BritNed v ABB: Defining “Overcharge” and a New Toolkit for Quantifying Cartel Damages (baked‑in inefficiencies, cartel savings, and market‑wide counterfactual)

Citation: BritNed Development Ltd v ABB [2018] EWHC 2616 (Ch), Chancery Division (Business List)
Judge: Marcus Smith J
Date: 9 October 2018

Introduction

This landmark judgment is the first English damages award following a cartel infringement under Article 101 TFEU/Article 53 EEA to be determined after a full trial on both liability (by reference to a European Commission decision) and quantum (para 423). The dispute arose out of ABB’s supply of the submarine cable (“Lot 2”) for the 1,000MW BritNed Interconnector between Great Britain and the Netherlands. After the Commission’s 2014 Decision in Power Cables (AT.39610) found a long‑running international cartel (1999–2009), BritNed—jointly owned by National Grid and TenneT—sued ABB for damages.

BritNed framed three heads of loss:

Overcharge on the cable price;
Lost profits from allegedly being deterred by cartelised prices from procuring a higher capacity (1,320MW) cable; and
Compound interest said to reflect higher capital costs caused by the overcharge.

ABB denied any overcharge and challenged the lost profits and interest claims. It also argued that any damages must be reduced in light of a regulatory internal rate of return (IRR) cap condition on BritNed’s revenues.

Summary of the Judgment

Overcharge awarded: €13,009,568 in total (para 464)—comprising:
- €7,516,639 for “baked‑in inefficiencies” in ABB’s product (principally excessive copper/thickness in the HVDC mass‑impregnated cable); and
- €5,492,929 for cartel savings (illegitimate savings enjoyed by cartelists from not having to engage in fully competitive tendering) (paras 454–458).
Lost profits claim dismissed: BritNed would still have chosen a 1,000MW cable in the counterfactual competitive world; the 1,320MW option remained commercially and technically disfavoured despite any price effects (paras 466–507).
Compound interest claim dismissed: Project was equity‑financed; any financing “cost” was borne by shareholders, not BritNed, so BritNed cannot recover it (paras 543–549).
Regulatory IRR cap (deduction) rejected: No deduction to damages. The cap only bites after 25 years and, even then, any “excess profits” benefit does not accrue to ABB or BritNed in a way that should reduce compensation (paras 529–541).

Along the way, the Court set out important principles on (i) what constitutes “overcharge” and actionable “gist” damage in cartel cases; (ii) the evidential/analytical approach to quantifying harm; (iii) what parts of a Commission infringement decision bind the domestic court; and (iv) how regulatory regimes and finance structure interact with private damages.

Analysis

Precedents and Authorities Cited

The judgment synthesises a wide body of authority to develop a pragmatic, evidence‑led approach to competition damages:

Commission decisions: binding scope—Netherlands v Commission (Case C‑164/02): only the operative part of a Commission decision is binding; recitals bind only if they are the essential basis of the operative decision. In English law, section 58A Competition Act 1998 and Article 16(1) of Regulation 1/2003 are the anchors (paras 67–68).
Quantification approach—Asda Stores v Mastercard [2017] EWHC 93 (Comm): courts use “sound imagination and the broad axe,” accept estimation, and, where compelled, err on the side of under‑compensation in the face of uncertainty (paras 12(8)–(9)).
Proof of damage (gist of action)—Mallett v McMonagle, Forster v Outred, UBAF v EAB, Barker v Corus: clarifies when actionable damage accrues and how future/hypothetical assessments are approached (paras 10–12, 422–429).
Loss of a chance, causation/quantification—Allied Maples v Simmons & Simmons: separates what must be proved on balance of probabilities from what is assessed as a matter of quantification and chance (para 469).
Collateral benefits and netting off—Parry v Cleaver, Hodgson v Trapp, Swynson v Lowick, Tiuta v De Villiers: informs net loss analysis and why IRR cap does not reduce damages (paras 531–541).
Damages for interest—Sempra Metals v IRC: interest is recoverable as a head of loss when pleaded/proved; but here claim fails because equity funding costs are shareholder‑side (paras 545–549).
Gourley principle and future contingencies—courts must take account of how external regimes might affect outcomes, but only where relevant and not speculative (paras 538–542).

Core Legal Reasoning and Doctrinal Developments

1) What counts as “overcharge” in a cartel case

The Court resolved a pivotal pleading point: “overcharge” is not restricted to the price ABB would have quoted absent the cartel. It is the difference between the price actually paid and the price that would have prevailed in a competitive market whoever won the tender (paras 15–18). This market‑wide counterfactual matters because cartel allocations can remove truly competitive third‑party bids—so the counterfactual is not “ABB without a cartel”, but “the market without a cartel.”

2) No presumption of harm before the Damages Directive

Directive 2014/104/EU’s presumption that cartels cause harm (and its UK transposition) did not apply to this pre‑2017 claim. The Court refused to create a freestanding presumption via the EU principle of effectiveness (paras 19–23). Instead, harm had to be proved on the evidence, albeit applying Asda’s broad‑axe pragmatism.

3) Binding effect of the Commission Decision

Only the operative part binds; recitals are binding insofar as they are the essential basis of the operative part; other recitals are persuasive only (paras 67–71). This constrained the use of the Decision to establish the existence of the cartel and its salient features, while leaving project‑specific effects for proof at trial.

4) “Gist” damage in cartel cases

Actionable damage accrues when the cartelist’s conduct reduces the claimant’s consumer benefit (e.g., price effects or a reduction in competitive alternatives). In this case, the cause of action was complete by the time BritNed contracted at a cartel‑affected price (paras 422–429). The Court recognised that in cartel claims the “wrong” is the collective failure to compete; “gist” damage is a reduced competitive field or price uplift—conceptually not limited to a direct price effect alone (para 427).

5) Evidence and quantification: rejecting a regression, preferring a margin analysis—but adding back two cartel effects

Two sophisticated but competing methodologies were advanced:

Claimant’s regression model (Dr Jenkins): a “during‑and‑after” econometric comparison of cartel‑period and post‑cartel projects (including underground cables), with proxies for cost (length, cross‑section, metals etc.) and variables for time trend and order backlog. The Court held the model was too fragile and insufficiently tailored to bespoke submarine projects:
- Underground projects were not comparable; a “submarine/underground” dummy could not control for fundamental technical/commercial differences (paras 388–397).
- Time trend and order backlog variables were poor proxies; their inclusion/exclusion swung the results and undermined statistical significance (paras 398–413).
- One‑sided hypothesis testing (assuming cartels cannot lower prices) was contested; sensitivity testing revealed wide confidence intervals and non‑robustness (paras 310–311, 418).
Defendant’s margin analysis (Mr Biro): a project‑level comparator analysis of ABB’s gross margins on post‑cartel submarine projects using actual internal costings (PPMs), thereby avoiding proxies and controlling for bespoke features. The Court preferred this as “anchored” in observed costs, with sensible controls (paras 321–342, 414–416).

Crucially, the Court did not simply accept “no overcharge” from the margin analysis. Instead, it identified two distinct cartel‑related components requiring adjustment to reach the competitive counterfactual:

Baked‑in inefficiencies (direct costs): ABB’s HVDC MI cable design used more copper/thicker cable than efficient rivals would have used. In a competitive tender ABB would have had to absorb that inefficiency (i.e., concede price) to win, so a price reduction equivalent to 15% of copper cost was warranted (paras 445–453).
Cartel savings (common costs): cartel allocation lowers tendering/coordination costs and improves factory loading certainty. These illegitimate “savings” raise margin in the cartel world compared with competitive conditions; a method was devised to allocate a portion (1.9%) to BritNed’s project (paras 454–458).

This two‑pronged framework—disentangling technical inefficiency from competition‑suppressed cost savings—is the judgment’s most practical contribution to the law and economics of cartel damages.

6) The negotiation narrative mattered

The Court examined the tender history and negotiation dynamics closely (Sections F and G). Notably, ABB extended a €10m “Lot 3” discount to “Lot 2” (cables only) even though it faced no rival cable bid at the end. The judge found this resulted from ABB’s senior decision‑maker (Leupp) misappreciating the competitive situation because he was unaware of the cartel; this supported the conclusion that direct costs were honestly compiled and margins were competed down, but did not negate the two cartel‑effects identified (paras 161–166, 441–444).

7) Lost profits: capacity choice would not have changed

BritNed contended that competitive pricing would have made a 1,320MW link viable, yielding higher auction revenues. The Court rejected this: non‑price considerations (unproven technology at 1,320MW, a system cap at 1,320MW limiting “dynamic rating,” and diminishing marginal revenues with size) already pointed to 1,000MW; on the pricing side, in the counterfactual world Base Case 2 would fall (due to removing ABB’s baked‑in inefficiency) and Base Case 3 would not—if anything strengthening the 1,000MW choice (paras 475–507).

8) IRR cap (regulatory cap) does not reduce damages

ABB argued that any overcharge increased BritNed’s allowed IRR ceiling, enabling BritNed to retain more profit before the cap bites (thus no “net loss”). The Court:

Held the cap only “bites” after 25 years, with a 10‑year report being informational (paras 529–530);
Applied net loss principles and The Albazero line: any “excess profits” are earmarked to capacity expansion or regulated network funding—these are not BritNed’s to keep, so there is no benefit that should be netted off the damages (paras 534–541); and
Suggested a practical undertaking to avoid double recovery depending on how the regulators treat any damages in the IRR computation (para 540).

9) Compound interest: no recovery where equity funded

Invoking Sempra Metals, BritNed sought compound interest on the basis that the overcharge increased its capital costs. The Court held that because BritNed was funded by equity from its shareholders (National Grid and TenneT), any financing “cost” sits with them, not BritNed; BritNed cannot claim its shareholders’ expected returns (paras 545–549). This underlines the need to align the pleaded interest loss with the claimant’s own balance sheet and financing reality.

Impact: Why this case matters

Overcharge is market‑wide: Claimants may frame overcharge by reference to the competitive price that would have prevailed in the market, not just the infringer’s hypothetical non‑infringing price (paras 15–18). This prevents a defendant from limiting counterfactuals to its own internal pricing.
No pre‑Directive presumption of harm: For legacy claims, harm is not presumed. But courts will compensate using the Asda “broad axe” where precision is impossible (paras 19–23).
Methodology matters: The Court’s scepticism of generic regressions for bespoke projects (mixing underground/submarine, relying on weak proxies) will influence expert practice. Project‑specific analyses grounded in contemporaneous cost data may be preferred—while still adjusting for distinct cartel effects (paras 343–421).
Two novel heads in the toolkit: “Baked‑in inefficiencies” (technical cost inefficiencies shielded by cartel conditions) and “cartel savings” (illegitimate common‑cost reductions from allocation/coordination) now have judicial recognition and a practical allocation method (paras 445–458). Expect these to be pleaded and tested in future follow‑on cases.
Negotiation evidence is probative: Courts may interrogate negotiation narratives and discount dynamics to test whether margins were competed down, even in a cartelised environment (paras 159–166, 441–444).
Regulatory regimes and damages: Where regulatory caps channel future “excess profits” away from the claimant, they will not reduce cartel damages; deductions for “benefits” will be resisted unless the claimant actually retains them (paras 531–541).
Financing structure shapes interest recovery: A claimant that is equity‑funded cannot recover “capital costs” that are in substance the shareholders’ expected returns (paras 545–549).
Gist damage clarified for cartels: Accrual of the cause of action is tied to reduced consumer benefit/competitive choice, not just a mechanical price uplift (paras 422–429).

Complex Concepts Simplified

Overcharge: Market‑wide counterfactual

Overcharge equals actual price minus the competitive price that would have prevailed without the cartel, whether the winning supplier would have been the defendant or a rival. This prevents a cartel member from shielding itself by assuming away competition (paras 15–18).

Direct vs common costs

Direct costs: Project‑specific inputs (materials, manufacturing, installation, risk contingencies). Here, ABB’s direct costs were largely accepted as honestly compiled (para 260), except for baked‑in inefficiencies.
Common costs: Selling, general and administrative overheads and profit. These can mask cartel effects (e.g., “cartel savings”); the Court looked at margins to detect inflation/savings (paras 266–272, 361–364).

“Baked‑in inefficiencies”

When a cartel suppresses competitive pressure, firms may drift into higher‑cost designs/processes. Even after the cartel ends, those designs may persist. In a counterfactual competitive tender, the inefficient firm must concede price to match efficient rivals. The Court quantified this by shaving 15% off copper cost (a conservative proxy), moving that amount from direct cost into a margin reduction (paras 445–453).

“Cartel savings”

The opposite effect: cartel allocation reduces competition costs (fewer genuine tenders, more predictable factory loading), inflating margins relative to a competitive world. The Court allocated a 1.9% adjustment to BritNed by spreading the observed cartel margin uplift across expected winning and losing bids (paras 454–458).

Why the regression failed

Mixing underground and submarine projects assumes away deep, structural differences; a simple “submarine dummy” could not fix that (paras 388–397).
Time trend and order backlog proxies were not persuasive; their presence/absence swung results, undermining robustness (paras 398–413).
Large confidence intervals and sensitivity to modelling choices reduced probative value (para 418).

IRR cap and damages

Even if an overcharge incidentally raises the IRR threshold, it does not mean the claimant is “no worse off” when “excess profits” are not the claimant’s to keep. Benefits that do not actually accrue to the claimant (e.g., earmarked for capacity expansion or to regulated networks) are not netted off (paras 534–541).

Conclusion

BritNed v ABB is a milestone in private competition litigation in England and Wales. It clarifies that:

Overcharge is assessed against a market‑wide competitive counterfactual, not just the defendant’s own hypothetical price;
No presumption of harm applies pre‑Damages Directive, but courts will use a pragmatic “broad axe” to quantify loss;
Only the operative part of a Commission decision binds; recitals bind only if they are the essential basis;
In bespoke project settings, a project‑level margin analysis anchored in actual cost data is preferable to broad regressions reliant on weak proxies;
Courts may recover cartel harm via two distinct levers: removing product‑level “baked‑in inefficiencies” and stripping out “cartel savings” embedded in common costs;
Regulatory caps that divert excess profits away from the claimant do not reduce compensation; and
Compound interest claims must track the claimant’s own financing reality—equity funding does not generate a recoverable financing cost for the company.

The result—an overcharge award of €13,009,568 with simple interest, the rejection of lost profits and compound interest, and no deduction for the IRR cap—sets a careful, evidence‑sensitive template for future cartel damages cases. Above all, the Court’s structured recognition of “baked‑in inefficiencies” and “cartel savings” offers a practical lexicon and methodology for quantifying harm in markets where projects are bespoke, data is imperfect, and counterfactuals are inherently complex.