       Secondly, it is important to note that as regards ABB’s bid for the BritNed Interconnector it would be an error to regard ABB as monolythic. Different persons associated with ABB’s bid for the Interconnector (and, specifically, the Cable element) had different knowledge in relation to the Cartel and in relation to ABB’s competitive position regarding the BritNed tender. It will be necessary to consider three distinct questions:

(1)         Who knew about the Cartel and who did not know about the Cartel? As I have noted, some of the people very directly associated with the compiling of ABB’s bid did not know of the Cartel. On the other hand, others – notably Mr Jönsson – did know. Mr O’Donoghue, QC, for BritNed, devoted some time to ascertaining how many people, and exactly who within ABB, knew of the Cartel in general terms. That, as it seems to me, is the wrong question. I am not concerned with knowledge, in the abstract, of the Cartel, but only with knowledge insofar as it affected ABB’s bid for the BritNed Interconnector.

(2)         What these persons actually knew regarding the BritNed bid? Once again, it is necessary to stress that general knowledge of the Cartel is nothing to the point. What matters is what those persons, cognisant of the Cartel, would have known regarding competition in relation to the BritNed bid.

(3)         Whether, and if so how, any “useful” knowledge regarding (the absence of) competition could have been deployed? Two very significant participants in the framing of the BritNed bid – Mr Leupp and Mr Larsson-Hoffstein – did not know about the Cartel or ABB’s participation in it. I found them to be honest witnesses, and I therefore do not accept that they would consciously have caused ABB to put forward an uncompetitive price. [255] That leaves three possibilities:

(a)            First , those who knew of the Cartel were able to circumvent those innocent of that knowledge so as to cause ABB’s bid price for the Cable element of the BritNed Interconnector to be higher than it otherwise would have been. I shall refer to this as “direct influence” on the ABB bid.

(b)           Secondly , those who knew of the Cartel were able so to influence those innocent of that knowledge (or the processes that such persons were involved in) so as to cause ABB’s bid price for the Cable element to be higher than it otherwise would have been. Such influence might have been exercised consciously or unconsciously by those knowing of the Cartel. It does not seem to me to be profitable to distinguish between states of mind. It is the effect on the bid that matters. I shall refer to this as “indirect influence” on the ABB bid.

(c)            Thirdly , although there were persons within ABB who knew of the Cartel and who were involved in the tender for the Interconnector, they were unable either themselves to inflate the price put forward nor influence others to do so. In short, there was neither direct nor indirect influence.

215.       This, third, possibility, does not however exclude the possibility of an overcharge arising in relation to the Cable element. An overcharge could, in theory, arise in two ways:

(1)         By way of baked-in inefficiency. It could, for example, be the case that ABB was an inefficient producer of cables and therefore tendered a higher (non-competitive) price for the Cable element which ABB actually considered to be competitive. The effect of the Cartel would be to cause ABB’s price to be accepted because of an absence of competition from other, more efficient, suppliers. Such inflation of price arises out of the natural inefficiency of cartels, whereby an uncompetitive supplier receives business it would otherwise not receive simply because of the absence of competition caused by the cartel. Such inefficiencies are structural within the business of the cartelist, who may not even be aware of such inefficiencies. During the trial, inefficiencies of this sort were referred to as “baked-in inefficiencies”, and that is a term that I use in this Judgment.

(2)         By way of cartel savings. The absence of or reduction in competition is, of course, a disbenefit to consumers, in that it may result in overcharges. One benefit to cartelists is the saving that they may incur as a result of not having to compete. To a supplier, competition is expensive, because it means incurring the costs of engaging with competing suppliers, with no assurance that a firm order will be placed. The advantage of a cartel is that such costs may be avoided or reduced. I shall refer to such savings as “cartel savings”. 

216.       I consider the three questions articulated in paragraph 214 above in the following Sections:

(1)         Section G(4)(d) considers who , within ABB, relevantly knew of the Cartel. By “relevantly”, I mean those persons who were involved in the tender for the BritNed Interconnector.

(2)         Section G(4)(e) considers what these persons knew.

(3)         Section G(4)(f) considers the extent to which such persons could (directly or indirectly) influence the bid price for the Cable element of the BritNed Interconnector.

I do not consider the question of baked-in inefficiencies or cartel savings until later on in this Judgment, as such inefficiencies do not arise as a result of direct or indirect human intervention in the tender negotiation process, although they may nevertheless have an effect on price.

217.       The difficulty underlying these questions of direct or indirect influence over the tender is illustrated by some of ABB’s internal documents, which suggest a perception, on the part of ABB, of a properly competitive process. Thus, an email dated 21 December 2006 from Mr Pääjärvi (who, as will be seen, I assume knew of the Cartel, but without making a finding to this effect) to various persons within ABB circulated a “brief update after our latest meeting on Cable installation and protection yesterday at Schiphol”. The final paragraph stated:

“I am concerned about the process. We are helping BritNed to develop the Contract with great efforts and costs and still we will be squeezed in the BAFO. It may well be that some of our competitors will be awarded the Contract in the end.”

This document was put to Mr Rose, who (unsurprisingly) was unable to say much about what was, after all, an internal ABB document. [256] He accepted that the email generally referred both to the Cable and Converter elements of the bid, but suggested that this comment made more sense in the context of the Converter element (where ABB did have competition) than the Cable element (where ABB did not have competition). [257] That, however, begs the very question of what ABB’s negotiating team actually knew. That is a point on which I will have to reach a view: for present purposes, I would only say that this email suggests a fear of competition – in relation to all aspects of the contract – on the part of ABB.

218.       Another ABB internal document is an email (sender and recipients are immaterial) dated 30 March 2007 stating:

“As we said on the phone yesterday, Staffan, it feels terrible that we will probably lose yet another HVDC project. We now need to do everything we can do to win BritNed, if this still is possible. My ambition for getting as competitive a price as possible from us is to convince the bosses in Zürich etc. that we have to come in with a reasonable margin and risk provision. However, to succeed, I need your help:

Staffan: When you have any more information about the Great Belt straight, such as who the low bidder is and what the difference is in price or greater differences on the technical solution for the Tfo’s, then I will gladly receive this.

Staffan/Bo: I’d be pleased to hear any news about BritNed as I will try to work on the bosses from a number of angles. It is highly unlikely that we will have time to change our business model on this. We then will have to ensure that we work closely together in terms of our updated quotes for BritNed. As soon as I have a technical quote from our estimators, we can talk about this.”

It was put to Mr Rose that this document evidenced real commercial pressure on ABB. Mr Rose’s response was as follows: [258]

“The pressure we used across the project, particularly in relation to the cable, because we knew we only had one competitor, one party for the cable, was that the project risked not actually happening at all if a price that came in was too high and damaged the business case.

So I accept that type of pressure was used, rightly I think, to try to create a feeling of competitive pressure.”

Mr Rose, of course, knew who had bid and who had not. He knew that the field was thin, and that as a result the commercial pressures open to him were limited. He deployed them to the extent he could. But, of course, he could not speak to ABB’s knowledge. Like the previous communication – this email suggests a concern on the part of ABB that the project could go ahead, but with ABB being unsuccessful in its tender.

219.       I reference these emails not because I am persuaded that within ABB everything operated on a competitive footing. I reference them because it is important that, when considering the facts, I keep in mind the potentially insidious nature of cartels. Cartels do not advertise themselves, and their influence can be extremely difficult to discern.

(d)     Who knew about the Cartel and who did not know about the Cartel?

220.       BritNed accepted that a number of ABB employees – notably Mr Leupp and Mr Larsson-Hoffstein – had no knowledge of the Cartel. No case was advanced that these persons had used the existence of the Cartel to inflate the bid price.

221.       If the bid for the Interconnector was inflated, then it was with the innocent involvement of such persons. In short, these persons considered that they were doing their duty – i.e. pricing competitively – and were effectively duped into producing a non-competitive price.

222.       It is necessary to begin by identifying those aware of the Cartel and aware of the impairment of the competitive process as regards the BritNed tender. Clearly, Mr Jönsson was aware of the Cartel and aware of the fact that the Cartel might impair the competitive process as regards the BritNed tender. It was also alleged that Mr Pääjärvi had this knowledge. The evidence in this regard was much more tenuous, and I am a little reluctant to make a finding of this nature without having heard from Mr Pääjärvi. I shall proceed on the basis of an assumption that Mr Pääjärvi was aware of the Cartel and aware of the fact that the Cartel might impair the competitive process as regards the BritNed tender. [259] For the reasons that I give in paragraph 225 below, I do not consider that proceeding by way of an assumption, in this way, actually prejudices ABB.

223.       I regard the knowledge of these two persons as relevant because their role in ABB can be related to ABB’s bid. Mr Jönsson was in charge of the Cable element of the bid, although he delegated to Mr Larsson-Hoffstein. [260] Mr Pääjärvi was in charge of the Converter element of the bid. [261]

224.       Clearly, each might well have been in a position to influence the nature of ABB’s bid, although I accept that in relation to the Cable element, Mr Pääjärvi’s influence would have been somewhat attenuated. Nevertheless, given Mr Jönsson’s evidence that Mr Pääjärvi had a broader role, [262] it is necessary to bear him in mind.

225.       Unfortunately, it is very difficult to get any sense of how the various individuals worked together during the course of the BritNed tender. I heard no evidence from Mr Pääjärvi, and neither Mr Jönsson nor Mr Larsson-Hoffstein was able to paint a picture of how the ABB tender team really worked. It follows that I cannot actually trace – with any degree of specificity – the sort of influence that Mr Pääjärvi might have had. The same is all the more true of the other persons said by BritNed to have known of the Cartel. Their role in relation to the BritNed tender was exiguous. It seems to me that I should proceed on the basis that Mr Pääjärvi had the same knowledge and the same ability to influence matters as Mr Jönsson, simply as a safeguard against a tendency to regard Mr Jönsson as a “lone wolf”, acting on his own within an otherwise “innocent” ABB. That was very much not the case, as BritNed emphasised. The fact is that knowledge of the Cartel was embedded within ABB, and it is important not to lose sight of this fact. At the same time, a number of persons having significant influence over the BritNed tender did not know of the Cartel. In these circumstances, I shall proceed to analyse what Mr Jönsson knew, without mentioning Mr Pääjärvi further. However, the assumption that I am making regarding Mr Pääjärvi’s knowledge and influence underlies my analysis throughout for the reason that I have given.

(e)     What these persons knew regarding the BritNed bid?

226.       I consider that Mr Jönsson would have known the following things.

(i)      The fact that projects were allocated by the Cartel and that the BritNed Interconnector was allocated to ABB

227.       As has been described, the cartelists kept track of allocations of work and monitored who got what. [263] I accept, of course, that sometimes the allocation of work was not followed, with resultant sanctions, [264] but that, in my judgment, cannot affect the broad expectation that I find all cartelists, including ABB, to have had: namely that when work was allocated to a particular undertaking, the other cartelists would behave in a manner that would assist, if not ensure, that that undertaking received the work.

228.       The fact that ABB was not a “core” cartelist, and may not itself have kept track of allocations, is nothing to the point. [265] The point is not whether ABB kept track but whether there was an expectation that if ABB was allocated a project, the other cartelists would “play ball” and do their best to see that ABB got the work. I find that there was such an expectation, based upon the Decision and Mr Jönsson’s general acceptance that the thrust of the Decision was right. [266]

229.       In fact, Mr Jönsson’s understanding of the competitive situation is likely to have more sophisticated and more specific than this.

(ii)     Likely absence of competition from Japan/Korea

230.       I consider that Mr Jönsson would have appreciated that there would be an absence of competition – at least so far as the Cable element of the bid was concerned – from Japanese and Korean suppliers. I reach this conclusion based upon the fact that such a geographic allocation of work was the starting point for the Cartel. [267]

231.       Mr Jönsson sought to suggest, in his witness statement, that he regarded the threat of competition from Japanese cable suppliers as a real one: [268]

“…I could not be confident that the Japanese cables suppliers would not compete for the cables contract since cartel discussions had broke off with Japanese companies in 2004. Any interactions I had with Japanese suppliers after that date…were not in the context of the Cartel but in the normal course of business. However, although they were developing technologies that could potentially be used for projects such as the BritNed project, the Japanese suppliers at the time lacked the technical expertise in certain areas to execute the BritNed project in its entirety…and were subject to significant transport costs.”

232.       The thrust of Mr Jönsson’s evidence appears to be that the Japanese (and, no doubt, Korean) cable suppliers were not competitive rather than that they chose not to compete. I do not accept this evidence. It may be that these suppliers were uncompetitive: but that was not tested in the European market. I find that, by reason of the Cartel and as described in the Decision, the Japanese and Korean suppliers chose not to compete, and that Mr Jönsson would have appreciated this effect of the Cartel. [269]

233.       I conclude that Mr Jönsson would not have seriously been concerned about competition from Japanese or Korean suppliers. I do not go so far as to say that such competition was impossible : it simply would not have been expected and would have been discounted by those in the know.

(iii)    Level of competition from European suppliers as regards cable projects

234.       As regards ABB’s European competitors, for cables , ABB’s main competitors were Nexans and Prysmian. [270] Both of these companies were cartelists.

235.       The contacts that Mr Jönsson had with Nexans have already been described, [271] as has Nexans’ submission of a non-compliant tender. [272] In these circumstances, the inference that Nexans did not compliantly tender because of the Cartel must be a strong one and I have found that this was indeed the case.

236.       Nevertheless, Mr Jönsson sought to suggest that – because ABB had not made provision in its tender for a sub-contraction to Nexans, and because Nexans has “reproached” ABB about this [273] – “[i]t could not…be ruled out that Nexans…would compete hard for the project”. [274]

237.       I do not accept this evidence. In cross-examination, Mr Jönsson accepted that the BritNed project had been allocated to ABB by Nexans. [275] The relevant passage in cross-examination was as follows: [276]

Q (Mr O’Donoghue, QC)

Mr Jönsson, come on. You went to Zurich in May 2006 with the specific purpose of allocating the BritNed project. Nexans told you that they would not meet the customer’s delivery requirements for the project?

A (Mr Jönsson)

Correct.

Q (Mr O’Donoghue, QC)

You agreed that with Nexans?

A (Mr Jönsson)

Correct.

238.       As regards the possibility of Nexans reneging on this, because of ABB’s failure in its bid to make provision for a sub-contract to Nexans, Mr Jönsson said this: [277]

Q (Mr O’Donoghue, QC)

You had at least five separate contacts with Nexans in the period following [the allocation agreement] and at no stage did they indicate to you that they would bid for BritNed or that they were irritated with you for not subcontracting.

You then find out in January 2007 that they are fully loaded and do not have capacity. Are you seriously expecting the court to believe that you considered Nexans a credible competitor for the BritNed project?

A (Mr Jönsson)

Well, possible contract. I could not rule anything out, because you should also remember at that point in time, when it comes to 2007, Nexans had procured and aligned themselves and had factory resources in Tokyo, in Japan, so – and that plant could at least do oil-filled cables, which has certain similarities to – even if they are not the same. So how Nexans could use that plant, I could not be sure…

…it was maybe not likely, but I could not be sure. Therefore, my statement, I stay with my statement.

239.       Clearly, it was not impossible that Nexans could tender – notwithstanding the Cartel allocation agreement. The question is how likely ABB would regard that eventuality. Mr Jönsson was pressed on this in cross-examination: [278]

Q (Mr O’Donoghue, QC)

I suggest to you, it is at least extremely unlikely, would you agree with that?

A (Mr Jönsson)

Yes, but then we are in the view of probabilities.

Q (Marcus Smith J)

Well, yes. That is what I was going to ask you, Mr Jönsson, because the question that kicked off this last dialogue was whether Nexans was considered by you to be a credible competitor for the BritNed project and I think your answer was, “Well, it was a possible competitor”?

A (Mr Jönsson)

Correct.

Q (Marcus Smith J)

So I think, right off, you have been talking about probabilities: so I want to get a feel for where, on the scale of certainty, you considered Nexans to be as a rival bidder?

A (Mr Jönsson)

I mean, to answer your question, it would be on the lower half.

Q (Marcus Smith J)

Right. I will press you a little on that. Let’s talk percentages. 100% is they will definitely bid, and 0% is they definitely won’t. Where are we, on that sort of scale, in your subjective view?

A (Mr Jönsson)

I have never thought about it from that perspective, of percentage points…as you put the question to be now…10-15%.

240.       There is, of course, a spurious and artificial precision to percentages. But I consider Mr Jönsson’s answer to be a helpful one as regards his own views regarding the likelihood of a truly competitive Nexans bid: not impossible, but quite unlikely.

241.       As regards Prysmian, Mr Jönsson’s evidence was again that he could not exclude them as a potential bidder, particularly in conjunction with a collaborative bid with Siemens for Lot 3, with Siemens tendering for Lot 1 and Prysmian for Lot 2. [279] Mr Jönsson’s evidence was, essentially, that he was cautious when assessing potential competition, and took nothing for granted. He did not accept that Prysmian could be classified as an “unrealistic” competitor, [280] but he did not put the chance of a real bid at much higher than this: [281]

A (Mr Jönsson)

…Therefore, I am always very cautious for taking anything absolutely for granted, and we have always looked on the possibilities that there would be competition and that we needed to act accordingly. So that is my general policy, that goes 20 years back, and has nothing to do with the Cartel, but it is how I look upon it, I never take anything for granted and take it from the point of view that there would be ways for competitors to act.

Mr Jönsson showed a no doubt laudable conservatism in his assessment of the risk of rival bids. However, I do not accept that his assessment was “nothing to do with the Cartel”. What Mr Jönsson would have been doing is assessing, given the existence of the Cartel , the risks of genuine competition from another cartelist. As with Nexans, I find that this was a risk, just not a very high one: [282]

Q (Mr O’Donoghue, QC)

Now, you accepted, in relation to Nexans at least, some of the tender team were made aware of the very low probability of them competing. [283] In relation to Prysmian, I would suggest that the tender team were also aware that there was a very low possibility of Prysmian competing. That is what [Mr] Agnevall is saying, isn’t it?

A (Mr Jönsson)

Yes, I mean after – with SAIPEI, but the probability is not zero.

242.       As with Nexans, I find that Mr Jönsson’s view was that Prysmian was unlikely to constitute real competition in relation to the Cables element of the tender.

(iv)    Level of competition as regards the Converter element and Lot 3

243.       I heard much less evidence in relation to the Converter element of the tender. Mr Pääjärvi did not give evidence, and the parties (quite rightly, given the nature of BritNed’s Overcharge Claim) focussed on the Cable (Lot 2) element of the tender, because it was in relation to Lot 2 that the question of the overcharge arose.

244.       In theory, competition in relation to the Converter element could only affect prices and competition in relation to Lot 1 and Lot 3. Lot 2, being purely concerned with the Cable element, ought to be insulated. However, as demonstrated by ABB’s own conduct in relation to Lot 2, it is not possible to regard the Cable and Converter elements as so isolated.

245.       The relevance of Lot 1 to potential competition in relation to Lot 2 (and so Lot 3) is not that BritNed overpaid in relation to Lot 1. The evidence before me was such, however, that I can only proceed on the basis that the Cartel did not affect the level of the Siemens bid in relation to Lot 1, even though, had there been greater competition, it is likely that BritNed would have been able to place further legitimate competitive pressure on Siemens.

246.       The importance of Lot 1 lies in – as I have described [284] – the commercial pressures it exerted on ABB. I consider that whilst ABB was assured – as it was with the Cable element of the bid – that it was the Cartel’s “choice” for this element, unlike with the Cables element of the bid, where Nexans and Prysmian represented the competition, [285] competition in relation to the Converter element might emanate from non-cartelists. That threat of competition was made good by Siemens’ participation in the process. I consider that Mr Jönsson would have appreciated that the supply-side for the Converter element of the project was wider than that for the Cable element.

(v)     Conclusion

247.       In terms of the knowledge that Mr Jönsson (and, inferentially, Mr Pääjärvi) had, I conclude that they appreciated that ABB would have a “clear run” in relation to Lot 2. By this, I mean that they appreciated (because of the fact that effective competition for Lot 2 only emanated from cartelists) that ABB would not face real competition from rival tenderers and that any tenders actually submitted would not be genuinely competitive.

248.       That was not quite the case in relation to Lot 1. Here – although Mr Jönsson and Mr Pääjärvi could be assured that ABB was the choice of the Cartel – there was potential for competition from suppliers outside the Cartel. I consider that Mr Jönsson and Mr Pääjärvi would have been aware of this risk of competition at the time the tender process commenced, and their view would have been confirmed by the participation of Siemens in the tender process.

(f)      Whether, and if so how, any “useful” knowledge regarding competition could have been deployed within ABB

(i)      Introduction

249.       I have found that the actual knowledge within ABB of the Cartel was limited. Some persons involved in formulating the ABB bid for the cable part of the contract did not know of the Cartel, whilst others did .

250.       The question arises as to whether – and if so how – those aware of the Cartel transmitted their knowledge acquired through the Cartel within ABB so as to derive an advantage. One anterior question is this: was it necessary to transmit this information at all? BritNed contended that a cartelist would not participate in a cartel without wanting to benefit. I accept that: but the benefit may be the obtaining of work properly tendered for , as opposed to the obtaining of work at an inflated tender price . Given that the Cartel originated as a response to under-capacity, the perceived benefits of the Cartel may simply have been the allocation of work, thereby ensuring work flows were directed as the Cartel wished and minimising costs thrown away through unsuccessful tenders.

251.       It seems to me, therefore, that I cannot presume that it was a necessary part of the Cartel to inflate tender prices by directly or indirectly influencing the bid. (As I noted earlier, the questions of baked-in inefficiency and cartel savings I leave for later.)

(ii)     Direct influence on the level of the bid

ABB’s process and the distinction between direct costs and common costs

252.       The process by which ABB’s tender was compiled was described in general terms in paragraphs 125 to 136 above. The question I consider here is the extent to which it would be possible directly to influence – in an upward direction – the level of ABB’s bid.

253.       For the purposes of analysis, it is necessary to differentiate between those costs which ABB considered to be directly attributable to the supply of a specific project (or part of a specific project – like the Cable element of the BritNed Interconnector) and all other costs incurred by ABB in the course of its business. This, as will be seen, represents a distinction drawn by Mr Biro as part of his analysis. [286] I shall refer to the former type of costs as “direct costs” and the second type as “common costs”. [287] It is clear from Mr Larsson-Hoffstein’s evidence regarding the production of costings during the “technical” stage of ABB’s tender process that this distinction between direct costs and common costs is one that informed ABB’s processes. [288]

254.       I consider the question of the direct influence on the level of ABB’s bid for the Cable element of the BritNed Interconnector in the following way:

(1)         First , I consider the integrity of Mr Larsson-Hoffstein’s assessment of direct costs, and the extent to which these could be inflated.

(2)         Secondly , I consider the extent to which common costs could be inflated. That involves consideration of, in particular, the margin that was allocated in relation to the common costs of ABB – referred to as the sales, general and administrative (or “SGA”) costs. These costs warrant greater examination, because they are inherently more subjective and less closely tied to the specific costs of a specific project.

I stress that I am simply seeking to gauge the extent to which ABB’s internal processes could be affected by the Cartel. It is a further question – to which I will come – whether any inflation of price could survive negotiation with BritNed.

Influencing the level of direct costs

255.       Mr Jönsson denied seeking to influence the amount of the BritNed bid. Mr Jönsson’s evidence was that Mr Larsson-Hoffstein’s process for costing would not be affected by the Cartel: [289]

Q (Mr O’Donoghue, QC)

[In Jönsson 1/para. 18] you say that “…Larsson-Hoffstein was experienced in pricing large submarine power cables projects, as he had been involved in the pricing of Gjøa and Estlink, two other large submarine cables projects.”

Now, first question: both of these projects were cartelised, weren’t they?

A (Mr Jönsson)

Gjøa, to the extent that I got that information that Nexans was loaded or that they would not bid, as has been informed. This is correct.

Q (Mr O’Donoghue, QC)

So, Larsson-Hoffstein, if he was basing his pricing for BritNed on these two projects, or his experience in these two projects, would have been pricing in the Cartel effect, wouldn’t he?

A (Mr Jönsson)

No. Let’s be clear, from what you have been alluding to, we have discussed as well, the pricing structure, the whole way that things were done with material costs, production times, machine hours, engineering costs, the various components that you also have seen – I think it has been given to the court in fairly great detail – was never tampered with. That was running and that was even audited generally.

So when Mr Larsson-Hoffstein was pricing the cable projects, he would run it perfectly, through those computerised calculation points, and they would come out what they were, and there were no changes to that…

…That goes for Gjøa, it goes for Estlink and all the others. This was something important also for me, we never changed those things…

256.       Unsurprisingly, given my view of Mr Jönsson’s evidence and – most importantly, the fact that he admitted influencing other bids (so that ABB lost these) – I treat this denial with a degree of scepticism. In other cases, Mr Jönsson admitted he had been able to affect ABB’s prices in specific bids for reasons that were not competition, but Cartel, related. Mr Jönsson admitted he caused ABB’s prices to be adjusted upwards , so as to ensure that ABB did not get work: [290]

“15.   From March 2004, when I became head of the Karlskrona factory, the individuals from the factory who were aware of the power cables cartel and who helped me implement cartel arrangements were Mr Lars G. Carlsson (Project Manager/Marketing Manager) and Mr Håkan Agnevall (Marketing & Sales Manager for AC cables from September 2003 to January 2007).

16.     Mr Carlsson was aware of the cartel before I became head of the Karlskrona factory. I do not recall how or when Mr Agnevall became aware of the cartel.

17.     Even where I was involved in, or aware of, allocation discussions in respect of certain power cable projects, I did not always need to provide instructions to other individuals within ABB to ensure the cartel allocation was adhered to. This could be, for example, because ABB did not in any event have the technical capability for the project concerned or was not invited to bid.

18.     When I did instruct other individuals to help implement the cartel arrangements, I nearly always did this in person, rather than by email.

19.     I gave instructions to Mr Carlsson and Mr Agnevall on a small number of occasions as they were involved in the preparation of tenders for large projects. If I was concerned that ABB would submit a bid for a project that risked disrupting the cartel allocation, I would advise Mr Carlsson or Mr Agnevall to adjust the tender response in some way, for example by altering the commercial terms, that would have the effect of making it less attractive to the customer. For the avoidance of doubt, I did not give any instructions about the BritNed project in connection with the implementation of the cartel.”

Mr Jönsson was able, when necessary, to ensure that ABB’s bids on projects not allocated to it by the Cartel were “cover bids” not intended to win the tender. [291]

257.       The question is whether the converse also pertained: was ABB able to inflate bid prices in order to increase its profit in relation to projects that were allocated to it? It was Mr Jönsson’s evidence that there was no explicit increase of ABB’s prices of the sort described by him in relation to tenders ABB was concerned to lose. [292] In other words, the benefit ABB derived from the Cartel was not business at a deliberately inflated price but simply business that it would not otherwise have got .

258.       As I have said, I treat Mr Jönsson’s evidence on this point with a degree of scepticism. Mr Jönsson’s evidence was, however, confirmed by Mr Larsson-Hoffstein, at least so far as the BritNed Interconnector was concerned. Mr Larsson-Hoffstein’s evidence was as follows:

(1)         His costings process was “self-standing”. It was not based on work done in relation to previous tenders or bids. The fact, therefore, that Mr Larsson-Hoffstein had been involved in pricing cartelised bids, which ABB was keen not to succeed on, would not affect the direct costs incorporated into the BritNed bid. As regards the compilation of the costings, Mr Larsson-Hoffstein stated that “[t]his did not involve starting with the pricing of any other project, but instead involved looking at the cost of each element of the project (e.g. conductors, insulation, and raw material prices) and putting together a costing on that basis. [293] I accept this evidence, so far as direct costs are concerned.

(2)         Mr Larsson-Hoffstein’s process was not altered by Mr Jönsson (or, for that matter, anyone else). Mr Larsson-Hoffstein said: [294]

“I ran my proposed price past my managers within ABB at this time, including Hans-Åke Jönsson, but no alterations were made to those figures. Similarly, no adjustments were made to my figures as a result of the risk review process for the project.”

(3)         Mr Jönsson’s general instruction to Mr Larsson-Hoffstein was to keep the bid price down: [295]

“From the very outset of pricing the BritNed project, the only instruction I was given from the management in the ABB Karlskrona factory (including Hans-Åke Jönsson) was to push the price and costs for the project down as far as they could go.”

Again, so far as direct costs are concerned, I accept this evidence.

(4)         This conclusion is borne out by an analysis of the PPMs that were produced by ABB in relation to the BritNed project over time, up to the first tender submission: [296]

Date

Event

Lot

PPM bid price

9 or 10 May 2006

Meeting between Mr. Jönsson and Mr. Romand. [297]

 

 

15 May 2006

“PPM 1” [298]

Lot 3

€310,056,062

18 May 2006

“PPM 2”

Lot 3

€310,056,062

18 May 2006

“PPM 3”

Lot 3

€310,056,062

20 May 2006

Business Unit risk review [299]

 

 

23 May 2006

Division risk review [300]

 

 

30 May 2006

“PPM 4”

Lot 3

€276,072,970

6 Jun 2006

“PPM 5”

Lot 3

€272,772,255

15 Jun 2006

“PPM 6”

Lot 3

€283,184,087

16 Jun 2006

“PPM 7”

Lot 2

€188,808,653

16 Jun 2006

“PPM 8”

Lot 3

€274,000,000

16 Jun 2006

ABB submits its Phase 1 bid [301]

 

 

Table 2 : ABB bid prices as evidenced by PPMs.

On the whole, as can be seen, the trend is downward. Significantly, to my mind, the risk reviews resulted in a decrease in the bid price, in particular after the risk reviews in May 2006. Although Mr Larsson-Hoffstein very fairly accepted that it was “possible” for those knowing of the Cartel to cause increases in the bid price at the risk review stage, [302] there is no suggestion that this in fact occurred. I reject the contention that the risk reviews were used to increase the level of the bid price for the BritNed project as being inconsistent with the evidence before me.

259.       I consider that the direct costs recorded in the PPMs represent an honestly and competently compiled statement of those costs, and that they were not inflated by the direct influence of Mr Jönsson or (for that matter) anyone else within ABB. I reach this conclusion essentially because I considered Mr Larsson-Hoffstein to be not only a transparently honest witness, but also an extremely competent compiler of the costs of the Cable element of the ABB tender. I do not consider that he would have allowed that process to be distorted and that if he had been required to include within the Cable element of the tender a cost that he did not consider to be justified, he would have told the court. Mr Larsson-Hoffstein was not cross-examined, in any detail, on exactly how he had compiled the BritNed tender. The bulk of his evidence was given in response to questions from me, set out in paragraph 132 above.

260.       I find that Mr Larsson-Hoffstein’s pricing of the direct costs of the BritNed Interconnector Cable bid were unaffected by the Cartel. They were properly calculated, competitive, costs. I also consider that any margin added by Mr Larsson-Hoffstein to represent risks relating to the project specifically (i.e. to the direct costs being underestimates or to contingencies regarding direct costs) were properly added and were not inflated.

261.       To an extent, I am fortified in this conclusion by the fact that Mr Biro reviewed the costs set out in the PPMs and considered them sufficiently robust to base his analysis on them. Again, I am satisfied that had he identified any anomalies, he would have drawn these to the court’s attention. However, I should say that I place only limited weight on the endorsement of Mr Biro. That is because Mr Biro gave evidence as an expert economist and did not hold himself out as having any expertise in relation to submarine cable projects like the Interconnector nor, indeed, any expertise in cable projects generally. Thus, whilst Mr Biro would no doubt be adept at spotting accounting anomalies in the PPMs, I doubt very much whether he would have been able to spot the illicit inflation of a direct cost for dubious and not well-founded technical reasons. 

262.       The same, of course, goes for Dr Jenkins. Like Mr Biro, Dr Jenkins was a formidable economist, but that was the limit of her expertise. Dr Jenkins did not consider the PPMs sufficiently reliable and relied on them as little as possible for the purposes of her analysis. In her first report, she set out her concerns regarding ABB’s reported costs. [303] Whilst I accept that the internal workings and costings of ABB were unsurprisingly complex and no doubt difficult to follow, I do not regard the points raised by Dr Jenkins as supporting a suggestion that the direct costs in the PPMs were inflated by reason of the direct influence of those knowing of the Cartel. In cross-examination, Dr Jenkins said this: [304]

“The difficulty with the PPMs that have been supplied by ABB is that it is very difficult to do any of that checking and sense-checking – certainly for myself, with no access to ABB itself – other than through written questions. And I think in the teach-in I did point out a few of those situations where some of the cells in these PPMs just look quite unusual. They are not the sort of thing you would expect to see. Now, as you rightly say, perhaps there is a very good reason for that. But it is very hard to glean that information in this process, and there are very, very many cells in these spreadsheets. So the ability for there to be issues with those and aspects that may not be reflective of true costs is quite high in my view and that is why I was not keen to rely on them.”

263.       This was explored further in cross-examination: [305]

Q (Mr Hoskins, QC)

What, then, is the actual issue you are concerned with? Are you suggesting the figures in the PPM may simply be false figures, put in to justify an increase in price? Is that what we are actually talking about, is that your concern, that it is a deliberate fixing of the figures because of the Cartel?

A (Dr Jenkins)

I don’t know what the meaning of deliberate is, in that sense, in that does – I think the people who are putting those price schedules together may feel they are doing the right thing, they may be following guidance that they have been given.

I think one of the aspects that I pointed out in the teach-in, which was the copper cost, and subsequent to that there has been disclosure around that cell and what underlies it, and it was stated that there is good evidence of where that copper cost comes from, even when the document that is then disclosed doesn’t actually match the numbers in that PPM and that document shows that in that copper cost, which has been described as a raw material cost, there are overhead elements included inside that cost.

Now, an employee might feel that they are following the rules about how do we measure what costs of copper or other materials are, but is that actually a good reflection of in a competitive environment if they had gone out and procured copper and just said “This is the amount of copper it is”? That is the bit that is quite hard to see, and I think in the Cartel I do think there would be incentives on the business to ensure that when they present their final results, whether that is internally or potentially to the customer, that they have – it is much easier to say, “Here are all the different costs that have been incurred”, including in the case of these PPMs a lot of risk elements, and therefore those costs are affected by the fact that it is during the Cartel and they are a way in which to increase the overall price of the project.

264.       BritNed did identify two specific elements in the direct costs which, it was contended, were indicative of Cartel inflation in the case of the BritNed project. One of these related to the cost of the copper recorded in the PPMs. The other related to the cost of the construction all-risk (“CAR”) insurance used by ABB in the tender. In neither case do I consider that these costs made good BritNed’s contention that the direct costs as recorded in the PPMs were inflated:

(1)         Inflated cost of copper. As to this:

(a)            As Dr Jenkins indicated in the teach-in, the BritNed PPM recorded a cost for copper that was above the price of copper on the open market. The price in the PPM appeared to contain an (unexplained) overhead cost. Mr Larsson-Hoffstein was not asked about this. This may be because the point only surfaced during the course of the teach-in, resulting in subsequent inter-solicitor correspondence, initiated by BritNed’s solicitors, during the course of which the point was further explored.

(b)           The point was, however, put to Mr Biro (at some length) during the course of his cross-examination. [306] Mr Biro, perfectly understandably, felt unable to provide a granular explanation as to how specific numbers were put together by the ABB engineers at the time. [307] It was put to Mr Biro that the figure recorded in the PPM was 20.26% higher than the market price for that quantity of copper, [308] and that there was no explanation for this increase. The inference was that this inflation of cost was Cartel induced.

(c)            Unsurprisingly, this caused ABB’s solicitors to look into the matter rather more closely. In the end, a convincing explanation of the reason was given by Mr Hoskins, QC, during the course of oral closing submissions (Day 15/pp.46 ff ). [309] This comprehensively explained the anomaly. Since Mr Hoskins, QC had provided an explanation of a matter of fact on instructions (albeit that the documents Mr Hoskins, QC showed me were very convincing in indicating how the copper cost had been compiled), I requested that the point be evidenced by a further statement from Mr Larsson-Hoffstein, which resulted in Larsson-Hoffstein 2.

(d)           In short, there was nothing in the point. Indeed, this delving into the detail of the PPMs rather underlined their essential reliability, at least so far as direct costs were concerned.

(e)            In a note dated 24 May 2018 submitted by counsel for BritNed, certain criticisms were made of ABB as to how the evidence on this point, culminating in Larsson-Hoffstein 2, had emerged. Nothing in this note undermined or caused me to alter the conclusion expressed in paragraph 264(1)(d) above. I do not regard the criticisms of the late-emergence of this evidence as well-founded. Of course, points can emerge late at trial, and have to be dealt with. However, in this case, the point began life as an example, from Dr Jenkins, as to how difficult it was to follow certain entries in the PPMs. As the matter was delved into further, it appeared to become a matter of altogether greater moment, and BritNed chose to pursue the point. ABB were entitled to respond and it seemed to me in all the circumstances appropriate to admit Larsson-Hoffstein 2 into evidence. The fact is that the sort of point that BritNed was seeking to make through Dr Jenkins and through the cross-examination of Mr Biro really needed to be made by an expert in cable costs called by BritNed, on which Mr Larsson-Hoffstein could then have been cross-examined. There is no reason why, had an application been made sufficiently before trial, such expert evidence could not have been adduced. In short, the reason I admitted Larsson-Hoffstein 2 – and, indeed, permitted Mr Hoskins QC to explain the position in closing – is because the point was raised by BritNed so late in the day.

(2)         Inflated cost of CAR insurance . As to this:

(a)            The contract between BritNed and ABB identified certain sums in the contract price as “provisional”, which rendered the pricing of such cost items adjustable in light of their subsequent actual cost. . [310] In such cases, the price payable by BritNed would reflect the true cost, and not the provisional cost recorded in the contract and in the anterior bidding documentation.

(b)           Both the CAR insurance and the copper referred to above were provisional items in this sense. Given this fact, it is difficult to see how an overstatement of cost in relation to such items could benefit ABB. Increasing such provisional costs would only serve to make ABB’s costs look high, without any overcharge actually being received by ABB. In short, it seems to me highly doubtful that any overcharge in relation to the CAR insurance was attributable to the Cartel.

(c)            The provisional sum for the CAR insurance was 10% of the project price. [311] The actual premium was considerably less, at 4.158%. [312] This was relied upon, by BritNed, as evidence of overstatement of costs by ABB. [313] I do not accept this. The whole point of provisional sums is that they are uncertain and they are adjusted in light of the actual prices. I can see no reason why a cartelist, seeking to benefit for the Cartel, would inflate such costs. [314]

265.       In conclusion, I find that BritNed’s attacks on the reliability of the direct costs recorded in the PPMs to be misconceived.

Influencing the level of common costs

266.       Mr Larsson-Hoffstein accepted that it was possible for others involved in the process to influence margins. [315] He did not accept that this had in fact happened, and I do not consider that it did happen so far as direct costs are concerned. I consider that margins added to direct costs, including contingencies for risks relating to the Interconnector project, would properly have been added by Mr Larsson-Hoffstein.

267.       However, the potential for inflation in relation to ABB’s common costs is less easily dismissed. Such costs were described by ABB as “selling, general and administrative costs” or “SGA” costs.

268.       In her first report, Dr Jenkins analysed the SGA mark-up on production cost (which she set out in a figure) and said this: [316]

“3.35 As shown in the figure, SGA allocation is not consistent across projects or over time. The SGA mark-up is quite volatile during the Cartel period, and increases markedly for the later cartelised projects. More importantly, the SGA mark-up falls once the Cartel collapses.

3.36   The figure above provides indicative evidence that cost reporting differed during and after the Cartel period. It might be the case that the Selling, General and Administrative divisions at ABB were less efficient as a result of the Cartel, leading to inflated costs and a higher SGA mark-up…”

269.       In the Joint Statement, Dr Jenkins’ response to Issue 16 – “There is evidence that ABB’s “SGA” costs were higher during the Cartel period” – was as follows:

“There is evidence that cost reporting differed during and after the Cartel period…It might be the case that the Selling, General and Administrative divisions at ABB were less efficient as a result of the Cartel, leading to inflated costs and a higher SGA mark-up…”

270.       Dr Jenkins very fairly put the issue no higher than this: that it was possible that the SGA costs were affected by the Cartel, either deliberately, indirectly or simply because of (for example) baked-in inefficiencies or cartel savings.

271.       The point was put to Mr Jönsson in cross-examination. Mr Jönsson was quite prepared to describe in general terms how ABB would have calculated its SGA costs: [317]

Q (Mr O’Donoghue, QC)

Can you explain how ABB would have calculated its selling, general and administrative costs for a project, the SG&A?

A (Mr Jönsson)

Normally, we would assign a certain percentage point to the SG&A as part of our budget process for the overall organisation at the year-end, as part of the budget, and then we would assign a certain percentage point that would be used. That could then be adjusted in some specific cases in discussion with senior management.

Q (Mr O’Donoghue, QC)

So, just to get this straight, was it a specific calculation made in each case or did you start with the general position and then that could be adjusted in an individual case?

A (Mr Jönsson)

The latter.

Mr Jönsson was not, however, prepared to comment on Dr Jenkins’ analysis, [318] which was – as all eventually agreed – really a matter of expert evidence. [319]

272.       I come to the question of indirect influence on price, baked-in inefficiencies and cartel savings in due course. Here I consider the possibility that SGA costs were inflated by direct influence so that ABB recovered more than its costs because of the Cartel. I consider that I cannot – on the basis of the factual evidence – exclude that possibility. Indeed, I note that Mr Biro accepts the possibility, but seeks to control for it in his analysis. [320] For the present, that is my finding in relation to SGA, or common, costs: that they could have been deliberately inflated within ABB. The extent of any such possible inflation (which will involve consideration of other matters, notably the competitive relation between ABB and BritNed) is a matter this is considered further in Section I below.

(iii)    Indirect influence on the level of the bid

273.       Although it is difficult to differentiate between baked-in inefficiencies and an indirect influence on the bid level, I nevertheless consider that such a distinction is worth drawing. Indirect influence, causing a bid to inflate, was well-described by Dr Jenkins: [321]

“Indeed, that individuals who may themselves not know officially about the Cartel, but if they operate in an environment where there is actually not very much competitive pressure and therefore a process has grown up whereby you add 10% into a lot of your material costs, if you then face competition you discover you can’t continue to sustain that.

Now, if you ask that person, they say, “We are just following standard practice, and this is how we price things, there is a 10% overhead on all these costs”, it doesn’t mean it is a good measure of the counterfactual competitive costs that would have been observed absent the Cartel.”

274.       There would, in my judgment, be a number of such influences operating within ABB.

Communication of market capacities

275.       Inevitably, within ABB, perfectly innocent conversations would have gone on as to which of ABB’s rivals was bidding, for what, and at what prices. It would be very surprising if that did not occur.

276.       It would be very easy for someone in the position of Mr Jönsson, knowing of the Cartel, to feed into such conversations information derived illegitimately from his participation in the Cartel.

277.       In his evidence, Mr Jönsson was not very forthcoming about this, partly because (as I have found) he was careful not to admit matters not clearly found by the European Commission in its Decision, but also because such exchanges would not necessarily be easy to remember: [322]

Q (Mr O’Donoghue, QC)

…But are you seriously suggesting that recognising that there was zero or a low probability of Nexans and Prysmian competing, that when the tender team came to you and said, do we have competition, you sat there and said nothing? Do you really expect the court to believe that?

A (Mr Jönsson)

No, I don’t think they came to me and asked this question at all. But what I remember when we discussed was that they felt that they had competition. I didn’t question that because I didn’t have such information – as I said, I would not agree that the probability for Prysmian was zero…

I accept that it would be most unlikely for persons innocent of the Cartel within ABB to ask Mr Jönsson point blank whether they had competition. That would imply guilty knowledge on their part, which I have found not to exist. Equally, Mr Jönsson would be most unlikely to volunteer information that would show such guilty knowledge. 

278.       The conversations would have been much vaguer than this, and their effect difficult to judge. But I find that they would have taken place. One example concerns the low probability of Nexans bidding: [323]

Q (Mr O’Donoghue, QC)

But this extremely low probability of Nexans’ bidding, you obviously made that known to the people involved in the bid within ABB, right?

A (Mr Jönsson)

Exactly how I did that make known, I must say I don’t remember. There is probably some evidence here how – that I did something. I cannot remember it, sitting here.

Q (Mr O’Donoghue, QC)

Do you deny making it known?

A (Mr Jönsson)

No, the fact is that Nexans’ factory, as it was before any extension of capacities, was smaller than ours, so that was well known. So, in that sense, it was known. But to – yes, this is what I can say.

Affecting the commercial approach of ABB

279.       For reasons that I have discussed, it would not be possible for Mr Jönsson to explain to an ABB officer innocent of the Cartel that ABB’s commercial approach could be more relaxed as regards costs because of the Cartel. I do not consider that such things would have been said overtly and explicitly. But I do consider that it would be possible for someone in the position of Mr Jönsson to affect the general approach ABB might take to a bid – to affect the “mood music” as it were, that would inform ABB’s commercial approach.

280.       It was suggested that one way of ensuring that a bid was higher or “fatter” than it might be was by ABB explaining (if only internally), in advance of any bid, why that bid might be “high”. An example of this is an internal email within ABB which read as follows:

“At our last HVDC meeting, Hans-Ake and I concluded it would be good that we both pay a visit to Mel Kroon (and possibly Lex Hartman). Each time I meet Mel Kroon he makes clear to me he is worried not to get competitive prices from ABB (factory full…). We believe we should create clarity around that issue now (in neutral time and proactively). I learned from Hans-Ake that it is anyhow to be expected that due to material price increases and a more sound cable market situation vs the days that NorNed was signed (nineties) the BritNed price could be some 20% higher versus the NorNed. The day Mel would hear from this, if we do not prepare that message upfront, he will definitely conclude: ABB are crooks! We want to avoid this.”

281.       This is a communication capable of a perfectly innocent reading. ABB was concerned that its prices for BritNed might appear too high and was proposing proactive steps to ensure that the (potential) client’s expectations were managed so that, when figures were put forward, ABB was not regarded as a “crook”. Yet the obvious effect within ABB would be to convey a sense that senior management within ABB were relaxed about a high bid. [324] Inevitably, this would have an effect on how the tender team approached costs.

282.       Another instance is when ABB was called by TenneT, complaining that:

“TenneT experiences ABB as too arrogant with its pricing of the cable of

·          BritNed – 70% price increase in comparison with NorNed

·          NorNed 10km additional cable – 100% price increase in comparison with NorNed”

283.       I take TenneT’s complaints about pricing with a pinch of salt. TenneT’s statements cannot, in my judgment, be regarded as an objective assessment of the situation. They were part of the bargaining process. ABB would have appreciated this. However, the reaction within ABB will have set the agenda for the tender team. Mr Jönsson’s response, disseminated within ABB, was that the comparisons made by TenneT were unsustainable. He concluded: [325]

“I have personally reviewed the pricing in both cases and these have been reviewed all the way up to CEO level (BritNed case). A key aspect in this review has been to have prices that are defensible and correct and should not be possible to label as arrogant.”

Here, therefore, we have the man in charge of the Cable element of the bid, aware of the Cartel and of the fact that ABB has an advantage in relation to the BritNed contract because of the Cartel, telling his team that the bid is not arrogant, but realistic. Leaving entirely on one side the question of whether the tender was inflated, if it was, then it was certainly not being challenged by a sense of competitive pressure coming from within ABB.

(g)      Conclusions

284.       In terms of the effect that the Cartel had on ABB’s business practices so far as the Cable element of the BritNed Interconnector was concerned, I have reached the following conclusions:

(1)         There were officers, within ABB, who knew of the Cartel and who were involved in the tender for the Interconnector. They were in a position to influence ABB’s conduct, using their knowledge that ABB had been “allocated” the BritNed Interconnector. Although there were others who knew about the Cartel, I find that only one such officer was involved in the BritNed Interconnector tender such as to be able to influence it. That was Mr Jönsson. So far as Mr Pääjärvi is concerned, I have (for the reasons I have given) assumed that he was – in terms of knowledge – in the same position as Mr Jönsson. [326]

(2)         The relevant knowledge of Mr Jönsson and Mr Pääjärvi was, in essence, that whilst competition for the Cable element could not be entirely excluded, the main competitors (for the Cable element) were Nexans and Prysmian, who were both cartelists and who had both acceded in the BritNed Interconnector being “allocated” to ABB. The position so far as the Converter element was concerned was different in one material respect: there were potential non-Cartel competitors, who could be expected to tender and to tender competitively. [327]

(3)         Apart from Mr Jönsson and Mr Pääjärvi, the other persons involved in the tender were ignorant of the Cartel and ignorant of the fact that the BritNed Interconnector had been “allocated” to ABB. I find that these persons – who sat both above and below Mr Jönsson and Mr Pääjärvi within ABB’s organisation – would have approached the tender on the basis that ABB was in proper competition for the Interconnector and that ABB would lose the tender if its bid was uncompetitive. The question therefore arises as to the extent to which Mr Jönsson and Mr Pääjärvi could deploy their knowledge that this was not in fact the case to cause ABB’s bid to be inflated. As to this:

(a)            Because the compilation of the detail of ABB’s tender for the Cable element was in the hands of others – like Mr Larsson-Hoffstein – I do not consider that Mr Jönsson and Mr Pääjärvi had the ability directly to influence the level of direct costs within the Cable element of ABB’s bid for the Interconnector.

(b)           As regards the level of common costs within the Cable element of ABB’s bid for the Interconnector, because of the rather subjective nature of these costs, and because Mr Larsson-Hoffstein and his team were less directly involved in the allocation of such costs to the Interconnector tender, I consider that Mr Jönsson and Mr Pääjärvi had the potential to affect ABB’s internal processes so that the tender price, so far as common costs were concerned, was inflated. The question of whether they in fact did so is a question that I will have to return to, in Section I, after I have described the analysis of the expert economists.

(c)            So far as indirect influence over the tender was concerned, I find that there was the potential for such influence within ABB. In terms of its effect, this is a matter that must be considered in the round, in Section I, after I have described the analysis of the expert economists. [328]

I have not yet considered the issues of baked-in inefficiency and cartel savings.

H.      THE EXPERTS’ ANALYSIS

(1)     Introduction

285.       The experts adopted very different approaches in terms of how they sought to analyse the effect of the Cartel and the overcharge that BritNed contended the Cartel had caused. It is necessary, first, to state the approaches of Dr Jenkins and Mr Biro, as well as their findings, in as neutral a way as possible. This is done is Section H(2) in the case of Dr Jenkins and in Section H(3) in the case of Mr Biro.

286.       Having articulated the different approaches of the experts, I then proceed to consider them more critically. The experts reached divergent views as to the amount of the overcharge. Dr Jenkins concluded that the overcharge was in excess of 20%: her precise estimates varied as Dr Jenkins refined her analysis during the course of exchange of expert reports and the compilation of the Joint Statement. Mr Biro put the overcharge at nil.

287.       It goes without saying that it will be necessary to understand why this divergence has arisen, and whose assessment is the more reliable. This critical evaluation is undertaken in Section H(4).

(2)     The approach of Dr Jenkins

(a)     A summary of Dr Jenkins’ approach

288.       Dr Jenkins’ general approach was described by her in her first report. Her expectation was that the Cartel would lead to higher prices, when compared with the competitive counterfactual. [329] She tested this expectation in the following way: [330]

“I test this expectation by analysing and estimating the size of the cartel overcharge using information on the prices ABB agreed with its customers at the point of the tender award for disclosed cable projects supplied in the period 2001 to 2016. My preferred approach to estimate the overcharge is a during-and-after price regression. This uses the pricing of post-cartel cable projects (i.e. projects “after” the cartel period) as the benchmark for the competitive counterfactual price, while controlling for factors that affect the price of projects, and which differ between projects and over time.”  

289.       Thus, Dr Jenkins was comparing the price of projects during the Cartel with the price of projects after the Cartel. The former were presumed by Dr Jenkins (although, as will be seen, Dr Jenkins tested for this) to be affected by the Cartel, such that their pricing was higher than it would be absent the Cartel. The latter were presumed to be competitive prices. [331]

290.       Dr Jenkins’ analysis proceeded through the following stages:

(1)         Identification of a sample dataset. Plainly, it is necessary – as a starting point – to identify a dataset of “during” and “after” projects that can be the subject of the analysis. This dataset must:

(a)            Be sufficiently large so as to enable a robust statistical analysis. Self-evidently, the larger the sample, the less likely it is that an outlier will distort the analysis.

(b)           Be sufficiently homogenous so as to enable meaningful analysis. Again, self-evidently, if the dataset comprises projects which simply cannot sensibly be compared, i.e. their differences cannot be adjusted for, then including such projects within the dataset will distort the analysis. There is a very close relationship between this aspect and the next stage in the analysis, which I shall call (Dr Jenkins did not use the word) “normalisation”.

(2)         Normalisation. Even assuming a sufficient and sufficiently homogenous dataset, it is the nature of underground and submarine projects not to be fungible but distinct. If, to take an extreme example, the projects in question were identical , any overcharge would be easy (or easier) to detect. But, as has been seen, there is in fact a considerable difference between underground and submarine cable projects [332] and submarine cable projects are each bespoke. [333] Thus, before any price comparison can be undertaken, it is necessary to control for the factors that will – apart from the effect of the Cartel – cause the price of any given project to vary. These factors are, as it were, the “legitimate” factors that would cause a competitive price to vary. This process of normalisation essentially seeks to account for factors that will cause the price of a project legitimately to change. As noted, there is a close relationship between normalisation and a sufficiently homogenous dataset. If a given type of project cannot be “normalised”, then it is insufficiently homogenous to be included in the dataset. Of course, whether this is the case is itself an exceedingly difficult question, and one on which the experts before me differed.

(3)         Ascertaining the price and other project specific data for each project in the dataset. Self-evidently, the price for and other data relating to each project included in the analysis needs to be ascertained. Again, there is a relationship between the data that is available and the normalisation that can be undertaken in relation to that data. To be clear, the analyst can only work with the data he or she is given and inevitably the nature of that data affects the nature of the analysis.

(4)         Regression analysis. Having ascertained (i) the relevant dataset, (ii) the information relating to each project comprising the dataset and (iii) the factors that will affect the price of a cable project, it is possible to conduct what is known as a regression analysis. In Dr Jenkins’ words: [334]

“3.88 To estimate the effect of the cartel on the prices of cable projects, I use my sample of comparable projects with information on the relevant factors…to control for measurable and systematic differences between them. Using my regression analysis. I identify the relationships between the price of a project and other factors that have influenced the project price.

3.89   The analysis I have undertaken uses a common and widely accepted form of regression analysis: ordinary least squares (“OLS”). OLS identifies a “line of best fit”, which minimises the difference between the actual observed values of the relevant cable projects and the values predicted by the regression…

3.90   While differences do exist between power cable projects, these differences affect the overall price in a systematic way. I consider that a reliable estimate of the overcharge in this case can be identified by applying regression analysis that controls for the differences across projects and includes an indicator variable for whether the project was procured during the cartel period. The coefficient on the indicator variable gives the overcharge estimate.”

Regression analysis involves specifying a model using specifically formulated variables based on the material described in paragraphs 290(1) to (3) above, inputting the variables and the data into a standard piece of software (known as “Stata”, which was used by both experts [335] ) and producing as an output various correlation coefficients, including a coefficient indicating that the Cartel had an effect on prices.

If – as Dr Jenkins sought to do – a regression model is designed to explain price levels, but it also includes in the model variables representing other supply and demand factors, along with a variable that captures the illegal behaviour alleged in the case, the coefficient associated with that behaviour variable then provides an estimate of the impact of the alleged illegal behaviour on prices, holding constant the effects of other market factors.

(5)         Considering the reliability of the regression through sensitivity analysis and confidence intervals. It is possible to have an estimated coefficient that indicates a relationship between the variable representing the challenged conduct and prices where none exists in fact. Statistical significance is a way of testing whether the results obtained in a regression analysis are due to a coincidence or whether they reflect a genuine relationship.

291.       In the next sections, Dr Jenkins’ approach is considered in a little greater detail. It is necessary first to explain my understanding of the various statistical tools and devices used by Dr Jenkins (and, indeed, Mr Biro). This is done in Section H(2)(b) below. Thereafter, I describe:

(1)         The dataset used by Dr Jenkins.

(2)         The information regarding each project forming part of Dr Jenkins’ dataset.

(3)         Dr Jenkins’ consideration of how the data can be normalised.

(4)         Dr Jenkins’ regression analysis and its outcome.

292.       Dr Jenkins’ model went through a number of iterations, generally speaking as a result of and in response to points made by Mr Biro. That this occurred in no sense reflects negatively on Dr Jenkins’ work: indeed, it is much more a reflection of her integrity as an expert that she recognised force in some of the points made by Mr Biro and adjusted her model to deal with them. Despite these changes, Dr Jenkins’ model remained in essence unchanged: variables were changed, Dr Jenkins at times selected different data to draw on and, obviously, the specific figures produced by the model were different over time. I will, of course, place principal reliance on the final figures produced by Dr Jenkins’ model, but I should stress that what matters more than the figures is the robustness of the model that produced them. Because Dr Jenkins’ more detailed explanations were made in relation to her earlier models, there are times in the Judgment when I refer to an earlier model and its output, rather than Dr Jenkins’ final model and its output. That simply reflects the fact that Dr Jenkins very properly was disinclined to repeat herself in her reports, and therefore I must at times have resort to earlier explanations made in relation to earlier models.

(b)     Statistical tools and devices

(i)      What is a coefficient?

293.       A coefficient is simply a number that multiplies a variable. Thus, 6 x means “6 times x ”, where x is the variable and 6 the coefficient . Lawyers would refer to it as a “multiplier”.

(ii)     Correlation

294.       Correlation refers to a relationship between two variables. That relationship is described by a correlation coefficient. It is important to appreciate the correlation is just that – co-relation. It simply means that variation in the scores on one variable correspond with the variation in the scores on a second variable. There is no causation necessarily implied. Causation means that variation in the scores on one variable causes variation in the scores on a second variable.

(iii)    Measuring correlation

295.       Correlation is measured through the coefficient of determination . Where two variables are related, the scores on one variable are associated with the scores on a second variable. That association means that one can predict – to a certain extent – what the scores on the second variable (“V2”) will be, if one knows the scores on the first variable (“V1”). In short, a variance in the scores of V1 may be able to explain or predict (without suggesting any causal connection) the scores of V2. The coefficient of determination measures this relationship.

296.       The coefficient of determination is essentially a measure of shared variance. If V1 changes, V2 also changes. The stronger the correlation, the greater the amount of shared variance, and the more changes in V1 can explain or predict changes in V2. Essentially, the coefficient of determination provides a measure of the strength of the association between two variables.

297.       The coefficient of determination is the square of the correlation between V1 and V2. It is the percentage of the response variable variation in V2 that is explained by a variation in V1. The coefficient is measured as a percentage or as a figure lying between 0 and 1. [336] Thus, if the coefficient of determination is 0 or 0%, there is no correlation. If it is 1 or 100% then changes in V1 are a perfect predictor of changes in V2. Usually, the relationship between V1 and V2 lies between these two extremes. Essentially, R 2 is the square of: [337]

R = Explained variation

Total variation

298.       Thus, if R 2 = 0.92 or 92%, if V1 moves by 100, V2 will on average increase by 92. [338]

(iv)     Regression analysis

299.       Regression can be either “simple” or “multiple”. Simple regression is rather like correlation, save that correlation does not distinguish between independent and dependent variables, whereas simple regression has a designated predictor variable and a designated dependent variable. Just as with correlation, however, regression analyses do not allow analysts to claim a causal association. There is correlation, from which causation may be inferred, at most.

300.       Multiple regression involves two or more predictor variables and a single dependent variable. Multiple regression analysis permits the analyst to see how the predictor variables, as a group, are related to the dependent variable. It is also possible to see the relative strength of each predictor variable, and the strength of the relationship between each predictor variable and the dependent variable, while controlling for other predictor variables in the model.

301.       Two other points, regarding variables, need to be made:

(1)         It is possible for variables to relate through what are known as “interaction terms”.

(2)         A “dummy” variable is one that takes the value of “1” or “0” to indicate the presence or otherwise of a categorical effect that may be expected to influence the dependent variable.

302.       Both experts agreed that, in order to be reliable, a regression analysis must:

(1)         Be based upon a sufficiently large data-set. Clearly, the more data that can be input and reviewed for purposes of correlation, the more reliable the model will be.

(2)         Be well specified. To be well specified, it would (in this case) need to take account of the main drivers of project prices and not take account of factors which were irrelevant.

(v)     Testing how good a model is: statistical significance

303.       Dr Jenkins explains: [339]

“…regression results can only be estimates of the true values associated with the coefficients. A result is referred to as being “statistically significant” if it is unlikely to have been observed by chance. In regression analysis, this is normally assessed by calculating the likelihood of the estimated coefficient being observed if the true underlying coefficient is actually equal to zero – i.e. if there is no true effect between the variable associated with that coefficient and the dependent variable. The significance is reported as “probability values” (“p-values”). Thus, significance is indicated by p-values that are close to zero (i.e. there is a low probability that these results would have been observed if the true value of the underlying parameter is zero). Standard thresholds are 10%, 5% and 1%.”

304.       In her report, Dr Jenkins uses symbols to designate these thresholds. Thus: [340]

(1)         “***” means significant at 1%.

(2)         “**” means significant at 5%.

(3)         “*” means significant at 10%.

305.       I shall eschew the symbols and stick to measures of statistical significance at the 1%, 5% and 10% levels. A variable is said to be statistically significant at the 5% level (for example) if there is less than a 5% probability of having observed this estimate from the evidence, when in fact the true value is zero.

306.       Another measure is the “t-statistic”. The higher the value of the t-statistic, the lower the probability that the given estimate of the underlying parameter would have been observed, on the assumption that the true value of the parameter is in fact zero. As a rule of thumb, a t-statistic higher than 2 or lower than -2 indicates that there is a 5% probability, under a two-sided test, that the coefficient estimate would be observed, if the true coefficient is zero. The actual threshold for the t-statistic depends on the number of observations and number of control variables included in the analysis.

307.       A final measure is the confidence interval or hypothesis testing. As has been described, the regression will produce a coefficient stating the correlation between the dependent variable and the predictor variables. Let us suppose a relatively high coefficient of determination, where R 2 = 0.92.

308.       Hypothesis testing seeks to measure the uncertainty around this coefficient. Essentially, it produces a range around the coefficient where the analyst can express a degree of confidence that the coefficient will lie within this range. Thus, the analyst may be able to say that he or she is 95% confident that the true coefficient will lie within the range 0.90 to 0.94, with 0.92 being the most probable outcome. On the other hand, the analyst might only be able to express this confidence to a standard of 60%. That would imply, in order to have a confidence of 95%, the range of value attaching to the coefficient would be much greater.

(vi)    The Practical Guide on Quantifying Harm

309.       The Practical Guide on Quantifying Harm draws together the various threads extremely clearly, and says this about regressions and statistical significance:

“81.   Carrying out a regression analysis requires knowledge of various statistical techniques to measure the relationship between variables, to construct an appropriate regression equation and to calculate the precision of the parameters in this equation. In addition, it is necessary to have a good understanding of the industry concerned, in the first place, to formulate the right hypothesis when constructing the regression equation and to make the right choice as to the factors that are likely to have significantly influenced the variable of interest (and which should therefore be included in the analysis). Industry understanding is furthermore necessary to make informed choices about which statistical techniques to use in a given situation, for instance, to account for unusual observations (outliers) or other specific features in data sets. In particular, where the influencing variables were themselves affected by the infringement, biased results may occur if this aspect is not taken into account, e.g. through applying specific statistical techniques or through using data observations that lie outside the infringement period or market.

82.     Without a sufficient number of data observations, statistical analysis cannot identify relationships between economic variables. To identify the effect of influencing variables on the variable of interest therefore requires that a sufficient range of data observations is available for all variables considered. Regression analysis therefore typically requires extensive data. However, statistical techniques may help to overcome some gaps in data or biases in their interpretation and there can be situations where also the analysis of a smaller number of data observations is meaningful.

83.     Data observations can, in principle, be gathered at different levels of aggregation. For example, where the relationship between price and input cost is to be analysed, data series either for the prices charged in individual transactions, for annual industry average prices or – in between – monthly data at firm level could be examined next to data series either for individual input costs per unit or for industry cost averages respectively. Using disaggregated data makes it possible to analyse a greater number of observations and therefore to obtain more precise estimates. Where such disaggregated data do not exist or are not accessible to the party carrying out the regression analysis, the analysis of aggregated data may still produce informative results, in particular if the aggregated data have a high frequency.

84.     Having a sufficient range of data observations and the level of data aggregation are examples of the importance of data reliability and data relevance for economic analysis. However, most datasets are incomplete, and not all relevant facts may be observed or measured with high accuracy. It is therefore proper to explicitly acknowledge those imperfections. Deficiencies in the data should not prevent an economic analysis from being given proper weight, though conclusions should be drawn with caution.

85.     Where used appropriately and on the basis of sufficient data observations, regression analysis can considerably refine the damages estimation through comparator-based methods. It should be stressed, however, that even very sophisticated regression equations rely on a range of assumptions and will (like any technique to predict a hypothetical situation) only be able to deliver estimates. It is good practice to consider the assumptions underlying a regression equation, because some assumptions may be more appropriate than others in a given situation and may lead to significantly different results.

86.     One way to deal with the uncertainty of the estimate is to indicate the results not as a point estimate (“the price in the non-infringement scenario is 10€”), but as an interval (“the price in the non-infringement scenario is between 9€ and 11€”). The notion of “confidence interval” – which is standard in statistics – is used to describe how likely it is that the true value is contained in an interval. By convention in economics, a 95% likelihood that a specific interval does in fact contain the true value is regarded as a high degree of certainty.

87.     A similar way of dealing with the uncertainty of estimates is to refer to the notion of “statistical significance”, which is a standard way of testing whether the results obtained in a regression analysis are due to a coincidence or whether they reflect in fact a genuine correlation. For this, a certain hypothesis is tested: in the field of damages actions, such a hypothesis could for instance be whether the cartel infringement did in fact have an actual effect on prices or not. The hypothesis that the infringement did not have an effect (and that therefore the non-infringement price does not differ from the price in the infringement scenario) is called the “null hypothesis”. Regression analysis is then used to test this null hypothesis. A result of a regression analysis is said to be statistically significant when it is possible to reject the null hypothesis, because it would be very unlikely that the results observed a due to chance. By convention, a likelihood of at least 95% that the null hypothesis is rejected is regarded in economics as allowing to judge that the results are “statistically significant”.

88.     As described above, it is a convention in economic science for both the notion of “confidence interval” and “statistical significance” to use a 95% threshold of probability. It should be stressed that this represents a pure convention and that more as well as less stringent thresholds (for instance: 99%, or 90% probability) may likewise provide useful information. This is because statistical significance is determined, in part, by the number of observations in the data set: other things being equal, the statistical significance increases as the sample size increases. It is good practice to indicate the probability threshold chosen. In a damages action, it is then for the court, under applicable law, to decide the probative value of such regression analysis and the procedural consequences (in particular with regard to the burden of fact-pleading and proof) which such analysis may entail.”

(vii)   One-sided and two-sided hypothesis testing

310.       When considering statistical significance and hypothesis testing, it is necessary to understand whether a one-sided or a two-sided test is being conducted. A two-sided hypothesis test will have no a priori view about whether the null hypothesis is that the coefficient will be positive or negative. Thus, for a 90% threshold of statistical significance, what is tested is whether the estimated coefficient lies in either of the two extreme areas of the distribution around zero (positive and negative) with 5% lying at each end of the distribution. [341]

311.       A one-sided test allocates the entire 10% to one end of the distribution. In this case, Dr Jenkins tended to use a one-sided test, with the 10% allocated to the positive side of the distribution. This is because Dr Jenkins’ null hypothesis was that the effect of the Cartel was nil, i.e. that the prices charged by ABB were unaffected by the Cartel. Dr Jenkins rejected, as implausible, the hypothesis that the Cartel would cause ABB’s prices to fall. Had Dr Jenkins considered such a hypothesis to be appropriate, then Dr Jenkins would have used a two-sided test. This aspect of Dr Jenkins’ approach was explored in cross-examination: [342]

Q (Mr Hoskins, QC)

…Under a one-sided test, and you look at just one end of the graph, and that is what you have shown here, so a 95% threshold or 5%, whichever way you want to put it, rather than having a two-sided test, where you would have 5% at each end, in the one-sided test you look at the 10% at the positive end? Is that correct?

A (Dr Jenkins)

So, just to clarify, if you are looking at a 90% threshold, and it is two sided, then it is 5% at each end, and if it is one-sided, it is 10% at one end, yes. If it were a 95% threshold, which I think you started the question with, that would be 2.5% at each end or 5% at one end.

Q (Mr Hoskins, QC)

Thank you. And again, the convention is to use a two-sided test, is it not? In the same way the convention is to use a 95% threshold, there is a convention in economics to use two-sided test?

A (Dr Jenkins)

The convention for one-sided or two sided depends on whether you have a prior belief. So it is conventional to use a two-sided test where you have no prior view about whether the estimate is going to be positive or negative, and where you have a prior view that you will have – the underlying coefficient that you are trying to estimate is either positive or negative, so it is either less than or greater than, but if you have a prior view of about which side, then one-sided testing is conventional.

 

Q (Mr Hoskins, QC)

Now, there has been a lot of factual evidence about the potential impact of this Cartel on the BritNed project. If the judge considers that it cannot be presumed that the Cartel raised prices on the BritNed project, then do you agree it would not be appropriate to use a one-sided test?

A (Dr Jenkins)

That would be correct. If the judge chose to the view that it was as likely to have a significant negative effect of the Cartel as a significant positive effect, then you would use a two-sided test.

 

Q (Marcus Smith J)

…[Dr Jenkins] is saying that she has adopted the one-sided test on the basis that she has presumed…that cartels raise rather than lower prices. Now, of course, we are concerned here with the effect of the Cartel specifically on a single project, the BritNed cable. And I think what Mr Hoskins was saying was that if I were to take the view that the Cartel could be neutral in this case, would a one-sided or a two-sided test be appropriate?

A (Dr Jenkins)

So, then, the one-sided test is appropriate, because the null hypothesis is that it is a neutral effect, that there is no effect of the Cartel.

If the question is what is your reply about the alternative, is that whether there is a positive effect or a positive or negative effect?

Q (Marcus Smith J)

So, if I were to take the view that the Cartel could actually move prices in either direction, at that stage I think you would be saying it would have to be a two-sided test that would be adopted?

A (Dr Jenkins)

That’s correct.

 

Q (Marcus Smith J)

What you do when you have adopted a one-sided test is effectively you put all of the probability on one side?

A (Dr Jenkins)

[Nods.]

Q (Marcus Smith J)

Is that in itself a reason for moving to a higher probability, that is to say, whereas if one was using a two-sided test with 5% on either side, is it an argument for having a 95% probability or 5% probability, so you just load the 5% on one side – or is that just not a very good question?

A (Dr Jenkins)

That is a good question. And as I said before, it is in a sense up to the reader’s view about what amount of probability they want to tolerate for having the probability that this variable – this result would have been observed if the underlying true effect was zero. So, if you say 95% one-sided, as you rightly say, you are putting 5% in the top there, and you are saying you are only willing to tolerate a 5% chance that this could be observed if the true effect is zero. If you use a 90% threshold, you’re saying I’m willing to tolerate up to 10% of a chance.

(c)     Dr Jenkins’ dataset

312.       Dr Jenkins’ dataset comprised: [343]

(1)         Successful ABB tenders for projects;

(2)         Comprising both underground and submarine projects; [344]

(3)         From both the Cartel and the post-Cartel periods. [345]

Specifically, the dataset comprised: [346]

 

Cartel

Post-Cartel

Total

Submarine

15

34

49

Underground

21

22

43

Total

36

56

92

Table 3 : The dataset used by Dr Jenkins

313.       Dr Jenkins regarded the size of this dataset as being at the lower end of what she would have hoped to have for the regression she performed. [347]

(d)     Information regarding each project forming part of Dr Jenkins’ dataset

314.       As regards each of these 92 projects, Dr Jenkins obtained the following information: [348]

(1)         The contract award date.

(2)         The contract price at the contract award date.

(3)         ABB’s expected installation costs.

(4)         ABB’s expected accessories costs.

(5)         The number of cables provided.

(6)         The length of each cable.

(7)         The conductor cross-section of each cable.

(8)         The number of cores within each cable.

(9)         The voltage of each cable.

(10)       Whether the cable was submarine or underground.

(11)       The conductor material of each cable.

(12)       ABB’s copper and aluminium input prices.

315.       Dr Jenkins’ first source for this information were the PPMs. [349] Dr Jenkins described her process as follows: [350]

“3.86 When compiling my dataset, I sought to draw all the relevant information for a given project from as few sources as possible, to ensure consistency within the project record. Therefore, I have drawn datapoints from the identified documents in the following way:

·          Where a selected pricing model includes all the required information, I have drawn all the information for a given project from that selected pricing model;

·          Where a selected pricing model does not include sufficient information on the cable specifications, I have drawn information from a combination of the pricing model and a detailed specifications document that is consistent with the selected pricing model.”

(e)     Dr Jenkins’ process of normalisation

316.       Dr Jenkins used the data described in paragraph 314 above to inform her analysis of the factors that would affect the price of a cable project. She did not use the price agreed by ABB with its customers. Dr Jenkins’ thinking was that it was more appropriate to model what a competitive price should be by conducting her “during and after” analysis. In this way – by assuming that post-Cartel projects would be competitively priced – her model would be able to ascertain the Cartel overcharge.

317.       Obviously, a critical element in this process was the “normalising” of the various projects being used as part of the assessment. [351] In this case, normalisation involved two significant exercises:

(1)         Identifying the legitimate factors that would go to affect price. In her first report, Dr Jenkins set out her understanding (derived from Mr Röstlund’s evidence) of the cost drivers of cable projects. [352] She then explained how she proposed to model these cost drivers, so as to create the inputs that would feed into a model that was independent of ABB’s actual reported costs. [353] In summary (and I do not propose to set out all of the relevant parts of Dr Jenkins’ report) her approach was as follows: [354]

“…I include controls in my overcharge analysis to take into account the drivers of the costs of raw materials and of manufacturing of HV cables…, which may vary across projects and over time. Specifically, I include cost drivers and use appropriate specifications to control for the length of cable, the size of the cross-section, the cost of the metals used for the conductor, the number of cores, the voltage, and whether the cable is for an underground or submarine project.”

(2)         Ensuring that the differences between underground and submarine cable projects were correctly weighed. The differences between underground and submarine cable projects have been described. [355] Dr Jenkins specifically considered whether underground projects were sufficiently comparable (or homogenous) to be included in her analysis. She concluded: [356]

“3.65 I understand that the cartel covered all types of underground power cables of 110kV and above. Based on my review of the Decision and the factual evidence, I consider that there are many underlying similarities in terms of the factories used, the approach to allocating projects between cartel members, and the cost drivers.

3.66   In my view, the differences between submarine and underground projects…are systematic (i.e., they are likely to have a predictable relationship with price) and I am able to control for these in my price analysis. For example, submarine projects:

·          Are generally greater in length and size of the cross-section – I control for the volume of cable in my analysis;

·          More commonly use copper than aluminium – I control for the price of the metal used for the conductor in my analysis;

·          Are more costly due to the greater amount of insulation, armouring, production process, type-testing, storage and transport costs and insurance costs – I control for this by allowing submarine projects to have a higher price for a given set of characteristics.

3.67   I therefore use information on both underground and submarine cable projects to estimate the cartel overcharge, controlling for the systematic differences between submarine and underground projects.”

(f)      Dr Jenkins’ regression analysis

318.       The project value at contract award is the dependent variable in Dr Jenkins’ regression specification. [357] The other variables – the so-called “control” variables – are as follows: [358]

 

Variable

Cartel Indicator

This is a “dummy” variable, having a value of “1” during the Cartel period and a value of “0” for the post-Cartel period. [359] This variable is intended to capture the cartel effect. Specifically, it captures the average effect of the cartel. [360] Dr Jenkins explained the point as follows: [361]

“By having the binary indicator variable, what I’m doing is getting an average effect during the Cartel of all the projects that ABB sold during the Cartel, and it is saying that given the factors that one would expect to see driving the costs of those projects, that on average it looks as if these projects were too high by this amount. That’s what the binary element is. It is saying, across all these projects, what is it saying about what seems to be the overall Cartel effect.”

It is the Cartel Indicator that is used to calculate the amount of the overcharge for any given project value. The manner in which the overcharge is calculated from coefficients for this variable is complex, but the process was agreed by both experts, and was uncontroversial. [362]

Total volume of conductor materials

This is derived by multiplying the cable length by the size of the cross-section and the number of cores. According to Dr Jenkins, this controls for costs that increase with the size of the cable. [363]

Submarine Indicator

To capture any differences in the level of costs between submarine and underground cable projects. [364]

Copper and Aluminium Prices

These are taken from the relevant time period, to control for the cost of the metal used for the conductor. [365]

Cable Voltage

To control for the differences across projects and the impact of the voltage on the thickness of insulation required. [366]

Cost of Installation and Accessories

These are controlled by scaling for the volume of conductor materials. For this variable, Dr. Jenkins relied on the reported costs of installation and accessories from ABB’s disclosure. [367]

Capacity Utilisation Control

Dr. Jenkins initially used ABB’s order backlog for the power technologies division in the quarter prior to the award of the contract to control for ABB’s “appetite” for a project, which (according to Dr. Jenkins) would be expected to affect the price for the project. [368]

Time Trend

To reflect the quarter and the year in which a project was awarded, to capture any long term changes in pricing over time, such as inflation or cost efficiencies. [369]

Table 4 : Control variables used by Dr Jenkins

319.       The results of Dr Jenkins’ analysis were as follows : [370]

Variable

Coefficient

P-value

Cartel Indicator

0.30

5%

Total volume of conductor materials

0.86

1%

Submarine Indicator

0.73

1%

Copper and Aluminium Prices

0.29

1%

Cable Voltage

0.21

1%

Cost of Installation and Accessories

0.37

1%

Capacity Utilisation Control

0.70

1%

Time Trend

0.0075%

5%

Table 5 : Outcome of Dr Jenkins’ regression analysis

It should be noted that Dr Jenkins’ p-value for the cartel indicator is based on a one-sided, not a two-sided, test.

320.       On this basis, Dr Jenkins concluded that “a reasonable and reliable estimate of the overcharge suffered by BritNed as a result of the cartel in HV cable projects is 25.4%”. [371]

(4)     The approach of Mr Biro

(a)     A summary of Mr Biro’s approach

321.       Mr Biro used what he called three complementary methodological approaches to assess what the price of the BritNed project would have been, but for the Cartel: [372]

(1)         A price comparison analysis controlling for ABB’s actual costs of supply.

(2)         An econometric analysis of the relationship between prices and ABB’s actual costs of supply.

(3)         A price comparison analysis which does not directly control for ABB’s actual costs of supply, but instead uses proxy measures based on the technical characteristics of the projects.

322.       The differences between Dr Jenkins’ approach and Mr Biro’s approach are considered in greater detail below, but one difference stands out immediately. Whereas Dr Jenkins sought to ascertain the overcharge generally caused by the Cartel by comparing cartelised and post-Cartel projects (including underground as well as submarine projects), Mr Biro compared the price of the BritNed Interconnector project alone with the prices of other submarine ( not underground) power cable projects in the post-Cartel period . [373]

(b)     Mr Biro’s first approach: a price comparison analysis controlling for ABB’s actual costs of supply

(i)      Overview

323.       Mr Biro described his approach as involving: [374]

“…a comparison of the price of the BritNed project at the point of tender award with the price offers made by ABB in relation to other submarine power cables projects which were tendered in a competitive environment after the end of the cartel, adjusting for differences in the costs that ABB expected to incur in supplying these projects. This methodological approach controls for the impact of project-specific costs on project prices by analysing and comparing the margin (or mark-up) over these costs that ABB sought to achieve on the BritNed project with the margins on comparable submarine power cables projects tendered outside the cartel period for which ABB submitted a price offer; this price comparator analysis can therefore also be referred to as a “margin analysis”. In particular:

·          If the margin on the BritNed project were not systematically higher than the margins on comparable non-cartel period projects, then I would conclude that the price of the BritNed project was no higher than would have been expected in a competitive environment, given the differences in costs of supplying the projects, and would indicate no evidence of an overcharge; and

·          In contrast, if the margin on the BritNed project were higher than those on comparable projects tendered outside the cartel period, then this would imply that the price of the BritNed project was higher than would have been expected under competitive conditions, taking into account differences in the costs of supply, and would indicate evidence of an overcharge. In this case, the size of any margin difference would provide a measure of the size of the BritNed overcharge.”

324.       Essentially, Mr Biro sought to compare the gross margin that ABB expected to achieve on the BritNed project with the margins that ABB sought to earn on comparable submarine power cable projects tendered after the Cartel. [375] Broken down, Mr Biro’s approach was as follows:

(1)         Identifying comparable projects. Like Dr Jenkins, Mr Biro needed a pool of comparator projects which would be subjected to his margin analysis.

(2)         Identifying and dealing with multiple margins earned by ABB. ABB comprises several business units. In this case, only one business unit was involved in the supply of the cable to BritNed. [376] However, “for some other post-cartel comparator projects which I have analysed, other ABB business units were involved in the supply of certain elements of the project, usually in relation to the logistics and organisation to the cable installation. The result of this internal structure is that an offer prepared by ABB for a submarine power cables project may have incorporated multiple margins.” [377] Mr Biro needed to deal with such multiple margins in a consistent way.

(3)         Calculating gross margins. Plainly, it is important that Mr Biro define “gross margin” and then apply this consistently across the comparable projects.

(ii)     Comparable projects

325.       Mr Biro did not consider that underground cable projects were suitable comparators for his purposes. [378] He was sufficiently concerned about potential differences between different types of submarine cable projects: [379]

“…any analysis that seeks to assess the BritNed overcharge by means of a comparison of project prices needs to ensure that the price comparisons are genuinely on a like-for-like basis, or otherwise make appropriate adjustments to account for differences in the costs of supplying these projects and other relevant drivers of project prices. Through comparing anticipated gross margins, my methodological approach directly controls for differences across projects in terms of their costs of supply. However, there are also other commercial and competitive considerations which influenced the pricing of power cables projects, and which cannot readily be quantified, with implications for which types of power cables projects should appropriately be used as comparators to the BritNed project.”

Mr Biro used submarine cable projects as his comparators, but noted the different types of project within this class. [380]

(iii)    Multiple margins and margin consolidation

326.       In relation to multiple margins earned by different units of ABB, Mr Biro used a single consolidated project margin which would reflect the return expected by all units of ABB involved in that particular project. [381]

(iv)    Calculation of gross margin

327.       Mr Biro noted that project margins were reported by ABB on both a “gross” and a “net” basis: [382]

“a.     gross margins refer to project profitability taking into account those costs which ABB considered to be directly attributable to the supply of the specific project in question; and

b.       net margins refer to the bottom-line profitability after also allocating to each project a certain proportion of the general business overheads incurred in relation to administration and management, non-order related R&D costs, and sales and marketing costs (including any commission paid internally to local ABB units). Although these common costs were not directly related to the manufacture and delivery of any individual power cables project, they were allocated by ABB across projects to produce reported net margins in recognition that they would ultimately need to be recovered by the power cables business as a whole for it to remain viable.”

328.       Mr Biro used gross margin for the purposes of his margin analysis: [383]

“For the purpose of assessing the extent of the BritNed overcharge, I consider it most appropriate to make comparisons across projects on a gross margin basis . Since gross margins should reflect only those costs which were directly attributable to individual power cables projects, conducting my comparator analysis at the gross margin level is the most economically meaningful approach. This is because one would, as a matter of economics, expect direct project-specific costs to constitute the primary driver of project prices. Moreover, this avoids any concern that the particular allocation by ABB of its business overheads and other shared costs may impact on the results of my overcharge analysis.”

329.       Mr Biro did not , however, base his analysis on the gross margins as reported by ABB but calculated his own. He did this using the information contained in the PPMs. [384] Specifically: [385]

“…I have therefore not based my comparator analysis on the gross margins as reported by ABB and have instead calculated a consistent measure of gross margins across all relevant projects. Specifically, I have calculated gross margins by deducting from the project revenues the following cost items:

a.       manufacture of the cables, including raw materials, use of machines employed in the manufacturing process and direct manufacturing labour;

b.       design of the cables, and any required tests conducted on the cables;

c.       cable accessories, whether supplied directly by ABB SEHVC or by any other ABB business unit or external supplier;

d.       transportation of the cables from the factory to the location where the cable was to be installed;

e.       activities related to the installation of the cable;

f.       project management;

g.       insurances, warranties, taxes, duties and similar items; and

h.       provisions for cost overruns and project risks.”

330.       Mr Biro excluded from his calculations of gross margin any cost item he did not consider to be directly attributable to the specific project in question. [386] Mr Biro considered that the costs items he was considering had been accounted for by ABB in a consistent manner over time. [387] In terms the figures he used, Mr Biro sought to use price and cost information that was as nearly as possible contemporary with contract award. In other words, he did not consider earlier cost and pricing data, nor post-contract cost and pricing data. [388]

331.       The outcome of Mr Biro’s margin analysis was as follows:

Project

Bid outcome

Price

Cost

Margin

Type of project

BritNed

Won

2,726,400,218

2,218,089,507

18.6%

18.6%

“Project 1”

Lost

2,080,093,029

1,712,813,775

17.7%

Post-cartel submarine interconnectors HVDC MI [389] turnkey

17.6%

“Project 2”

Lost

1,745,886,408

1,502,048,752

14.0%

“Project 3”

Lost

3,521,665,480

2,906,391,434

17.5%

“Project 4”

Lost

3,187,543,377

2,637,194,484

17.3%

“Project 5”

Lost

2,325,662,065

1,922,135,626

17.4%

“Project 6”

Lost

2,372,222,799

1,956,437,196

17.5%

“Project 7”

Won

3,884,278,656

3,218,018,932

17.2%

“Project 8”

Lost

1,014,491,836

803,520,465

20.8%

“Project 9”

Lost

852,059,995

716,180,583

15.9%

“Project 10”

Lost

5,247,786,554

4,177,060,126

20.4%

“Project 11”

Won

391,404,858

313,653,215

19.9%

Post-cartel submarine interconnectors HVDC XLPE [390] turnkey

19.3%

“Project 12”

Won

3,246,651,510

2,594,326,728

20.1%

“Project 13”

Lost

2,442,564,617

1,979,100,213

19.0%

“Project 14” [391]

Won

3,111,467,997

2,487,839,113

20.0%

“Project 15”

Won

2,758,616,823

2,273,843,923

17.6%

“Project 16”

Won

223,953,709

180,114,752

19.6%

Post-cartel submarine interconnectors HVAC [392] turnkey

20.3%

“Project 17” [393]

Won

245,321,695

196,360,692

20.0%

“Project 18”

Won

103,974,615

80,426,644

22.6%

“Project 19”

Won

49,391,218

40,411,403

18.2%

“Project 20”

Won

59,615,868

46,907,790

21.3%

“Project 21”

Won

838,282,244

670,437,450

20.0%

Post-cartel submarine for offshore wind farms and oil and gas platforms, turnkey

22.1%

“Project 22”

Won

1,202,622,900

874,706,902

27.3%

“Project 23”

Won

763,348,721

597,246,549

21.8%

“Project 24”

Won

448,289,147